Grocery store software in Baltimore: SNAP wants 28 staples by November 4, and Maryland’s new pricing law covers the delivery app, not the corner store

Three grocers in Baltimore have a busy month ahead. The first runs a corner market in West Baltimore, a rowhouse store with two coolers, a deli slicer and a steady line of SNAP customers. He replaced his card reader in September because the State was switching off the old stripe-only EBT cards, and he assumed that was the paperwork for the year. It was not. A federal rule that retailers must implement by November 4 requires him to carry seven varieties in each of four staple food categories, and by its counting rules his dairy case, which looks full, holds five. An inspector can arrive unannounced, and the only defense the rule allows is an invoice dated within 21 days of the visit.

The second owns a Latin American grocery in Highlandtown with her brother. Half their online orders come through two delivery apps, where the same plantains, masa and queso fresco cost more than they do at the counter, and customers have started to ask why. A Maryland law that took effect on October 1 forbids a delivery app to use a shopper’s personal data to charge that shopper more for groceries, but it does not apply to her store at all, because it only covers food retailers of 15,000 square feet or more. She would like her own online store, one where families can pay with SNAP, and has no idea what that involves.

The third operates an independent supermarket in Baltimore County, well above the 15,000-square-foot line, with a loyalty app and a quote for electronic shelf labels on his desk. Since October 1 every price his store or its app shows a shopper is subject to the new law, and he cannot say, for any price on any day, why it was what it was. His loyalty program is probably fine. The model his coupon vendor uses to decide who gets which offer, he is less sure about.

Each of them has a software problem dressed as a compliance problem. The questions the rules now ask, how many varieties are on the shelf, which items are SNAP-eligible, which barcodes the WIC program recognizes, which prices were personalized and why, what a bag costs in the City and in the County, are all answered by the way a store keeps its item file, its prices and its invoices, and the answers have to exist before anyone asks. This article goes through those questions for Baltimore City and Baltimore County, using the statutes, the federal rules and the State’s and City’s own documents, a look at how 59 independent Baltimore grocers and 85 independents nationwide show up online, and what the software costs. It covers the online store too, because federal regulators now say outright that they expect online SNAP purchasing to fill some of the gaps their own rule may leave. It is our reading of the rules for the products we build, not legal advice.

Baltimore’s grocery market, by the numbers

The best single picture of who sells food in Baltimore comes from the federal government’s list of stores that accept SNAP, because nearly every grocery store, corner store and convenience store in the city is on it. The U.S. Department of Agriculture publishes it as a retailer locator layer, updated every two weeks, and as a historical file of every authorization and withdrawal since 2005, current to the end of 2025; we counted both. As of October 7, 2026 it listed 584 SNAP-authorized retailers in Baltimore City and 498 in Baltimore County. The two jurisdictions sit side by side on a map and look nothing alike in the data.

Two numbers in that table explain most of this article. In Baltimore City, convenience stores are 63.5% of all SNAP retailers, against 43.8% nationally, and there are 7.1 of them for every supermarket or super store, against 1.7 in Baltimore County and 2.7 in the country as a whole. And by simple name rules (no chain, franchise, dollar or drug store in the name) about 434 of the City’s 584 retailers, roughly three in four, look independently owned, including 320 of its 371 convenience stores and every one of its 65 grocery stores. In Baltimore County the independent share is closer to one in three. A rule written for corner stores, or a cost that falls on independents, falls on the City’s food system far more heavily than on the County’s.

The trend is the other half of the story. The City’s count of SNAP retailers peaked at 1,000 at the end of 2015 and fell to 609 by the end of 2025, and the category that shrank fastest was the grocery store between a corner store and a supermarket: small, medium and large grocery stores together went from 214 to 68 over those ten years, while the count of supermarkets and super stores barely moved, from 52 to 55. Between the end of 2025 and October 7, 2026, 103 City retailers dropped off the list and 78 joined or rejoined it, most of them convenience stores in both directions. The County’s count has been roughly flat at about 500 for a decade. Census Bureau employer data tell the same story from the payroll side: in 2023, the latest year published, Baltimore City had 231 supermarket and grocery establishments with paid employees, 16% fewer than in 2018, and 72.7% of them had fewer than five employees, against 40.2% in Baltimore County, where the average grocery establishment employs about 48 people to the City’s 17. The County Business Patterns file also counts 166 City convenience stores with employees, down 21% since 2018, and it leaves out the many stores run by their owners alone. The churn underneath is mostly ownership. In every year from 2019 to 2025, between 72% and 91% of new SNAP authorizations in the City were at an address that already had a SNAP store, which is to say a new owner behind the same counter, with a new authorization, a new set of terminals and, as the WIC section below explains, no WIC.

The City’s own planners describe the same market from the shelf. Baltimore’s 2024 Food Environment Brief, published by the Department of Planning in 2025, counts 53 supermarkets run by 22 different operators, 428 independently owned small grocery and corner stores whose average Healthy Food Availability Index score is 9 on a scale that runs to 28.5, and five public markets. It found that in 2024 an average of 27% of City residents were enrolled in SNAP and 3% in WIC; that 479 of the 703 small grocery, corner and convenience stores in its count accept SNAP, “yet most only carry the minimum stocking requirements and serve primarily as places for snacks and drinks”; and that only 19 of the 428 independent small grocery and corner stores accept both SNAP and WIC, down from 103 in 2018. Stores that accept SNAP scored about 30% higher on its healthy food index, and stores that accept WIC about 40% higher.

The share of residents living in what the City calls Healthy Food Priority Areas, which it called food deserts until 2018, fell from 23.5% (146,077 people) in 2018 to 16.2% (94,472) in 2024, partly because the City’s population fell by about 5% over the same period. About 84% of the people who live in those areas are Black. Benefits make these stores matter more than their size suggests. In August 2026, according to the Department of Human Services’ monthly report, 138,208 City residents in 86,618 households received $27.2 million in SNAP benefits, 22.8% of Maryland’s total, while the City has about 9% of the state’s population; that works out to roughly 2,660 SNAP recipients for every City supermarket or super store, against about 810 in Baltimore County. Nationally, the independent grocer’s margin for error is thin: the industry association FMI reports an average food-retail profit margin of 2.1% for 2025, and the National Grocers Association’s 2026 study of independents found shrink at 3.9% of sales and online orders at just 1.1% of revenue, in a year when online orders reached 20.9% of all US grocery spending in the third quarter, according to Brick Meets Click. The USDA’s Economic Research Service forecasts food-at-home prices to rise 2.4% in 2026, and beef 9.4%, which means more price changes on more shelf tags.

We want to be fair to the software companies that already serve this market, because most of them are good at what they do. Grocery point-of-sale vendors handle scales, scanners, EBT, eWIC, age checks and tax tables that general retail systems still get wrong; the delivery marketplaces bring customers to stores that could never afford their own drivers; and a few online ordering platforms built for independent grocers do a solid job of putting a store’s catalog online. What none of them models is the part of a Baltimore store’s week that this article is about: the federal shelf count, the State’s WIC file, the new pricing statute, the City and County line on a bag, and the difference between what a family pays at the counter and on an app.

What grocery software costs, and who pays the meters

A Baltimore grocery runs on four pieces of software that are sold separately and priced in four different ways: the register, the payment processing behind it (with EBT and WIC as separate rails), the online store or ordering platform, and the delivery marketplaces. Electronic shelf labels are becoming a fifth. We checked each vendor’s own pricing page on October 18, 2026, and counted a price only if the vendor publishes it today.

The register, and whether it takes EBT

General retail systems publish their prices, and most of them are reasonable for a small store. Square for Retail is free, $49 or $149 a month per location, with in-person card fees of 2.4% to 2.6% plus 15 cents; Shopify POS Pro is $89 a month per location; Lightspeed Retail is $89, $149 or $289 a month on annual billing; KORONA POS is $59 to $99 per terminal. Systems sold to grocery and convenience stores are often cheaper to start: IT Retail and POS Nation each advertise a $49 starter plan, Modisoft $69 or $89, NRS $19.95 to $79.95 a month when the store also uses its payment service, and Toast’s retail point of sale from $90 a month with the first terminal included. Clover sells retail bundles that run from $16 a month over 36 months to $240 a month, or hardware up front plus $84.95 or $104.90 a month in software. The systems built for supermarkets, ECRS’s Catapult and LOC Software’s SMS among them, publish no price at all. One published plan is a trap for a real grocery: Erply’s $49 to $59 grocery plan caps net sales at $25,000 a month, and a $2.5 million store sells about eight times that.

The question that sorts this list for a Baltimore grocer is not price but EBT, and the answer changed this summer. Square added SNAP EBT at checkout on July 30, 2026, as an open beta on its Register, Terminal and Handheld in retail mode for sellers with an FNS number; it removes tax from SNAP items, splits a mixed cart automatically, and its pricing page lists EBT at 1.8% plus 5 cents a transaction. Much of the web still says the opposite, including a 2022 thread in Square’s own community that ranks first for the question, and until this year sellers who needed EBT on Square used an add-on such as TotilPay, at $62.95 a month plus a $155 card reader for SNAP and WIC. Shopify’s point-of-sale pages do not mention EBT at all; its route to SNAP is online, through a third-party app. Clover’s developer documentation lists EBT on six of its devices, including EMV chip and contactless cards “if supported by the specific EBT program in the state,” but also says plainly that online and offline EBT transactions are not supported and that taxes must not be applied to SNAP purchases. Toast says EBT runs at its retail register without a separate terminal. NRS sells an EBT add-on for a flat $49.95 a month with no per-transaction charge, and offers eWIC only in some states. The difference between a percentage and a flat fee is bigger than it looks: in the model store described below, with about $575,000 a year of SNAP sales at the register and an average SNAP ticket of $30, Square’s EBT rate comes to about $11,300 a year, and NRS’s flat plan to $599. For WIC, the State’s own manual distinguishes integrated registers from stand-beside terminals, and few small-store systems publish whether they are certified for Maryland’s eWIC at all.

