Custom specialty food retail software in Baltimore: two numbers you cannot know at the moment of sale

There is a question we have taken to asking early when we sit down with a food shop, and the answer is almost never a number. It is this: how much did you sell to restaurants last year? Not roughly, not as a share of the business — the dollar figure, meat separately from poultry, wholesale separately from retail.

Nobody has it. What people have is a feeling. A fair bit. More than we used to. The Italian place takes a lot. Then somebody offers to pull the invoices, and an hour later there is a spreadsheet with a number in it that is probably right within twenty percent, and everyone agrees that they should really keep better track of that.

They should, and not for the reason they think. That figure is not a management metric. In a shop that cuts, grinds, cures or smokes anything, it is the number that decides whether you are legally a retail store or an uninspected federal establishment operating without a grant of inspection, and there is a specific dollar amount attached to it that changed on the 28th of May this year.

This is the twentieth trade we have taken apart in this series, and it is the first where the business is selling something that does not have a price at the moment the customer decides to buy it. That single fact — the price arrives after the transaction begins, not before — turns out to explain most of what is wrong with the software, and it rhymes with the legal problem in a way we did not expect when we started. Both are numbers that only exist once the thing has already happened. Let us start with the market, because Baltimore's is stranger than its size suggests.

What Baltimore's specialty food market actually looks like

We counted rather than quoting a trade association. The Census Bureau's County Business Patterns program publishes establishment counts by county and industry code, and specialty food retail sprawls across seven of them: 445210 for meat markets, 445220 for fish and seafood markets, 445230 for fruit and vegetable markets, 445291 for baked goods stores, 445292 for confectionery and nut stores, 445299 for other specialty food stores, and 311811 for retail bakeries, which is where a shop that bakes what it sells is counted rather than under retail at all.

Taken together the 2023 file gives Maryland 520 specialty food establishments with paid employees, 3,826 employees and about $101.1 million in annual payroll. That is a small trade by the standards of this series — roughly a fifth the size of Maryland's salon and barbering sector, and spread very thin across twenty-four jurisdictions. It is also, in one particular corner, a trade where Baltimore City leads Maryland outright.

Baltimore City sits fourth on count and pays $32,883 per employee against Montgomery's $24,313, Anne Arundel's $25,025 and Prince George's $26,071 — the best pay in the metro, and an inversion of the pattern we found in salons, funeral homes and flower shops, where the city consistently pays worse than its ring. Worcester County sits above it on that column, and the reason is charming rather than economic: it holds 15 confectionery and nut stores, more than any other jurisdiction in Maryland, because that is what an Ocean City boardwalk is made of, and 28 seasonal shops carrying only 77 employees between them make payroll per head look generous in a way that would not survive a second glance.

Break the city's 64 establishments apart and the interesting one appears immediately.

Baltimore has more butcher shops than anywhere else in Maryland

Under NAICS 445210, meat markets, Baltimore City has 13 establishments with paid employees — the largest count of any jurisdiction in the state. Montgomery has 11, Prince George's 7, Baltimore County 6, Washington 5, Anne Arundel 4 and Howard 3, out of 49 statewide. The city has the most butcher shops in Maryland and, at $50,625, the highest payroll per employee of the four counties with the most of them — more than double Montgomery's $23,603 — though with only 48 employees across those 13 shops that particular average rests on a very small cell and we would not lean on it hard.

The establishment count itself, though, is solid, and it is not what most people would guess about a city with Baltimore's grocery reputation. It is also a portrait of a very small business: 3.7 employees per shop, the smallest average of any trade we have measured in this series — smaller than florists, smaller than funeral homes, smaller than one-doctor veterinary practices. The city adds 21 retail bakeries and baked goods stores, 10 other specialty food stores, 9 confectionery and nut stores, 7 fish and seafood markets and 4 fruit and vegetable markets.

Some of that is the public market system, which is worth naming because it is genuinely unusual. Baltimore runs the oldest public market system in the United States. Lexington Market was founded in 1782 and moved into a new 60,000 square foot shed in 2022 with about fifty merchants; Broadway Market in Fell's Point dates to 1786; Cross Street Market in Federal Hill has carried a butcher counter through several rebuilds. A stall in a public market is a specialty food business with all of the regulatory obligations of a standalone shop and none of the square footage, which is exactly the size at which buying software that assumes a back office starts to feel absurd.

The grocery number that explains the rest

Now the contrast, and it is the most striking single pair of figures we have pulled out of the CBP file in twenty posts. Under NAICS 445110, supermarkets and other grocery stores, Baltimore City has 231 establishments with paid employees — more than any other jurisdiction in Maryland, ahead of Montgomery's 212, Prince George's 191 and Baltimore County's 174. On employment it has 3,897 against Montgomery's 10,811.

Baltimore City leads Maryland in the number of grocery stores and has the smallest average grocery store in the region, at roughly a third the staffing of Montgomery County's. Those two facts are the same fact, and it is the statistical fingerprint of a corner-store food system.

We are not going to pretend a piece about point of sale software is going to fix that. But it does change what the software has to be. A trade made of 16-employee grocers, 3.7-employee butcher shops and market stalls is not a trade that is going to run an enterprise fresh-item management suite, and any recommendation that assumes otherwise is a recommendation for somebody else's city.

The number the scale decides

Here is the problem that separates this trade from every other retailer we have written about. A hardware shop sells a hammer for $24.99. A pharmacy sells a box with a price on it. A florist sells an arrangement priced by design. A butcher sells a ribeye, and a ribeye costs $18.99 a pound, and nobody — not the shop, not the customer, not the till — knows what this particular ribeye costs until it has been cut and put on a scale.

In the shop this is a solved problem and has been for a century. The scale is a legal instrument, inspected by the state, and it prints a label with a weight, a unit price and a total. The till takes the total. Everyone is happy.

Online it is not solved, and the reason is structural rather than lazy. A shopping cart is a promise: this is the price, click here and we will charge you exactly that. A variable weight item cannot make that promise, because the fulfillment step is what determines the amount. Shopify's own staff have said in its community forums that there is no native support for selling by measurement or weight in either the admin or its point of sale. The ecosystem answer is third-party apps — Filljoy, IzyUnit, Unitpricer and others, several of which name butcher shops in their own descriptions — and the standard workarounds are to add an overage of about ten percent and refund the difference after fulfillment, or to authorize the card and capture the true amount later.

Both of those work. Both of them also mean that for a period of days your order record, your payment record and your accounting record disagree, and somebody reconciles them by hand. On a shop doing forty online orders a week that is a chore. On a shop doing four hundred it is a job.

The platforms built for this trade solve it properly, and it is instructive to read how one of them describes the mechanism:

"Your Farm can sell anything online by taking an initial Deposit, then entering in the final Weight when closing the order. With Barn2Door's patented technologies, the final charge automatically deducts the initial deposit." — Barn2Door, Sell-by-Weight

That is the correct architecture: a deposit at checkout, a weight at fulfillment, a second charge for the balance. It is also, plainly, a different piece of software from a shopping cart — two payment events, an order that stays open, a customer who has agreed to a price per pound rather than a price. And it is worth noticing that a company felt the mechanism was novel enough to patent, in 2026, for a problem butchers have had since scales were invented.

The same shape recurs everywhere in this trade once you look for it. A whole or half animal share is a deposit now and a settlement in eleven weeks, against a cut sheet the customer filled in. A standing bread order is a subscription whose quantity changes weekly. A cheese counter sells by the piece, by the weight and by the board. A crab order in July is priced by the bushel at a number that was set that morning. None of these are exotic. All of them break a stock catalogue whose central assumption is that a product has a price.

What the software actually costs

We checked twenty-nine products on 5 August 2026 across the four families a specialty food shop actually shops in: general and specialty retail point of sale, grocery-specific systems, bakery production software, and the farm-direct commerce platforms that have quietly become the default answer for anyone selling meat online. Nine publish a subscription price outright, which makes this a more transparent market than most we have surveyed — pharmacy managed none out of thirty, hotels one out of sixteen.

Two things in that table are worth more than the prices. The first is GrazeCart's, which prints a dollar sign followed by an ellipsis where the number should be — "$…" — under a heading inviting you to talk with an expert. We have now seen a pricing page whose only figures were gift cards, a page titled "Affordable & Transparent Pricing" with no figures at all, and a directory listing. A literal typographic ellipsis is a new one, and it is at least honest about what it is doing.