EBT equipment is the store’s cost now

Many owners still believe the State hands out free EBT terminals. It used to, and the federal regulation still reads that way: 7 CFR 274.3 says authorized retailers “shall not be required to pay costs” of EBT equipment the State supplies for SNAP. The statute changed in 2014, though, and the statute wins. Under 7 U.S.C. 2016(f)(2)(A), stores must pay all the costs of acquiring EBT point-of-sale equipment and supplies, and the Food and Nutrition Administration’s factsheet for new retailers says that “most retailers are required to pay for their EBT equipment and services,” with exceptions only for farmers markets, direct-marketing farmers, military commissaries, nonprofit food-buying cooperatives and meal programs. A corner store in Baltimore pays for its own terminal, including the update that lets it read the new chip cards described below. The same statute contains one piece of good news: no interchange fees apply to EBT transactions, so whatever a processor charges for EBT is its own price, which is why one vendor can sell unlimited EBT for a flat monthly fee while another charges a percentage.

The online store, and the three processors that can take SNAP

A store that wants to take SNAP online has exactly three payment processors to choose from. The Food and Nutrition Administration’s online requirements page says the only approved third-party processors for SNAP online are Fiserv, Worldpay and Forage, and the National Grocers Association’s list dated June 16, 2026 names the same three. Only one publishes a price: Forage charges 3.5% plus 30 cents per transaction with no setup fee, which is 3.9% of a $75 order, and its Shopify app is listed as free, with eligibility tags on each product. Shopify’s own plans are $39, $105 or $399 a month with online card fees of 2.9% plus 30 cents on the cheapest plan. The platforms built for grocery e-commerce do not publish prices: Instacart’s Storefront Pro asks retailers to get in touch, and says its retail media revenue “can offset or fully cover your platform fees”; Freshop, which now belongs to NCR Voyix, and Mercatus quote privately. The category also has a churn problem worth knowing about before signing anything. Of the grocery e-commerce brands we checked, one domain now redirects to a gambling site, one no longer resolves, and one does not answer at all.

The marketplaces, and the parity clause in their pricing

Delivery marketplaces charge the store a share of every order and the shopper a set of fees. DoorDash’s published plans take 15%, 25% or 30% on delivery and 6% on pickup; Uber Eats takes 20%, 25% or 30% on delivery and 7% on pickup. Both pages are written for restaurants, and neither company publishes a grocery rate, so treat them as the closest public numbers rather than a grocer’s contract. The detail that matters is in the pickup terms. DoorDash’s 6% pickup rate requires that pickup prices match in-store prices, and Uber Eats charges 7% only with proof that in-app pickup prices equal in-store prices, 10% without it, which means both platforms already make a store document the relationship between its counter price and its app price. Instacart, which does not publish what it charges retailers, says in its terms of service that retailers, not Instacart, set the prices of goods, and that those prices may differ from in-store prices and vary across storefronts.

That is where higher app prices come from, and it is arithmetic rather than malice. To keep the same $100 after a 15% commission, a store has to sell the basket for $117.65 on the app; after 25%, for $133.33; after 30%, for $142.86. A markup applied to every shopper on an app is not what Maryland’s new law prohibits. A price raised for one shopper because of who that shopper is, is.

Shelf labels, the newest meter

Electronic shelf labels are the newest line in a grocer’s budget, and the least transparent. The large vendors, VusionGroup, Pricer and Hanshow among them, quote privately. One manufacturer’s published cost guide puts a system of 10,000 tags, enough for a mid-sized supermarket, at about $61,700 to $102,100 including the gateways and software, roughly $6 to $10 a tag, and the blogs of point-of-sale vendors cite $5 to $20 a label. What a store buys with that money is the ability to change a price on the shelf as easily as in the register, which is useful for markdowns on dated food and a liability if nobody can say afterward who changed what and why. Maryland did not ban them this year, and New Jersey’s new law freezes new installations for a year from February 2027.

What the stack costs a real store

To make the meters comparable, we modeled one store: a single 5,000-square-foot Baltimore grocery with $2.5 million in annual sales, 25% of them paid with SNAP, and 8% of sales, $200,000, taken online, with an average online order of $75. The online orders are where the choices diverge.

The comparison flatters the store’s own site in one important way, and we would rather say so than let a reader discover it. A marketplace commission buys drivers, customers and marketing, and the store’s own site buys none of those; the $7,368 option assumes the store either runs pickup only or delivers with its own staff. Even so, the gap is large enough to fund a lot: against a 15% marketplace, a store-owned site with Forage could spend up to $23,100 a year on building, hosting and delivery before it cost more, and against a 25% marketplace, up to $43,100. The register side of the same store, on a Clover retail bundle with card fees at its published rate, comes to about $28,900 a year, nearly all of it card processing that no software choice will remove.

Maryland’s new pricing law, read as a data model

On April 28, 2026, Governor Moore signed House Bill 895, the Protection From Predatory Pricing Act, as Chapter 154 of 2026. It took effect on October 1 as section 13-321 of the Commercial Law Article, and it is the first state law in the country aimed specifically at personalized pricing in food retail. The version that passed is much narrower than the one the Administration introduced, and the gap between the two is where most of the confusion about it comes from. When Governor Moore announced the bill in January, it would have required grocery prices to remain fixed for at least one business day and defined dynamic pricing as varying prices within a business day based on demand. The House added the delivery apps and a requirement that any merchant using an algorithm or personal data to set a price print the words “THIS PRICE WAS SET BY AN ALGORITHM OR BY USING YOUR PERSONAL DATA.” The Senate Finance Committee struck that disclosure, replaced the definition with one about personalized prices and limited the ban to higher prices for a single consumer, and a Senate floor amendment moved the exemptions into a list of their own. The House passed the bill 96 to 32, the Senate 41 to 1, and the House accepted the Senate’s version 100 to 31 on April 11. A separate Senate bill that would have banned electronic shelf labels in large grocery stores, SB 889, got a hearing in March and no vote. The United Food and Commercial Workers, which had endorsed the original, asked the Governor to veto the final version, as did several advocacy groups that objected to the exemptions and the bar on private lawsuits. Attorney General Anthony Brown’s review letter found the bill constitutional while warning that its mandatory cure period would “fundamentally weaken enforcement.” The Governor’s January and February press releases, still online, describe a rule that is not in the law.

What remains fits on a page, and every clause of it maps onto something a store’s software either knows or does not. It is worth reading slowly, because three words in it decide whether it applies to a given store at all: “15,000 square feet,” “exempt” and “higher.”

Who is covered: big stores and the apps, not the corner store

The law reaches two kinds of business. A “food retailer” is a merchant that operates a business establishment that “has a minimum of 15,000 square feet” and sells food that is exempt from Maryland’s sales tax under section 11-206(c) of the Tax-General Article. A “third-party delivery service provider” is a merchant that “facilitates as a consumer service the delivery of” that same tax-exempt food, and the definition expressly excludes food retailers. Nothing else is covered. A store below the size threshold is outside the law whatever it sells, and so is a restaurant, while a supercenter or a warehouse club that sells tax-exempt groceries is a food retailer just as a supermarket is.

For Baltimore that line falls in an interesting place. Supermarkets clear it easily. A corner store or a neighborhood market of two or three thousand square feet is nowhere near it, and most independent groceries in the city are below it too, so a small grocer selling from its own counter or its own website is outside the law entirely. But the delivery apps that sell the same store’s groceries are inside it. When a family in Park Heights orders from a corner market through Instacart, DoorDash or Uber Eats, the store is not a food retailer under the statute and the app, as we read the definition, is a third-party delivery service provider, which may not use the family’s personal data to raise the price of that order’s groceries. The law does not say how square footage is measured, whether as gross building area or sales floor, and because it defines the merchant, not the store, as the food retailer, a grocer with one large store and three small ones should assume, as we read it, that the ban follows its prices everywhere, including on its own website and app. For the handful of independents that sit near the line, those are questions for counsel, not for software.

What is covered: your tax table decides

The second limit is on products. The prohibitions apply only to food that is exempt from sales tax under section 11-206(c), and that subsection exempts “a sale of food for consumption off the premises by a food vendor who operates a substantial grocery or market business at the same location,” except food the vendor serves on the buyer’s premises and food for immediate consumption. Maryland’s tax code defines the second category broadly: hot food, sandwiches, salad and soup bars and party platters are all taxable food for immediate consumption. The same section says soft drinks, candy and alcohol are not “food” at all. So the hot bar, the deli sandwich, the soda cooler, the candy rack, the household goods and the over-the-counter aisle are all outside the pricing law, and the same flag that tells a register not to charge 6% on a gallon of milk is the flag that tells it the milk is protected. We wrote about how that tax line runs through a deli counter in our article on specialty food stores; for this law it becomes the scope of a consumer protection statute, which means a store whose tax flags are wrong has a second problem now, not just a tax one.

What is prohibited: higher, personal, and only for one

The operative clause has two parts. A covered business may not “engage in dynamic pricing to set a higher price” for tax-exempt food “for a specific consumer,” and may not “use personal data to set a higher price” for that food “for a single consumer.” Dynamic pricing is defined, unusually, not as prices that move during the day but as “the discriminatory practice of offering or setting a personalized price for a good or service that is specific to a consumer based on the consumer’s personal data, regardless of whether the seller collected or purchased the personal data.” Personal data borrows the definition from Maryland’s Online Data Privacy Act: any information linked or reasonably linkable to an identified or identifiable consumer, excluding de-identified data and publicly available information.

Three things follow, and none of them is what the headlines suggested. Prices that change for everyone, by the hour or by the day, are not dynamic pricing under this definition; an electronic shelf label that drops the price of bananas at 6 p.m. for every shopper is untouched. Prices that are lower for one person are not prohibited either, which is why loyalty discounts and personalized coupons survive. And the earlier drafts’ reach to “a group of consumers” did not survive into the final text, so the prohibition is about a price aimed at one identifiable person. What the law forbids is the thing a modern pricing system can do and a paper tag cannot: notice who you are and charge you more for the same can of beans.

Nine exemptions, read as reason codes

The law then lists nine situations to which the prohibition does not apply. Read by a lawyer they are defenses; read by an engineer they are the reason codes a covered business should attach to every price that differs from its shelf price, so that the explanation exists before anyone asks for it.

The statute imposes no record-keeping duty of its own, but the Attorney General can subpoena records under the Consumer Protection Act, and a 45-day cure period is only useful to a business that can find out quickly what its system did. A pricing system that writes one line per price change, with the product, the store or channel, the old and new price, the reason code and the evidence, can answer any question this law asks. A system that sets prices with a model nobody can explain cannot, and that is the practical difference between the two kinds of software a large grocer might buy.