The second is Barn2Door's own explanation of how it sets tiers, which is the clearest statement of the meter-is-the-product model we have found in any trade:

"We're not the cheap, generic DIY commerce solution. Pricing tiers correspond directly to your Farm size, order volume and customer base to align with the value realized by your business." — Barn2Door, pricing page

To be fair to them, that is exactly what it says on the tin, the rate card is complete, and a fixed 2.9 percent plus 30 cents on every tier — with no "as low as" and no surcharging product bolted on the side — is more straightforward than most of what we found in salons or hotels. We would rather a vendor say plainly that the price is a function of how well you are doing than bury it in a call.

Add it up for a real shop. A Baltimore market running Markt POS at $49 and a Barn2Door Business storefront at $159 is paying $2,995 in the first year and $2,496 a year after that, plus Maryland's 3 percent tax on information technology services. On a shop doing $1.2 million that is two-tenths of one percent of sales.

Hold that number, because everything that follows is bigger than it.

Where the money actually is

Model a Baltimore shop at the larger end of the trade: $1.2 million in annual sales, eight people, a full service counter, a small wholesale side and a website. Three numbers matter and the subscription is not one of them.

Shrink comes first, and it is not close. The Food Marketing Institute's fresh-department figures put shrink at 5.7 percent in meat, 7.6 percent in seafood, 8.5 percent in bakery and 8.7 percent in deli — the highest rates anywhere in a food store, and the reason grocery chains staff whole analytics teams against them. At 5.7 percent on $1.2 million that is $68,400 a year walking out in the bin, in the trim, in the reduced-to-clear tray and in the yield you did not get off the primal. One percentage point is $12,000, which is roughly five times the entire annual software bill.

Card processing comes second. At seventy percent card volume and a $32 average ticket, $840,000 across about 26,250 transactions at 2.6 percent plus ten cents is $24,465 a year, an effective 2.91 percent once the per-transaction cents are folded in. That is ten times the software bill, and — as in every trade we have costed — the published rates have converged so tightly that shopping processors moves a few hundred dollars, not tens of thousands.

Third, and last, the software. $2,496 a year. It is the number everybody negotiates and the only one on the list that cannot move enough to matter.

On a $1.2 million specialty food shop, total shrink runs about twenty-seven times the annual software bill and card processing about ten times it. The single line every owner shops hardest is the one with the least leverage in the building.

The reason shrink stays invisible is the same reason it stays invisible in every trade in this series: nothing joins the two halves. The invoice from the wholesaler is in pounds of subprimal. The till is in dollars of retail cut. There is no shared key between them, so nobody can answer the only question that matters — we bought 480 pounds of chuck this week; how many dollars of chuck did we actually sell? — without somebody sitting down with a calculator. The systems that could answer it are the ones that already hold both sides, and neither of them is asked to.

That is a small piece of software. It is also, at $12,000 a point, the highest-return small piece of software in the shop.

The ceiling nobody sells you a way to see

Now the second unknowable number, and this is the part of the article we would ask you to read even if you skip everything else.

If your shop cuts, grinds, cures, cooks or smokes meat or poultry, you are doing work that the Federal Meat Inspection Act says requires federal inspection — unless you fit inside an exemption. The exemption is at 9 CFR §303.1(d), and it is written in a way that will feel unusually generous until you get to the conditions:

"The requirements of the Act and the regulations in this subchapter for inspection of the preparation of products do not apply to operations of types traditionally and usually conducted at retail stores and restaurants, when conducted at any retail store or restaurant or similar retail-type establishment for sale in normal retail quantities or service of such articles to consumers at such establishments." — 9 CFR §303.1(d)(1)

Section 303.1(d)(2)(i) then lists the five operations that count as traditional and usual: cutting, slicing and trimming carcasses and wholesale cuts into retail cuts and freezing them; grinding and freezing; curing, cooking, smoking, rendering or refining of livestock fat, or other preparation of products, except slaughtering or the retort processing of canned products; breaking bulk shipments; and wrapping or rewrapping. That is a whole butcher shop, including the smoker. Note also §303.1(d)(2)(v), which sweeps in "any delicatessen which meets the requirements for a retail store" — so a deli counter is inside this regime whether or not anybody there thinks of themselves as a butcher.

The conditions are at §303.1(d)(2)(iii), and there are three that bind.

Condition one: a quantity limit measured in animals

No single sale may exceed a normal retail quantity, which §303.1(d)(2)(ii) defines as not more than one-half carcass in the aggregate, and then — because a regulation has to be operable behind a counter — gives you the table.