The clause that reaches advertising

One subsection is broader than the rest. A covered business may not use “protected class data,” meaning information that alone or in combination directly or by implication identifies a characteristic protected from discrimination under state or federal law, to offer, advertise or sell any consumer good or service if the effect is to withhold an accommodation, advantage or privilege accorded to others. That clause is not limited to tax-exempt food, so it reaches digital coupons, app-only offers and the audiences a grocer builds for its ads. A coupon that only some neighborhoods ever see, chosen by a model fed with demographic data, is the kind of thing it was written for.

Enforcement: a letter first, then the Consumer Protection Act

A violation is an unfair, abusive or deceptive trade practice under the Maryland Consumer Protection Act, but with two softenings. Before the Attorney General’s Consumer Protection Division can bring an enforcement action, it must send a notice of violation and give the business 45 days to cure; a cured violation cannot be the subject of an enforcement action. And the law “may not be construed to authorize a private right of action,” while a conforming amendment to section 13-411 takes the Act’s criminal penalties off the table. What remains are the Act’s civil penalties, which the fiscal and policy note puts at up to $10,000 for each violation and $25,000 for each repetition. The Attorney General’s office told legislators it would need an assistant attorney general and a technologist, about $325,000 a year, to enforce it, which says something about what the office expects to be investigating: pricing systems, not shelf tags.

Maryland was first, and it is not alone

The Governor’s office called Maryland the first state in the country to ban data-driven grocery price hikes, and the states that followed chose different designs. New York’s Algorithmic Pricing Disclosure Act, in force since November 10, 2025, bans nothing; it requires the statement “THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA,” and a broader New York ban passed both houses on June 4, 2026 and, as far as we could find, still awaits the Governor’s action. Connecticut enacted a surveillance pricing law in June 2026 that takes effect on July 1, 2027. New Jersey’s, signed on July 23, 2026, applies to groceries and household goods from August 1, 2027, allows private lawsuits, and freezes the installation of new electronic shelf labels for a year from February 1, 2027. At the federal level, the FTC published initial findings from its surveillance pricing study in January 2025, reporting that the pricing intermediaries it examined had at least 250 clients “ranging from grocery stores to apparel retailers,” and has issued nothing final since.

The most consequential event was not a law. In December 2025, Consumer Reports, the Groundwork Collaborative and More Perfect Union published a study in which 437 shoppers in four cities saw the same Instacart items at different prices, an average gap of 13% between the highest and lowest price and up to 23%, and on December 22 Instacart announced that it was ending all item price tests and would no longer let retailers use its Eversight tool to run them. Retailers on Instacart can still price differently from store to store and above their shelf prices. What changed is that two shoppers looking at the same store on the same app at the same moment should see the same price, which is also, roughly, what Maryland now requires of the app.

What this means for a Baltimore grocer’s software

For a store under 15,000 square feet, the honest answer is that this law asks nothing of its own systems. It still matters, for two reasons. The first is the delivery apps: a small grocer that sells through a marketplace should know that the marketplace is covered, should know whether its own item prices on that marketplace are marked up and by whom, and should be able to compare what its customers pay there with what they pay at the counter, which few stores can do today. The second is that the law sets the vocabulary every pricing tool sold in Maryland will now be built around. A store that keeps a dated price history with reasons, keeps its tax flags correct, runs its loyalty program on documented enrollment, and never lets a model set a price it cannot explain, is compliant with this law at any size and ready for whatever the next one says.

For a grocer above the line, the software checklist is short and specific. Keep the tax-exempt flag accurate on every item, because it is the law’s scope. Log every price change and every price that differs from the shelf price with a reason code from the table above. Keep loyalty prices tied to an enrollment record. Keep personal data out of the inputs of any model that sets a regular price for exempt food, including data bought from someone else, since the definition reaches data “regardless of whether the seller collected or purchased” it. Audit coupon and advertising audiences for protected-class attributes and their proxies. And make sure the answer to a notice from the Consumer Protection Division can be produced in 45 days, which is a matter of how the data is kept, not of how quickly a lawyer can type.

Seven of everything by November 4: SNAP’s new stocking rule is an item master

Most rules that reach a grocery store tell it what not to do. The one SNAP retailers must meet this fall tells them what to have on the shelf, item by item, and it is written so precisely that it reads like a database schema. The Food and Nutrition Service published the final rule on May 8, 2026, made it effective on July 7, and gave retailers until November 4, 2026 to implement it. It finishes a job Congress started in the 2014 Farm Bill, which raised the minimum number of staple food varieties a SNAP store must carry from three to seven in each of four categories. The Department wrote those numbers into a rule in December 2016, and then every appropriations act from 2017 onward forbade it to enforce them until it redefined “variety” in a way that gave stores more ways to count. This rule is that redefinition, so the seven-variety standard now becomes real.

The arithmetic is in the regulation itself. Under 7 CFR 278.1(b)(1)(ii), a store that qualifies under Criterion A, which is how almost every corner store and small grocery qualifies, must offer at least seven distinct varieties in each of the four staple categories, “for a minimum of 28 distinct staple food varieties”; at least three stocking units of each, “for a minimum of 84 stocking units”; and one perishable variety in three of the four categories, so that at least three of the 28 varieties and nine of the 84 units are perishable. The four categories were renamed along the way and are now protein (including plant-based sources), grains, vegetables or fruits, and dairy (including plant-based alternatives). The items have to be offered “on a continuous basis,” which the rule defines as displayed for sale in a public area on any given day the store is open.

How the rule counts

The interesting part is how a variety is defined, because it decides whether a shelf that looks full counts as three varieties or seven. Foods are distinct when they differ by kind of plant, kind of animal or main ingredient. Brands, flavors, package sizes and preparations do not create a new variety, and the rule gives its own examples: different cuts of meat, whole and cut fruit, vanilla and plain yogurt, and a gallon and a half gallon of milk are each one variety. Different types of the same food do not count separately either, so brown and white rice are one variety, as are pinto and kidney beans and Granny Smith and Gala apples. For anything with more than one ingredient, the variety is set by the main ingredient, which the rule defines as the first ingredient on the label other than water, broth or stock. The preamble’s own example is worth remembering: fresh tomatoes, 100% tomato juice, a can of diced tomatoes and a jar of pasta sauce whose first ingredient is tomatoes are all one “tomato” variety.

On top of that default, the rule designates specific foods that always count separately. Shell eggs, perishable meat, poultry or fish (one per kind of animal), dry beans, dry peas, dry lentils, raw grains and flour (one per kind of grain) and perishable liquid milk count as their own varieties when they have no other ingredients, so seasoned chicken is a different variety from plain chicken and chocolate milk a different variety from plain milk. A second list counts as one variety each regardless of brand or main ingredient: tofu and tempeh, whole grain bread, other bread, whole grain pasta or noodles, other pasta or noodles, breakfast cereals, shredded cheese, other cheese, fermented dairy drinks, yogurt, sour cream, infant formula and infant cereal. A third list covers shelf-stable meat, poultry or fish, shelf-stable liquid milk and powdered milk, which count separately from their fresh versions. Plant-based milks and other dairy alternatives can supply up to three dairy varieties. And a long list of “accessory foods” never counts at all: chips, crackers, cookies, candy, soda, energy drinks, condiments, jams, sugar, spices and cooking oils, joined in this rule by snack bars, every kind of jerky, cheese and fruit spreads, and butter. A food whose first ingredient is an accessory food is itself an accessory food, which means a breakfast cereal that lists sugar first does not help the grains count. SNAP households can still buy accessory foods with their benefits; they just do not make a store eligible.

Here is what that does to a typical corner-store dairy case. Under the old three-variety standard, milk, cheese and yogurt were enough. Under the new one, the same case falls short:

None of the fixes is expensive, and that is the Department’s own view. What makes them easy to miss is that a person looking at a full cooler sees eight products and a regulator’s counter sees five varieties. Eggs, which many shoppers think of as dairy, count in the protein category, and a cooler full of sodas and sports drinks counts for nothing.

Who it hits, by the Department’s own numbers

The rule’s regulatory impact analysis is unusually candid. It counts about 269,000 SNAP-authorized retailers nationally as of June 2025, of which about 188,500 are smaller stores authorized under Criterion A: 119,688 convenience stores, 60,335 “combination grocery/other” stores, such as pharmacies and dollar stores, and 8,489 small grocery stores. Using the inventory counts from each store’s most recent eligibility visit, the Department estimated that 73% of convenience stores and 77% of small grocery stores would not meet the seven-variety standard as they stand, against 8% of combination stores, for about 99,200 stores nationally that need to add something. Fewer than 1,000 lack the third perishable category. The cost it estimates is small: about $15 a year in extra stock for a store that needs it, roughly $400 of a manager’s time to review inventory and reorder in the first year, and an opportunity cost of about $17 a year for shelf space given up. It also expects up to 5% of the stores that must add varieties, around 5,000, to be denied at their next review, against roughly 2,000 a year under the old standard.

The Department is just as candid about why it does not mind. In fiscal 2024, it says, convenience stores were 44% of all SNAP retailers and 76% of all SNAP sanctions but took in only 5% of SNAP redemptions, and 93% of all retailer sanctions were imposed on small-format stores. About 74% of redemptions happen at supermarkets and superstores, which are about 15% of retailers, and 10.2% already happen at internet retailers, which are 0.2%. The rule says in so many words that some currently authorized stores may become ineligible, and that it expects online purchasing to cover part of the gap. That sentence is the reason this article spends a whole section on selling groceries online with EBT.

In Baltimore the stakes are concentrated. At the end of 2025 the federal retailer file listed 390 convenience stores, 39 small grocery stores and 75 combination stores authorized for SNAP in Baltimore City. If they fall short at the national rates in the impact analysis, 73%, 77% and 8%, about 320 City stores need to add varieties, more than half of the City’s 609 SNAP retailers; the same arithmetic gives about 170 of Baltimore County’s 505. The City’s planners had already noticed the pattern: their 2024 food environment brief says most of the 479 small grocery, corner and convenience stores that take SNAP “only carry the minimum stocking requirements.” Dairy is where they are most likely to fail. Nationally, 48% of small SNAP stores have fewer than seven dairy varieties, against 37% short in protein, 19% in grains and 17% in fruits and vegetables.

What a store visit checks, and the 21-day receipt rule

Eligibility is checked in person. According to the impact analysis, a USDA contractor makes an unannounced visit, photographs the aisles and the overall inventory and records counts on a survey form, without reading ingredient lists. The regulation lets the Department rely on “visual inspection, sales records, purchase records, counting of stockkeeping units, or other inventory or accounting recordkeeping methods,” and it contains one provision that matters more to a store’s software than anything else in the rule: if a store does not meet the standard at the time of the visit, it may show documentation that it ordered or received the missing stock no more than 21 calendar days before the visit. An invoice from day 22 does not count, and neither does one nobody can find.

Timing is less dramatic than the November date suggests. From November 4, a new store, or an existing retailer opening a new location, is assessed under the new standard when it applies. Existing Criterion A stores are assessed at reauthorization, which generally comes every five years, but the regulation lets the Department review any store at any time, and it encourages every retailer to comply by the implementation date. A policy memo the agency issued on October 13, 2026 confirms that the standard does not apply to specialty stores such as butcher shops, fish markets and produce stands, which qualify on their sales instead, and that final determinations are made when a store applies for authorization or reauthorization. A store denied for falling short of the stocking standard has to wait six months before it can apply again, which for a corner store where SNAP is a large share of sales is a long time to go without it. The other route to eligibility, Criterion B, does not count varieties at all; it requires that more than half of a store’s total gross retail sales, food and non-food together, come from staple foods, which a store can only show from its sales records.

Be careful where you read about all this, including on the government’s own site. The agency’s main “Is My Store Eligible?” page, updated on October 5, 2026, still describes the old standard of 36 items, three varieties per category and perishables in two categories, still says that chocolate milk and regular milk count as one variety, and carries a banner that begins “Starting Nov. 4, 2024,” two years off. The calculator page says a specialty store qualifies if staple food sales are “at least 50%” of total sales, while the regulation says more than 50%. And many summaries call November 4 the date the rule takes effect, when it took effect on July 7 and November 4 is the date by which retailers must implement it. Search engines and AI assistants repeat whichever page they find first, so a store owner asking one of them how many varieties a store needs may well be told three.

What the software has to know

The Food and Nutrition Administration, as the Food and Nutrition Service has been called since June 2026, now publishes a stocking calculator on its website, which is useful and also a form an owner fills in by hand. Everything the rule asks for can be computed from data a store already has, if its item file carries a few extra fields. Each product needs its staple category, its variety key (the main ingredient, or the designated variety it belongs to), whether it is perishable, whether it is an accessory food, and whether it is a plant-based dairy alternative, so the count can stop at three. With those fields, a daily report can show the 28, the 84 and the 9 for each store, flag any variety that has fallen below three units on the shelf, and warn before a delivery cancels the only perishable in a category. The purchase side matters as much: invoices filed by date and searchable by item turn the 21-day rule from a scramble into a printout. And if a store qualifies under Criterion B, the same item file, with the staple flag set correctly, produces the sales mix it would have to prove.

The same fields do double duty at the register. A product’s SNAP eligibility is not the same thing as its staple status, since SNAP households can buy accessory foods and cold prepared foods but not hot prepared foods, alcohol, tobacco, vitamins or anything that carries a supplement facts label instead of a nutrition facts label. A point-of-sale system that already stores the eligibility flag correctly needs only a few more columns to answer the stocking question as well, which is the cheapest piece of compliance software a SNAP store will ever buy.

The EBT card changed this fall, and the register has to keep up

Maryland has been trying to put a chip in its EBT cards for three years. Chapter 172 of 2023 told the Department of Human Services to work with its vendors on “the inclusion of an embedded microchip in all Electronic Benefits Transfer cards issued on or after October 1, 2023.” The cards arrived in July 2026. First the State changed processors: on Sunday, May 17, 2026, every EBT card in Maryland stopped working for twelve hours while the system moved from Conduent to FIS, the old ConnectEBT portal became ebtEDGE, and cardholders gained a switch that lets them freeze a card from their phone. Then, after a pilot in Anne Arundel County, Baltimore City and Dorchester County in which 92% of chip transactions were approved and 98% of cardholders lived within ten miles of a store that could take a chip, the Governor announced the statewide rollout on July 17 and said that all magnetic-stripe cards “will be deactivated on Sept. 30, 2026.”

That date did not land cleanly. By August 23, about 58% of the nearly 507,000 new cards had been activated. On October 1, the Department said about 85% had been, that the new cards had been used for 4.8 million transactions since July, and that the stripe-only cards “will deactivate soon,” while promising validated fraud figures later in the month. By that arithmetic, roughly 76,000 of the new cards had still not been activated a day after the deadline. Officials describe Maryland as the fourth state to issue chip EBT cards; the federal agency’s own map listed eight states issuing them by September 2, Virginia among them, which says more about how fast this is moving than about who was first.

Why the State did it

The reason is skimming, and the numbers are large. According to the Department’s stolen-benefits data, as of October 16, 2026 it had received 126,926 claims since March 2023, approved 102,958 of them for 81,472 households and replaced $52.5 million in stolen benefits, about $42 million of it SNAP. Federal money covered thefts only through December 20, 2024. Since then Maryland has paid from its own budget under section 5-609 of the Human Services Article, which the same 2023 law created: the Department must decide a claim within 10 days and may not demand a police report, and the 2025 budget law capped restorations at $30 million a fiscal year. When the U.S. Secret Service swept Baltimore businesses on October 22 and 23, 2025, it checked about 3,000 payment terminals, gas pumps and ATMs at 493 locations and pulled out 22 skimming devices. A chip does not stop every scheme, but it stops the one that drains a card at the counter.

What the terminal has to do

The new cards still carry a magnetic stripe, so a store with an old swipe terminal was not cut off on September 30. Federal guidance for retailers says to insert or tap first, swipe if the terminal asks, and key in the card number only as a last resort, and to keep accepting the older cards. The technical rules a payment provider has to load are public: EBT has its own chip application identifier, every transaction still needs an online PIN, there is no offline approval and no PIN bypass, phone wallets are not supported, and tap is optional for the merchant. Certification runs through the store’s acquirer or processor, which receives test cards from the EBT processor, and there is no State program that hands grocers a free chip terminal; Maryland’s equipment grants cover farmers markets only. In practice the failure that hurts a store is not an old swipe reader but a chip terminal that was never updated for EBT and keeps rejecting the chip, which is why federal guidance asks stores to log the error, the time and the amount and send them to their provider. In September the federal agency also asked payment processors to push retailers to allow refunds only to a card that made a matching purchase, to require manager approval for refunds and to watch for card numbers being tested in quick succession.

The rules a register applies to a SNAP tender

A grocery register treats SNAP differently from any other tender, and each difference is a line of configuration. Food bought with SNAP is never taxed: under section 11-206(b) of Maryland’s Tax-General Article, food stamp eligible food bought with SNAP is exempt, so a soda that carries 6% sales tax for a cash customer carries none for a SNAP customer. Prices and terms must be the same as for cash. A valid card from any state must be accepted. Hot food is not eligible and neither is anything sold with a supplement facts label, which is how the agency decides whether an energy drink counts as food or as a supplement. And a store’s bag fee can never be paid with SNAP, because a bag is not food, so in Baltimore City and Baltimore County, neither of which exempts benefit shoppers from the five-cent charge, a customer paying entirely with SNAP needs a second tender for a nickel. A 2025 bill that would have stopped local bag fees from applying to SNAP and WIC purchases died after one hearing; Montgomery County has exempted any transaction paid partly with food assistance since January 1, 2026, which is one more rule a grocer with stores in more than one county has to encode.

Restrictions next door, not here

Eighteen states now hold federal approval to bar SNAP purchases of soda, candy or other items, according to the agency’s food restriction waiver page. Maryland is not one of them, and neither are Delaware, Pennsylvania or the District of Columbia. Virginia’s restriction on sweetened carbonated drinks starts on November 1, 2026, with a 90-day grace period for retailers. West Virginia’s was vacated, with those of Colorado, Iowa, Nebraska and Tennessee, by a federal court in Washington on June 22, 2026 in Aragon v. Rollins, and the government has appealed. The agency’s compliance memo answers the question a border-region grocer will ask: a walk-in store applies the rules of the state it is in, for every card, while an online order filled from a warehouse follows the rules of the state that issued the card, identified by the card number’s prefix. A Baltimore store therefore blocks nothing at its register, even for a Virginia family; a Baltimore business that ships groceries from a warehouse to Virginia cardholders would have to block what Virginia blocks.

The money moving through the card

The benefits on those cards are the largest single source of revenue many Baltimore stores have. In August 2026, Baltimore City and Baltimore County households received $45.3 million in SNAP benefits between them, and the two jurisdictions account for 38% of Maryland’s SNAP cases. That money is less certain than it used to be. During the federal shutdown in November 2025, the Department of Agriculture cut November allotments to half before the government reopened on November 12, and Maryland put $62 million of its own money onto cards in the meantime. SNAP is currently funded by a stopgap law through December 11, 2026. Under the 2025 federal budget law, states with high payment error rates will pay part of the cost of benefits from fiscal 2028, and Maryland’s fiscal 2025 error rate of 13.08% puts it in the bracket that pays 15%. For a corner store where SNAP is a third of sales, software that shows sales by tender by day is not a luxury; it is how an owner sees an issuance problem on the day it happens instead of at the end of the month.

WIC keeps its own item master, keyed by barcode

If SNAP’s stocking rule reads like a schema, WIC’s is an actual database. Maryland WIC’s authorized vendor manual describes the Approved Product List as “a database that contains Universal Product Codes (UPCs), and Price Lookup Codes (PLUs), of all Maryland WIC approved foods,” maintained by state staff, and it states the consequence plainly: “A WIC transaction cannot be completed unless the UPC or PLU of the item being purchased is in the APL.” Manufacturers change package sizes and barcodes constantly, so the list is never finished; a store can submit a missing product through an online form or the State’s APL checker app, and a newly approved code can take up to 72 hours to reach the store’s file. When loose produce fails to ring up as WIC-approved, the manual’s advice is that the store’s IT staff or point-of-sale provider map it to the right PLU. In other words, whether a mother can buy a bag of apples with her benefits depends, at the register, on whether a grocer’s item file and the State’s file agree.