Which means that a shop selling quarter and half beef shares online is comfortably inside the exemption — a half is 300 pounds and that is the line — and a shop that sells somebody a whole steer is outside it. That is a rule about the contents of a shopping cart, expressed in pounds of animal, written in 1970s regulatory prose, and there is not a general-purpose e-commerce platform on earth that can express it.

Condition two: where the product came from

Section 303.1(d)(2)(iii)(c) requires that only federally or State inspected and passed product is handled or used in preparation. In most states that phrase gives you two doors. In Maryland it gives you one, and this is a fact about our state that surprises nearly everyone we mention it to.

Maryland does not run a state meat inspection program. Twenty-seven states operate an "at least equal to" state meat and poultry inspection program in cooperation with FSIS, covering about 1,900 establishments — all of them small or very small — with roughly $50 million a year in federal support. Maryland is not among them. Commercial meat processing here is federal jurisdiction, full stop.

That removes the middle rung. In Virginia or Ohio a shop that outgrows the retail exemption can graduate to state inspection, which is a real regulatory step but a manageable one. In Maryland there is nothing between the retail exemption and a full federal grant of inspection with an inspector on the premises.

It also closes off something the rest of the country spent this summer opening up. The Cooperative Interstate Shipment program lets a state-inspected plant operate as a federally inspected one under specific conditions and ship across state lines — but it is limited to plants located in states that run their own inspection program. FSIS has CIS agreements with Indiana, Iowa, Maine, Missouri, Montana, North Dakota, Ohio, South Dakota, Vermont and Wisconsin, and added Georgia on 27 July 2026, nine days before this piece was written.

Every few months another state's small processors get a new door into interstate commerce. Maryland's cannot walk through any of them, because the program starts on a rung this state removed.

We are not arguing Maryland should stand up an inspection program; that is a legislative question with a real budget attached and it is well outside our lane. We are saying that if you are a Baltimore shop planning to ship your own product, the ladder you have been reading about in the trade press is not one you can climb, and you should plan around the exemption rather than around a graduation that is not on offer.

Condition three: the dollar cap, and the arithmetic nobody has done

Here is the one that decides the year. Section 303.1(d)(2)(iii)(b) sets two tests at once:

"At least 75 percent, in terms of dollar value, of total sales of product represents sales to household consumers and the total dollar value of sales of product to consumers other than household consumers does not exceed the dollar limitation per calendar year set by the Administrator." — 9 CFR §303.1(d)(2)(iii)(b)

The same paragraph explains that the limitation adjusts automatically in the first quarter of each year whenever the Consumer Price Index moves the price of the same volume of product by more than $500, and that notice of the adjusted figure is published in the Federal Register. What it does not do is tell you the number. The regulation's footnote instead directs you to obtain the current limitation by telephoning a USDA office — at an area code and exchange that has not been USDA's for decades.

So here is the number, from the primary source. 91 FR 22789, published 28 April 2026 and applicable from 28 May 2026, raised the limitation to $109,600 for meat and meat food products and $76,100 for poultry and poultry products, on the basis of CPI series CUUR0000SEFF and CUUR0000SEFG accessed on 13 January 2026 — increases of $6,000 and $1,300 respectively, driven by a 5.86 percent annual average rise in meat prices against 1.73 percent in poultry.

Two caps, tracked separately, moving at different speeds, resetting each spring. Now do the arithmetic, because the interaction between the percentage test and the dollar test is where it gets genuinely counterintuitive.

The crossover is at $438,400 of annual product sales. Below it the percentage test binds and the shop has a proportional allowance that grows as it grows. Above it the fixed dollar figure binds and never moves again, which produces the result in the last column: the larger your shop gets, the smaller the share of it you are permitted to sell wholesale. A $2 million market may sell 5.5 percent of its product to restaurants. A $300,000 shop may sell 25 percent. Success narrows the door.

And in the units a shop actually thinks in, $109,600 is $2,108 a week. A restaurant taking $400 a week of your product uses 19 percent of your entire annual allowance; five such accounts and you are at the ceiling. One good steakhouse taking forty pounds of dry-aged ribeye a week at $18 is $37,440 a year — 34 percent of the cap on its own. Three of those and the exemption is gone.

The federal government does not cap how much a Baltimore butcher may sell. It caps how many restaurants a Baltimore butcher may sell to — at roughly five — and it publishes the exact figure once a year in a place no point of sale system reads.