The other half of the item master is a minimum stock list. Maryland has two tiers, and the one for stores with one to three registers, revised in February 2026, reads like a picking ticket: two gallons of whole milk and three of 1% or fat-free; three pounds of cheese in two varieties; six dozen eggs; five cans of beans in two varieties; three jars of peanut butter; fifteen ounces of canned fish; cereal, juice, whole grain bread or its substitutes, infant formula, infant cereal and infant fruits and vegetables, each with its own size range and count; and at least $16 worth of fruits and vegetables across two kinds of fruit and three of vegetables. Expired food does not count, and the stock has to be in the store or its stockroom during business hours. Like SNAP’s rule, it is a list a store’s inventory system could check every morning, and almost never does.

The price rules a register has to respect

WIC does not pay whatever a store charges, and the limits are written as arithmetic. The State groups stores into peer groups by size within regions and computes the average price vendors charge for each WIC food every month. Under COMAR 10.54.03, the price of each food package at an authorized store must stay at or below 125% of its peer group’s average. A store over the line is notified and has 10 days to lower its shelf prices; a store over it in two consecutive reporting months loses its authorization for 12 months, and so does a store that fails to return a requested price list within 10 business days. The same regulation sets the checkout rules: no sales tax on WIC purchases, no rain checks, scan the barcode actually on the item, accept cents-off coupons, and let WIC shoppers use “bonus or club cards, buy one, get one free, and any other promotion offered to store customers that does not involve cash back.” The vendor manual adds that a vendor “shall not charge WIC participants a higher price for food than other customers,” a principle Maryland’s new pricing law has now extended, in its own narrower way, to every shopper of a large grocer or a delivery app. It also requires that the shelf price of every WIC-approved food be displayed “at, on, or near the item,” with no stickers or labels on the WIC items themselves. The manual also says Maryland does not allow stores that take in more than half of their food sales from WIC; the regulation is slightly more forgiving and allows such a store only where shoppers would otherwise lack access, paying it the statewide average price.

The stakes run across programs. Under SNAP’s regulations, a WIC disqualification can lead to a SNAP disqualification that cannot be appealed through SNAP’s own review process, and a change of ownership or a move to a new address ends a store’s WIC authorization outright, so a new owner who buys a corner market cannot simply keep taking WIC. Vendors must also attend annual training each August to stay authorized.

Two kinds of WIC register

The manual distinguishes the two setups a store can have, and the difference is a software decision. Most chains run an integrated system in which eWIC is simply another tender: WIC and non-WIC items go into one transaction, the system matches each scanned barcode against the APL and the family’s benefit balance, prints a review slip, and takes the rest by card, cash or SNAP. Smaller stores usually have a stand-beside terminal, on which the cashier separates WIC foods from everything else, scans the eWIC card, has the shopper enter a PIN, scans each item and keys its price, then rings up the rest of the order on the regular register. A stand-beside terminal works, but it doubles the cashier’s work at the moment the line is longest, and it keeps WIC sales outside the store’s own sales data unless someone copies them back in.

WIC online is still a proposal

Families can use SNAP online; they cannot yet use WIC. The federal rule that would allow online WIC ordering was proposed in February 2023 and, as of this month, has not been finalized; the government’s regulatory agenda had targeted July 2026 for a final rule, and that date has passed without one. The proposal would remove the requirement that a WIC purchase happen in front of a cashier, would bar stores from charging the WIC program for delivery, service, convenience or bag fees while letting WIC shoppers pay those fees with another tender, and would require the shopper to see the total before the purchase is completed. We found no Maryland pilot. For a store, the practical advice is to build the online catalog so that the WIC list status of each item is already a field: when the rule arrives, the stores whose item files already know which barcodes the State recognizes will be the ones ready to sell.

What the City and the County add

The federal and state rules apply the same way on both sides of the city line. The local ones do not, and a grocer with one store in Highlandtown and another in Dundalk is running two sets of rules through one register.

A bag fee that works like a small tax return

Baltimore City banned thin plastic checkout bags in October 2021 and put a five-cent surcharge on every other bag a store provides, whether at the register, at pickup or with a delivery. Under Article 28, Subtitle 32 of the City Code, the surcharge “must be itemized on any receipt,” and the store keeps 4 cents of each nickel and remits the other cent by the 25th of the following month. The Finance Department’s regulations add the details a point-of-sale system has to know: each store registers on the City’s bag-tax portal and chains need a separate tax ID per location; a monthly report is filed per location; no payment is due in a month with fewer than 300 bags, but the count carries over until it reaches 300; partial payments are not accepted; and the City’s cent is excluded from the sales tax base. A late payment costs 1% a month plus a 10% penalty, and a missing report $1,000 a month. A store that hands out 2,000 bags in a month collects $100, keeps $80 and sends the City $20, and the code offers no exemption for SNAP or WIC shoppers. Bags used only to carry meat, fish, loose produce or nuts, unwrapped baked goods, ice or prescriptions are outside the ban.

Baltimore County went a different way. Its Bring Your Own Bag Act, Bill 1-23, has required stores since November 1, 2023 to charge at least five cents for each paper or reusable bag, keep all of it and show the number of bags and the charge on the receipt. An exemption for SNAP, WIC and pandemic EBT customers was in the bill as introduced and was struck before it passed. And it exempts “small retail establishments,” stores with three or fewer locations operating only in Baltimore County or the jurisdictions next to it, but not convenience stores or gas stations with convenience stores. A family grocer with one store in the City and two in the County therefore charges and remits the nickel in the City and may charge nothing in the County, while a convenience store on either side of the line charges it. Neither rule is hard; both have to be in the register, by store, before the first receipt prints.

The code still says “food desert”

The City stopped using the term “food desert” in 2018 and now maps Healthy Food Priority Areas, places where the average Healthy Food Availability Index score of the stores is low, median household income is at or below 185% of the federal poverty level, more than 30% of households have no car, and the nearest supermarket is more than a quarter mile away. The latest map, dated November 2024, has 55 such areas, and the City estimates that 94,472 people, 16.2% of residents, live in them. The tax code never caught up with the vocabulary. Article 28, section 10-30, still titled “Food Desert Incentive Areas,” gives a newly built or substantially renovated supermarket in or within a quarter mile of such an area a credit of 80% of its personal property tax for ten tax years. To qualify, the store needs all the major food departments, more than half of its sales and more than half of its floor space in food, at least 500 square feet each for produce and for other perishables, and a minimum investment in new equipment: the greater of $150,000 or $25 a square foot for a new store, and $5, $8 or $10 a square foot for a renovation, depending on size. The application is due at least 90 days before the first tax year. Every one of those tests is a number a store’s systems can produce, from sales by department to floor area by category, and a store that cannot produce them will struggle to keep the credit for ten years.

A new grant that pays for registers

The newest local rule is a grant program, and it takes effect on the same day as the federal stocking deadline. The Community Healthy Food Access Act, City Council Bill 26-0163, signed by the Mayor on October 5, 2026, becomes effective on November 4. It creates a Healthy Food Priority Area grant program in the Planning Department’s Food Policy and Planning Division, with grants of up to $250,000 for a “small box” grocery under 12,000 square feet or a community food retailer and up to $100,000 for a larger grocery store, for stores in or within a quarter mile of a priority area. Applicants must accept both SNAP and WIC, or show a plan to do so within 90 days; the money may pay for refrigeration, build-out and “point-of-sale systems or electronic benefit transfer equipment”; and each recipient’s annual report must include the number of SNAP and WIC transactions it processed. Applications must open within 180 days. The catch is the budget: the Finance Department’s memo shows $175,088 in general funds for these grants in fiscal 2027, less than one maximum grant. For a corner store that wants to take WIC again, the grant is still the first public money in years aimed at exactly the equipment and records this article describes.

Baltimore sold groceries online to SNAP shoppers in 2010

The City has tried this before, and early. In 2010 its Baltimarket program launched a Virtual Supermarket: residents ordered groceries online at libraries and housing sites, picked them up there on a set day each week with no delivery fee, and could pay with SNAP years before the federal government allowed SNAP to be used for online orders. When its only grocery partner closed in 2013 the program paused, and it came back in 2014 with ShopRite; by 2018 it served more than 530 households at 13 sites. We found no sign that it operates in 2026, and the program’s old web address now belongs to an unrelated Swedish moving and cleaning site with casino links, although AI search summaries still send people to it. The City’s current effort, B’more Fresh, used federal recovery funds to promote online SNAP purchasing through Giant, Safeway and Amazon. The idea Baltimore pioneered, ordering online with benefits and picking up nearby, is now available nationally to any SNAP store that sets it up; the census below shows how few of the City’s independents have.

Licenses, dates and hours

A few more local rules belong in the item file and the calendar. A City food service license renews on its own anniversary rather than on a common date, at $520 a year for a high-priority facility, $285 for moderate and $65 for low, while the County’s annual permits cost $540, $340 and $135. The City’s Health Code forbids selling perishable dated foods such as milk, cheese, meat, eggs and baby food past their date, allows other past-date food only in a separate area under an 11-by-14-inch sign announcing it, and prohibits changing or covering a date, which is a rule an inventory system with expiry dates can enforce and a clipboard cannot. And the zoning code shapes what a rowhouse corner store can be: a neighborhood commercial establishment in a residential zone may not sell alcohol or tobacco and must keep its displays indoors, and any business in a residential zone or one of the City’s lower-intensity commercial zones that operates between midnight and 5 a.m. needs a late-night license, which matters to a store that wants to offer pickup windows around the clock.

What 59 independent Baltimore grocers show a shopper, and a machine

To see how Baltimore’s independent grocers show up online, we built a frame from the same federal retailer data. We took every SNAP-authorized grocery store, supermarket, super store and combination store in Baltimore City and Baltimore County as of October 7, 2026, and removed national and regional chains, any banner with more than ten SNAP stores of the same name in Maryland and its neighbors, gas stations, pharmacies and duplicate addresses. We set aside 17 stores that operate under a licensed banner such as Save A Lot or ShopRite, because their websites belong to the banner, and kept local operators with ten stores or fewer. That left 142 independents, 80 in the City and 62 in the County. We drew a random sample of 60 with a fixed seed, found that one belonged to a regional chain after all, and read what the other 59 put online on October 18, 2026. Every one of them takes SNAP at the register.

The first finding is about presence. Only 11 of the 59, about one in five, have a working website of their own: 5 of 34 in the City and 6 of 25 in the County, and 6 of the 13 larger-format stores against 3 of the 42 smaller grocery stores. Our method, which probed likely domains, read OpenStreetMap and searched by name, can miss a site, so treat the rest as an upper bound on stores without one. Several of the rest exist online only as a directory listing, a Facebook page or a marketplace storefront, and for one store with no site of its own the top search results were spun pages on hijacked subdomains describing it as a “global leader” with certifications it does not claim. A store that does not publish its own facts leaves the job to whoever does.