The asymmetry inside the exemption

One more clause, and it is the sharpest thing in the regulation. Section 303.1(d)(2)(iii)(e) says preparation for sale to household consumers is limited to the operations in §303.1(d)(2)(i) — all five of them, curing and smoking included. The very next subparagraph says preparation for sale to anyone else is limited to §303.1(d)(2)(i) paragraphs (a), (b), (d) and (e).

Paragraph (c) is the one left out. Curing. Cooking. Smoking. Rendering.

Read plainly, the same batch of house bacon is an ordinary exempt retail product when it goes across the counter to a neighbor, and falls outside the exemption when it goes out of the back door to the restaurant on the corner. Same smoker, same cure, same morning. The thing that decides it is a single attribute of the buyer — household or not — which is a field that exists in no till in this trade and in very few e-commerce platforms either.

We are describing what the regulation says, not giving legal advice, and any shop with a wholesale line should put this in front of FSIS or counsel rather than in front of a blog post. But it is worth understanding the shape of it before you build anything, because the shape has a direct consequence: household-or-not is not a report you run at year end. It is a required attribute of every single sale, and it has to be captured at the moment of sale or it cannot be reconstructed at all.

There is a hint in the regulation itself as to what "reconstructed" looks like when it goes wrong. Section 303.1(d)(3) provides that where the Administrator has reason to believe a store has operated outside the conditions, they may order the operator to keep complete, accurate, and legible records of total monthly purchases and of total monthly sales, in dollar values, "separately show[ing] total sales to household consumers and total sales to other consumers." That is a description of exactly the ledger we build for shops in this trade, offered by the federal government as a remedy after something has gone wrong. It is considerably cheaper to have it beforehand.

Maryland's tax boundary runs through your own counter

The federal layer decides what you may make and who you may sell it to. The state layer decides what you charge tax on, and Maryland draws its line in a place no tax engine looks.

Start with the exemption. Tax-General §11-206(c)(1) 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 where the food is sold. §11-206(a)(6) defines that term:

"'Substantial grocery or market business' means a business at which at least 10% of all sales of food are sales of grocery or market food items, not including food normally consumed on the premises even though it is packaged to carry out." — Md. Tax-General §11-206(a)(6)

Ten percent is a low bar and essentially every butcher, baker, cheesemonger and market clears it. So your meat is exempt, your bread is exempt, your cheese is exempt. Good. Now read §11-206(a)(4), which defines the category that is not exempt — food for immediate consumption — as food from a salad, soup or dessert bar; party platters; heated food; sandwiches suitable for immediate consumption; and ice cream and other frozen desserts sold in containers of less than one pint.

Set that against the Comptroller's own enumerated item lists in the List of Tangible Personal Property and Services Subject to Sales and Use Tax, and the boundary stops being a category and becomes a series of shelf-level distinctions that no product-level tax field can carry.

That crab subsection deserves a moment, because it is very Maryland and it is also a genuine trap. §11-206(f) exempts, for consumption off the premises, "(1) crabs; or (2) seafood that is not prepared for immediate consumption." The immediate-consumption qualifier is attached to clause (2) and not to clause (1). Read as written, steamed crabs sold to go are exempt and steamed shrimp sold to go are not, and the two are frequently sitting in the same case, sold by the same person, on the same afternoon.

The taxability of a pound of roast beef in Maryland is not decided by what it is. It is decided by what shape it left the building in — and that is a property of the transaction, not of the product, which is exactly the field a stock tax engine does not have.

There is one more provision worth knowing if you bake. §11-210(c) exempts equipment used by a retail food vendor to manufacture or process bread or bakery goods for resale, but only where the taxable price of each piece of equipment is at least $2,000 and the vendor operates a substantial grocery or market business at the same location. A $2,100 mixer is exempt. A $1,900 mixer is taxable. The threshold is per item, so the way a quote is itemized changes the tax on it — and whether you qualify at all depends on a sales-mix test computed from your own till.

None of this is expressible in a product catalogue, and that is the practical point rather than a complaint about Maryland. A stock storefront gives every SKU a tax class. Maryland needs the tax decided at the line, from how the item was prepared and presented, which means the tax logic has to live in the order rather than in the catalogue. It is not difficult to build. It is simply not a thing you can configure.

What custom software costs

We price fixed, and we publish the numbers, which in a piece that has just spent two thousand words on vendors who do not feels like the least we can do.