The second is about what the sites say. Of the stores with a website, only three run a real grocery cart; the others offer a list form, catering or sandwich ordering, a delivery service with no checkout, or a “Shop Now” button that leads to an error page. None of the sites mentions SNAP or EBT anywhere, although every one of these stores accepts it in person, and one mentions WIC. One shows its delivery and pickup fees, four post a weekly ad, none mentions a bag fee, and none uses the schema.org GroceryStore type that tells search engines what kind of business it is. Two have an llms.txt file, both generated automatically by their platforms, and none blocks AI crawlers. Shopify’s default robots.txt file now addresses AI shopping agents directly and points them to an endpoint for agent instructions, which means one store in our sample is better prepared for AI shopping assistants than its owners may know.

The third finding is on the marketplaces, and it is the sharpest. Instacart’s Baltimore page listed 106 stores on the day we read it. Of 42 chain grocery, club, drug and dollar store listings, 32 carried an “Accepts EBT” tag; of 17 independent grocery listings, none did, and the help pages of the independents we opened said each store “does not currently accept EBT SNAP,” while Giant’s said it does. Only 2 of our 59 sampled stores are on Instacart at all, and about 10 of the 142 in the frame. The federal government’s own list of retailers accepting SNAP online in Maryland tells the same story from the other side: of its 39 entries, the only independent grocers are two Baltimore County stores, Seven Mile Market in Pikesville and Geresbeck’s Food Market in Middle River, alongside the national chains, two farms that sell through a farm marketplace, and Rite Aid, whose last Baltimore-area stores had dropped off the federal retailer file by the end of 2025. In a city where 24% of residents receive SNAP, an independent grocer’s online customers effectively cannot pay with it.

Some stores do things well, and they deserve to be named. Harvest Fare runs pickup ordering on Freshop, posts a weekly ad and says plainly that it is WIC approved, and on Instacart it is one of only two Baltimore listings with a “No markups” tag. Eddie’s of Roland Park offers free pickup and delivery “starting as low as $7” depending on the ZIP code, publishes a weekly circular and has an llms.txt file. Koko Market and Swadesh both run priced catalogs with a cart. Graul’s Market posts weekly circulars and digital coupons. Harford Seafood Market publishes LocalBusiness markup and its hours, and All Halal Meat & Groceries lists its full weekly hours and takes orders through a form. None of them is yet the store this article describes, with SNAP online, eligibility on every product and its shelf visible to search engines, but each has done the first part of the work.

What 85 independent grocers across the country show online

To see whether Baltimore is unusual, we ran the same exercise nationally. From the federal retailer file, current to the end of 2025, we took every independent supermarket and grocery store in the core counties of 22 metro areas, from Philadelphia, Pittsburgh and Washington to Chicago, Denver, Phoenix and Seattle, plus seven metros in states with SNAP purchase restrictions, removing 1,601 chain stores by name and leaving 3,226 independents. We drew four larger stores per metro at random, 88 in all, and found a website for 28 of them, 32%; searching by hand for some of the rest suggests the true share is nearer 40%. To have enough sites to read, we added stores from the same frame whose websites are recorded in OpenStreetMap, which tilts toward stores that are already online, and read 85 sites in all on October 18, 2026.

The comparison with Baltimore depends on store size. Baltimore’s larger independents are, if anything, more often online than the national sample’s, with 6 of 13 running a site of their own against 28 of the 85 larger stores in the national random sample; its small grocery stores, at 3 of 42, are much less. Among the 85 national sites, 36, or 42%, offer some way to order online, but few run grocery software to do it. Ten use Instacart, six by linking to their marketplace page and four through an Instacart storefront under their own name; eight use Shopify, four WooCommerce, five only link to a delivery app, two use Square Online, and Freshop, Mercato and Toast run one each. Software built for grocers powers 6 of the 85 sites.

SNAP online is almost as rare nationally as in Baltimore. Of the 36 stores with online ordering, one says explicitly that it accepts SNAP online: Forty Acres Fresh Market in Chicago, which takes SNAP for its online produce boxes and states a $5 delivery fee for those orders. Every Instacart marketplace page we opened for these stores said the store “does not currently accept EBT SNAP,” including two stores whose own websites say they accept SNAP, a reminder that accepting it at the register and accepting it online are separate permissions. Four of the six marketplace pages carried Instacart’s own label “Higher than in-store prices,” and one said “No markups.” Only 8 of the 85 sites mention SNAP or EBT anywhere, and 2 mention WIC.

On pricing, the national sample suggests that independent grocers are not the stores Maryland’s new law was written for. Sixteen of the 85 run a loyalty or rewards program, seven offer digital coupons, the only language about personalized offers we found was in one cookie policy, and none mentions electronic shelf labels. Thirty-three show prices online, 27 post a weekly ad, and 15 show their delivery or pickup fees. None of the 18 readable sites in the six sampled states where SNAP soda or candy restrictions were in force mentions the restriction, although every one of those stores has to enforce it at the register.

What machines can read is decided mostly by platforms. Twenty of the 85 sites have an llms.txt file, and every one was generated by the website builder, eight by Wix and seven by Shopify, not written by the store. Thirty-nine carry structured data of some kind, but only 11 say what kind of business they are, two declare themselves a GroceryStore, and two that Wix built call themselves restaurants. Three block at least one AI crawler, two of them because their e-commerce host’s default robots.txt blocks all twelve AI bots we tested and describes product catalogs and per-item pricing as commercially sensitive. A grocer that never chose any of these settings has made a decision about AI search anyway.

A few stores stand out. Kuhn’s Market in Pittsburgh puts its weekly ad, loyalty card with digital coupons and pickup and delivery on its home page. Viengchan Oriental Market in Brooklyn Park, Minnesota runs a Freshop store with a written online pricing policy and lists EBT and eWIC among its payment methods. Canseco’s Market in New Orleans and Seward Co-op in Minneapolis sell through Instacart storefronts under their own names, Georgetown Market in Indianapolis carries the “No markups” label on Instacart, and Baiz Market in Phoenix publishes GroceryStore markup with its location and hours.

Selling groceries online with SNAP

Federal regulators have said, in a final rule and its impact analysis, that they expect online SNAP purchasing to cover part of the access that stricter stocking standards may cost. In Baltimore, as the census above shows, that online access currently runs almost entirely through chains. An independent grocer that wants to change that has a better path than most owners realize, and a handful of rules to build around.

How a small store starts taking SNAP online

The process is set out on the Food and Nutrition Administration’s online requirements page and its onboarding page. A store must already be authorized for SNAP in person. It contacts the agency’s online team with its authorization number, and every store under the same ownership goes through reauthorization first, which after November 4 means every location will be checked against the 28-variety standard before any of them can sell online. The store’s e-commerce platform must process SNAP through one of the three approved processors, Fiserv, Worldpay or Forage; if the platform is not already on the agency’s list of 26 approved e-commerce platforms, which includes Shopify, Instacart, DoorDash, Uber Eats, Mercato, Local Express and ECRS’s Webcart, the retailer submits a business requirements document. Then come end-to-end tests with EBT test cards before the store goes live. The list carries the same warning as the equipment rules: SNAP retailers pay for any costs or fees involved.

What SNAP pays for online, and what it does not

The checkout rules are specific, and each one is a requirement for the software. SNAP can pay only for eligible food, and the agency’s online purchasing page says that “delivery fees and other associated charges may not be paid for with SNAP benefits,” so the bag fee, the delivery fee and any tip go on a second tender. The customer chooses how much of the order to pay with SNAP, and enters a PIN through the processor rather than on the store’s own page. Only one EBT card may be attached to a customer account. To stop thieves from testing stolen card numbers, the store must either remove the balance check or keep it and disable guest checkout, limit changes to the stored card and allow no more than two balance checks per session. A September 2, 2026 memo from the agency asks processors to encourage retailers to allow refunds only to a card that made a matching purchase. The older rules apply online as much as in the store: food bought with SNAP must be sold at the same prices and on the same terms as for cash, without sales tax, and a store must accept a valid card from any state. We found no federal rule that requires a store’s online prices to match its shelf prices, which is exactly the gap the delivery platforms’ own pickup terms and Maryland’s new law each fill in a different way.

Marketplace, storefront or your own store

There are three ways to sell groceries online, and many Baltimore independents will want more than one. A marketplace listing on Instacart, DoorDash or Uber Eats brings customers and drivers the store could never afford on its own, at the cost of a commission, fees the shopper pays, prices the store often raises to cover them, and a customer relationship that belongs to the platform. A white-label storefront, such as Instacart’s Storefront Pro or NCR Voyix’s Freshop, puts the store’s own name on a platform-run site, usually at a price that is quoted rather than published. A store-owned online store, on Shopify with Forage’s app or built to order, costs the least per order and keeps the customer list, but the store has to bring its own customers and run its own pickup or delivery.

The SNAP question cuts across all three. On Instacart’s Baltimore page, the chains take EBT and the independents we checked do not, and SNAP acceptance on a marketplace is a per-store setting that the store has to apply for and the platform has to support. On a store-owned site, it is a feature the store controls. For a store where a quarter or more of sales are paid with SNAP, an online store that cannot take it is selling to the minority of its own customers, which is one reason so few independents bother.

Maryland’s new pricing law adds a reason to watch the marketplaces closely. A delivery app is covered by it even when the store is not, so the app may not use a shopper’s personal data to charge that shopper more for groceries. Uniform markups, fees that vary with distance and prices the store itself sets higher on the app all remain lawful. A store that sells on a marketplace should be able to see, item by item, what its customers pay there compared with its own counter, both because DoorDash and Uber Eats already tie their cheaper pickup rates to proof of in-store pricing and because a customer who notices a difference will ask the store, not the app.

How shoppers, and their AI assistants, find a grocery

Google has been unusually direct about what it wants. Its guidance on AI features says there are “no additional requirements to appear in AI Overviews or AI Mode,” and its AI optimization guide, updated in July 2026, says an llms.txt file will “neither harm nor help” a site’s visibility in Google Search. What it does ask a store to keep current is its Business Profile and, if it has one, its Merchant Center data, and for a grocer that is where the work is. A local inventory feed tells Google, at least once a day, which items each store has and at what price, with the non-member price in the price field, a loyalty price in its own field, and stock levels translated into in stock (three or more), limited (one or two) or out of stock. Google’s free Local Inventory app, the former Pointy, builds that feed from the register automatically, but only for items with manufacturer barcodes sold through one of the point-of-sale systems it supports, a list that includes Clover, Square and Lightspeed and no grocery specialist, and it guesses stock from how often an item is scanned. Loose produce, meat and deli items, the things a neighborhood grocer is often best at, fall out of it unless someone enters them by hand.

One setting deserves an owner’s attention this week. Google’s Business Profile help says that in-store inventory from Instacart, DoorDash and Uber Eats “automatically shows on Google,” and an owner can unlink those sources. It does not say which price Google displays, so a store whose items are marked up on an app should check what its Google listing shows before a shopper does. The same applies to AI shopping assistants. Google’s new agent checkout and OpenAI’s both pay with stored cards, which a SNAP shopper cannot use, and ChatGPT’s in-chat checkout was scaled back in March 2026, with Instacart’s ChatGPT app now sending shoppers to Instacart to pay. For an independent grocer, an AI assistant will mostly learn about the store from the same places a search engine does: its own website and structured data, its Business Profile, its Merchant Center feed if it has one, and whatever the marketplaces publish on its behalf. The federal SNAP retailer locator, which shows physical stores and refreshes every two weeks, does not say who accepts SNAP online; that is a separate list of names.

The catalog is the store

Every one of these channels is fed by the same thing: the store’s item file. The eligibility flag decides what SNAP can buy online; the staple category and variety fields decide whether the store stays authorized; the WIC list status decides what rings up on an eWIC card; the tax flag decides the receipt and, now, the scope of Maryland’s pricing law; the price and its history decide what the marketplace shows and whether the store can explain it; and the same rows, published as structured data, decide what a search engine or an AI assistant can tell a shopper about the store. A grocer with a clean item file can open a new channel in weeks. A grocer without one pays to clean it every time a new channel, rule or terminal comes along, which in 2026 has been every month.

A note for our readers in Lithuania

Part of our team works from Lithuania, and the comparison is instructive because Europe answered Maryland’s question about personalized prices in the opposite way. Maryland restricts: a large grocer or a delivery app may not use your personal data to charge you more for food. The European Union discloses. Since May 28, 2022, Article 6(1)(ea) of the Consumer Rights Directive has required a trader to tell the consumer, before a distance contract, “that the price was personalised on the basis of automated decision-making,” and recital 45 of the directive that added it says the duty does not cover dynamic or real-time pricing that changes for everyone. Lithuania carried it into Article 6.228⁷ of the Civil Code, where it became point 22 of the list in a restatement that took effect on September 27, 2026, and the trader bears the burden of proving the information was given. The rule applies to online, phone and off-premises sales only, so a price on a shelf label in a Vilnius supermarket needs no such notice at all. It also seems to have changed little: in August 2023, more than a year after it took effect, LRT reported that the State Consumer Rights Protection Authority (VVTAT) had not received a single complaint or inquiry about personalized prices. The European Commission’s planned Digital Fairness Act, which the Commission has scheduled for the fourth quarter of 2026, may go further: its call for evidence names personalized pricing that targets consumers’ vulnerabilities and “starting from” prices under dynamic pricing. As of this writing it has not been proposed.

The pricing rule Lithuanian shops actually meet is a different one. Under Article 6.353(8) of the Civil Code, every announcement of a price reduction must state the previous price, defined as the lowest price the seller applied in the 30 days before the reduction, and Lithuania applied that rule from July 1, 2021, eleven months before the European deadline, with an exemption for perishable goods. VVTAT has read loyalty prices into it: in a 2023 decision fining the e-shop Varle €2,475, it said that when a loyalty program is open to all customers or used by most of them, the lowest price over 30 days includes the loyalty price, a view it says the European Commission confirmed. In 2024 it fined a wholesaler’s online store €100,000 for presenting crossed-out recommended retail prices as discounts. For software, the Lithuanian rule and the Maryland one point to the same object: a dated price history per product and per channel, with the reason for each change, kept long enough to answer a regulator.

Electronic shelf labels, which a Maryland bill proposed to ban for large grocers this year, are already standard in Lithuania. Lidl finished installing them in all of its Lithuanian stores in 2024 and said it would save about 5.2 million sheets of paper a year, and in April 2026 TV3 asked each major chain how it uses them: Rimi has them in all its Express stores, Iki says it installs them to reduce pricing errors, and Norfa changes prices only on instruction from management. In February 2024 a price-monitoring firm filmed a Lidl label changing a tea’s price in the middle of its survey, and Lidl answered that it adjusts prices to competitors, several times a day if needed. The market those labels serve is concentrated: in 2025 Maxima’s Lithuanian business had revenue of €2.33 billion and reports a market share of 29.7%, Norfa €1.19 billion, Lidl €956 million and Rimi €496 million. Online grocery is still a small slice. Barbora, Maxima’s online store, sold €152.9 million of goods across the Baltic states in 2025, 3.8% of Maxima’s Baltic retail turnover, while in the United States online orders reached 20.9% of grocery spending in the third quarter of 2026.

Taxes run the other way too. Lithuania charges its standard 21% VAT on most groceries; a proposal for a 5% rate on fresh fruit, vegetables and berries won a Seimas committee’s support in February 2026 over the finance minister’s objection, and we found no sign that it has been adopted. Maryland taxes neither grocery food nor anything bought with SNAP, and taxes candy and soft drinks at 6% when they are paid for any other way. For a Lithuanian reader, the lesson of this article is that American food retail is regulated through the item file, which foods a store must stock, which barcodes a benefit program recognizes, which items a register may tax, while European food retail is regulated through the price display: the unit price, the previous price and, online, the personalized-price notice. A grocer’s software that keeps a clean item file and a dated price ledger satisfies both.

What custom actually costs, and when it pays

We price every project at a fixed fee, agreed before we start, and publish the starting points on our pricing page. For a Baltimore grocery, a corner market or a small group of stores, the packages map onto this article like this.

Most of our projects land between $6,000 and $30,000, and none requires you to give up a register, a processor or a marketplace that already works; the demos page shows finished work. In Maryland, the state’s 3% tax on software and IT services applies to a custom build, so the Online Store package starts at $6,180 with tax and a Custom App at $12,360.

Whether custom pays depends on what it replaces. The online store pays on volume. In the model store above, moving $50,000 a year of orders from a marketplace at a 15% commission to the store’s own pickup site saves $7,500 in commission and costs about $1,725 in card and SNAP processing at the published Shopify-style and Forage rates, a net saving of about $5,775 a year, so the Online Store package with tax pays for itself in about 13 months before hosting, and faster at a higher commission. The stocking report pays in a less visible way. For the same store, with a quarter of its $2.5 million in sales paid with SNAP, a denial at reauthorization followed by the rule’s six-month wait puts about $312,500 of SNAP sales at risk, and for a corner store where SNAP is a larger share of a smaller business, the loss can decide whether the store survives. A $3,500 prototype that tells an owner, before the inspector does, that the dairy case counts five varieties and not seven is the cheapest insurance in this article.

What we would build for a Baltimore grocer

We would not build a point-of-sale system. Grocery registers have to be certified with payment processors, with the State’s EBT processor and with the WIC program before they can take a single benefit transaction, and the companies that sell them have spent years on scales, scanners, cash drawers and the odd tender rules of a grocery checkout. We would not build an EBT terminal, a payment gateway or a delivery fleet either. What we would build is the layer the registers, the regulators, the marketplaces and the store’s own website all depend on and that nobody sells to a store this size: a single, correct description of what the store sells, at what price, to whom and why.

The first piece is an item master that knows what the regulators know. Every product carries its barcode, its Maryland tax treatment (grocery food, food for immediate consumption, soft drink or candy, non-food), its SNAP eligibility, its SNAP staple category, variety key and perishable and accessory flags, whether its barcode or PLU is on the Maryland WIC approved product list, and, for produce and deli items, how it is sold by weight. Kept in one place, those fields produce the 28-84-9 stocking report for a SNAP store visit, the WIC minimum stock check, the tax lines on every receipt, the eligibility flags on the website and the feed that goes to a marketplace, all from the same row, so they cannot disagree with each other. The store’s purchase invoices are filed against the same items by date, which turns the 21-day rule into a search.

The second is a price ledger. Each store and each channel, the counter, the website and every marketplace, has one dated price list, and every change is a line with the old price, the new price, who or what made the change and a reason code. For a grocer above 15,000 square feet, that ledger is what an answer to the Consumer Protection Division looks like. For a corner store, its most useful report is the comparison nobody runs today: what a customer pays for the same item at the counter and on each delivery app, so that the store knows how much of the markup is its own and how much is the platform’s, and can decide what to do about it.

The third is an online store that takes EBT. It sells from the item master, shows SNAP eligibility on each product, lets a family pay for eligible items with SNAP through a USDA-approved payment processor and for everything else, including any delivery fee, with a card, offers pickup windows the store can actually staff, asks before substituting, charges the bag fee where the jurisdiction requires it, and publishes Product and Offer markup and a local inventory feed so that search engines and AI shopping assistants can see what is on the shelf. It is the store’s own front door, with its own customer list, and it can sit next to a marketplace listing rather than replace it.

The fourth is the binder a store owner keeps forgetting to keep: licenses and permits with their renewal dates, the SNAP authorization and its reauthorization date, the WIC agreement, training and peer-group price notices, EBT terminal records, the bag-fee remittances, and the stocking report printed on the morning of every unannounced visit. None of it replaces the register, the processor or the marketplace. It connects them through the store’s own data, built on its own database, exportable in full and the store’s to keep.

Build, buy, or leave it alone

Most grocers should buy almost everything and build almost nothing. The register, the scale integration, card processing, EBT and eWIC acceptance, and the terminals that read the new chip cards should come from vendors who are certified to do it and answer the phone on a Saturday night. A marketplace listing should stay as long as it brings customers the store would not otherwise reach, and for many corner stores in Baltimore it does. A store that sells mostly snacks, drinks and household goods and has no interest in SNAP online or WIC has no reason to pay for anything we describe here, and we would say so on a call.

Choose vendors by the meter rather than the logo. Grocery point-of-sale systems are priced by the register or the location, and the ones built for independents are usually quoted rather than published; payment processing is priced per transaction, and EBT often through a separate processor; online ordering platforms charge a monthly fee, a share of each order or both; marketplaces take a commission from the store and fees from the shopper, and some allow the store or the platform to price items above the shelf; electronic shelf labels are priced per tag plus the system that drives them. Ask for the full price in writing, including what renews automatically and what the price becomes in year two, and then ask every vendor these questions before you sign:

• Can my register take Maryland’s chip and contactless EBT cards and eWIC in the same transaction as everything else, and who certified it?
• Can I store SNAP staple category, variety and perishable flags on each item and print the report a federal store visit checks?
• Does my online store accept SNAP online through a USDA-approved processor, and how are delivery fees and non-eligible items split onto another card?
• Who sets the prices my customers see on your platform, how much above my shelf price can they be, and can I export them?
• Do you, or any model you run for me, use a shopper’s personal data to set a price, and can you show me every price change with its reason?
• Who owns my item file, my customer list and my sales history, and can I export all of it, in full, whenever I like, including after I leave?

A vendor that answers all six well has earned the contract. Build when the answers are “by hand” for the questions that carry your risk: when the stocking report is a clipboard on the day of the visit, when the website, the register and the marketplaces show three different prices and nobody knows which is right, when SNAP shoppers cannot pay online, when WIC items fail to scan and nobody maps them, or when the only list of your customers belongs to an app. Leave alone what is cheap and works: the register, the processor, the scale and the marketplace listing that still brings in orders.

Our guide to custom software in Baltimore explains how we work, our article on butcher shops, bakeries and specialty food stores covers selling by weight and Maryland’s food tax line, our article on online ordering for Baltimore restaurants covers the delivery-app economics from the restaurant side, and our article on Maryland’s Online Data Privacy Act covers the law that sits under the new pricing statute. If you want to find out which of these would pay for you, bring us your item file, last month’s marketplace statement and the date of your next SNAP reauthorization, and we will show you.

Frequently asked questions

What are the new SNAP stocking requirements for retailers?

Under a USDA final rule published May 8, 2026 (91 FR 25082), a store authorized under Criterion A must offer at least seven distinct varieties in each of four staple food categories (protein, grains, vegetables or fruits, and dairy), 28 in all; at least three stocking units of each variety, 84 in all; and one perishable variety in at least three of the categories, so that 3 of the 28 varieties and 9 of the 84 units are perishable. The items must be displayed for sale on any day the store is open. Accessory foods such as chips, candy, soda, condiments, butter and jerky do not count. A store can qualify under Criterion B instead by showing that more than half of its total gross retail sales are staple foods.

When do the new SNAP stocking standards take effect?

The rule took effect on July 7, 2026, and retailers must implement it by November 4, 2026. From November 4, new stores and new locations are assessed under it when they apply; existing stores are assessed at their next reauthorization, which generally comes every five years, although the agency may review any store at any time. A store denied for failing the stocking standard must wait six months before applying again. If a store falls short during a store visit, it can show invoices or receipts proving it ordered or received the missing stock no more than 21 calendar days before the visit.

Does chocolate milk count as a separate variety under the new SNAP rule?

Yes. Under the new rule, flavored perishable milk is a distinct variety from plain perishable milk, shredded cheese is distinct from other cheese, and whole grain bread and pasta are distinct from other bread and pasta. Different fat contents, package sizes, brands and flavors of the same product still count as one variety, and butter is now an accessory food that does not count at all. The agency’s “Is My Store Eligible?” page still says chocolate milk and regular milk are one variety, which described the old rule.

Did Maryland ban surveillance pricing?

Partly. The Protection From Predatory Pricing Act, Chapter 154 of 2026, codified as Commercial Law §13-321 and in effect since October 1, 2026, bars food retailers with at least 15,000 square feet and third-party delivery service providers from using dynamic pricing or personal data to set a higher price for a single consumer on food that is exempt from Maryland sales tax under Tax-General §11-206(c). It lists nine exemptions, including promotions, voluntary loyalty and membership programs, subscriptions, cost and location differences, consensual data-for-price offers, error corrections and outage resets. It does not ban electronic shelf labels or prices that change for everyone, does not require an algorithm disclosure, and gives no private right to sue; the Attorney General’s Consumer Protection Division enforces it after a notice and a 45-day period to cure.

Does Maryland’s pricing law apply to corner stores and small grocery stores?

Not to the store itself if it is smaller than 15,000 square feet, because the law defines a food retailer as a merchant that operates a business establishment of at least that size. It does apply to the delivery apps that sell a small store’s groceries, because the definition of a third-party delivery service provider has no size threshold. The law does not say how floor area is measured, and because it defines the merchant rather than the store, a grocer with one large store and several small ones should assume, on our reading, that the ban covers all of its prices.

Does Maryland’s pricing law apply to Instacart, DoorDash and Uber Eats?

On our reading, yes. A third-party delivery service provider is a merchant that facilitates as a consumer service the delivery of tax-exempt food, which describes the grocery delivery apps, and such a provider may not use a shopper’s personal data to set a higher price on that food for that shopper. Uniform markups, delivery fees that vary with distance and prices a store sets higher on an app for everyone are not prohibited. Instacart ended all item price tests on its platform in December 2025.

Did Maryland ban electronic shelf labels?

No. A 2026 bill, SB 889, would have required food retailers of 15,000 square feet or more to use non-digital price tags, but it received a hearing and no vote. Maryland’s pricing law does not mention electronic shelf labels; it prohibits using personal data to raise a price for a single shopper, so a label that changes a price for everyone remains lawful. New Jersey’s new law freezes the installation of new electronic shelf labels for a year starting February 1, 2027.

Do Maryland’s new chip EBT cards still work with a swipe terminal?

Yes. Maryland’s chip-and-tap EBT cards, mailed from July 2026, still carry a magnetic stripe, and federal guidance tells retailers to insert or tap first, swipe if the terminal asks, and key in the card number only as a last resort. The State said the old stripe-only cards would be deactivated on September 30, 2026; on October 1 it reported that about 85% of nearly 507,000 new cards had been activated and that the old cards “will deactivate soon.” Every EBT transaction still requires an online PIN, there are no offline approvals, and retailers pay for their own EBT equipment.

How can a small grocery store accept SNAP EBT online?

It must already be authorized for SNAP in the store. It contacts the Food and Nutrition Administration’s online team with its authorization number, every store under the same ownership goes through reauthorization first, and its e-commerce platform must process SNAP through one of the three approved processors, Fiserv, Worldpay or Forage, before end-to-end testing with EBT test cards. Forage publishes a price of 3.5% plus 30 cents per transaction. SNAP can pay only for eligible food, so delivery fees, bag fees and tips need another form of payment, and the shopper enters a PIN through the processor.

Does Square, Shopify or Clover accept EBT?

Square added SNAP EBT at checkout as an open beta on July 30, 2026, on its Register, Terminal and Handheld in retail mode, for sellers with an FNS number, and lists EBT at 1.8% plus 5 cents per transaction. Clover supports EBT on several of its devices, including chip and contactless cards where the state supports them, but not for online or offline transactions. Shopify’s point of sale does not advertise EBT; Shopify stores can accept SNAP online through Forage’s app. Many older answers online, including a 2022 Square community thread, are out of date.

Can you use WIC online in Maryland?

Not yet. The federal rule that would allow online WIC ordering was proposed in February 2023 and has not been finalized, although a final rule had been targeted for July 2026, and we found no Maryland pilot. In stores, Maryland WIC purchases are made with the eWIC card, and an item rings up only if its barcode or PLU code is on the State’s Approved Product List.

Is sales tax charged on SNAP purchases in Maryland?

No. Food bought with SNAP is exempt from Maryland’s 6% sales tax under Tax-General §11-206(b), and SNAP rules require that purchases be made without sales tax. That includes soft drinks and candy, which Maryland taxes when they are paid for with cash or a card. Grocery food sold for consumption off the premises by a store with a substantial grocery business is exempt for every customer, while hot food, sandwiches and other food for immediate consumption are taxable.

Does Baltimore’s bag fee apply to SNAP and WIC purchases?

Yes, in both Baltimore City and Baltimore County. The City charges five cents for each bag provided at the register, at pickup or with a delivery, itemized on the receipt; the store keeps four cents and remits one cent each month. The County requires at least five cents for each paper or reusable bag, kept by the store, but exempts independent retailers with three or fewer locations, other than convenience stores and gas stations with convenience stores. Neither exempts SNAP or WIC shoppers, and a bag fee cannot be paid with SNAP, so those customers need another form of payment for it. Montgomery County does exempt them.

Which states ban buying soda with SNAP, and does it apply in Baltimore?

As of October 2026, 18 states had approved SNAP food restriction waivers, with start dates running into 2028, and Virginia’s restriction on sweetened carbonated drinks begins November 1, 2026. Maryland, Delaware, Pennsylvania and the District of Columbia have none, and in June 2026 a federal court vacated the approvals for Colorado, Iowa, Nebraska, Tennessee and West Virginia. A walk-in store follows the rules of the state it is in, so a Baltimore store restricts nothing for any state’s card, while an online order filled from a warehouse follows the rules of the state that issued the card.

How many grocery and corner stores accept SNAP in Baltimore?

USDA retailer data listed 584 SNAP-authorized retailers in Baltimore City and 498 in Baltimore County as of October 7, 2026. In the City, 371 were convenience stores, 65 grocery stores, 23 supermarkets and 29 super stores, so convenience stores outnumber supermarkets and super stores by about seven to one, against under two to one in the County. The City’s count of SNAP retailers fell from 1,000 at the end of 2015 to 609 at the end of 2025.

How much does grocery store software cost?

Published register prices run from free to about $290 a month per location for general retail systems such as Square, Shopify POS and Lightspeed, and from about $20 to $110 a month for systems sold to small grocery and convenience stores, while supermarket systems such as ECRS’s Catapult are quoted privately; card processing usually costs far more than the software. EBT processing ranges from a flat $49.95 a month to 1.8% plus 5 cents per transaction. Online, Forage charges 3.5% plus 30 cents per SNAP transaction and marketplaces take a commission. Our fixed prices start at $3,500 for a one-week prototype, $6,000 for an online store and $12,000 for a custom app or an operations system, plus Maryland’s 3% tax on software services.