One line to add to any budget because it is new and it catches people: Maryland now applies a 3 percent sales and use tax to information technology services, which reaches most of what you spend on software and on work like ours. It is small against the totals above, and it belongs in the model rather than on the first invoice as a surprise.

What we would actually build

Concretely, for a Baltimore food shop of three to fifteen people, in the order we would do it.

First, the storefront, because it is the one that makes money rather than saving it. Variable weight done properly — a deposit at checkout against a published price per pound, the real weight entered at the counter when the order is packed, the balance charged automatically, and one receipt that shows the customer the same arithmetic your scale printed. Alongside it, the things your case can do that a catalogue cannot: quarter and half shares with a cut sheet the customer fills in and a settlement date eleven weeks out, standing weekly bread and produce orders, holiday pre-orders with a hard cutoff tied to your production capacity rather than a guess, and pickup windows that reflect when things actually come out of the oven or off the block. All of it on your own merchant account, at your own processing rate, with the customer record belonging to you.

Second, the exemption ledger, and we would build this even for a shop with only two restaurant accounts. Every sale carries a household-or-not flag captured at the point of sale rather than inferred afterwards. Meat and poultry run against their own separate ceilings, because the regulation treats them separately and the two figures move at different rates. The screen shows one thing: how much headroom is left this calendar year, in dollars and in weeks at the current run rate. When the 75 percent test and the dollar cap cross over — at $438,400 of product sales at today's figures — the ledger switches which one it is warning you against. And on the second of January it resets and starts again, because the cap is a calendar-year cap and nobody's accounting year knows that.

Third, the yield and shrink record, which is where the $12,000 a point lives. What came in, by weight and by cost, from the wholesaler's invoice. What went out, by weight and by price, from the till. The difference, per primal and per item, per week. That is the join nobody has, and building it is mostly a matter of agreeing a shared key between two files that were never meant to meet. It is unglamorous work and it is the highest-return thing on this list.

Fourth, Maryland's tax logic at the line rather than in the catalogue, so that the same tray of cookies is taxed one way loose and another way arranged, the crab and the shrimp behave differently in the same order, and the half pint and the pint are not the same product with two prices. This is a day of work if you do it while building the storefront and a fortnight if you do it afterwards.

Build or buy

Our rule for this trade is narrower than usual, so here it is plainly.

• Buy the till. Markt POS, IT Retail, Square and Rain handle scale integration, label printing, tender types and end-of-day well, at $49 to $99 a month. Rebuilding a point of sale is a bad use of anybody's money.
• Buy the bakery production system if you run a production bakery. Recipe costing, batch scaling, allergen declarations and distribution runs are solved, and FlexiBake publishes a real price for them.
• Build the storefront if you sell by weight. This is the one place where the specialist platforms charge a real premium and the generic ones genuinely cannot do the job — and it is also the only channel where you keep the customer record.
• Build the exemption ledger. It is calendar-year, two-ceilinged, buyer-classified and federal, which is why no national product will ever ship it, and it is the only item on this list where being wrong costs you the business rather than some money.
• Build the join. Your wholesaler's invoice knows pounds and your till knows dollars, and until something reconciles them nobody in the building can tell you what a point of yield is worth.

Notice what is not on that list. We are not suggesting anybody rebuild inventory, accounting, payroll or e-commerce hosting. The specialist platforms in this trade are, on the whole, built by people who understand it — Barn2Door and GrazeCart both publish features that only somebody who has stood behind a counter would think to build, and a shop that is happy on one of them and does not have a wholesale problem has very little reason to talk to us.

Where to start

If you take one thing from this piece, make it the count we opened with. Pull last year's sales and split them into two piles: sold to a household, and sold to anybody else — restaurants, caterers, institutions, other shops. Do it separately for meat and for poultry. Then compare the second pile against $109,600 and $76,100.

Most shops we have done this with land somewhere between comfortable and startled. A few discover they crossed a line eighteen months ago and nobody noticed, because nothing in the building was counting. Either way it takes an afternoon, it costs nothing, and it will tell you more about what software you need than any feature comparison ever will.

And while you are at it, open your own website on a phone and try to buy a pound and a half of something. If you cannot — if the only options are one pound or two, or if there is no way to order the thing your case is actually full of — then the gap between what you sell and what your website sells is the whole project, and it is smaller than you think.

Common questions

How much does POS and online store software cost a specialty food shop in 2026?

Less than most owners fear, and far less than the numbers that actually decide the year. We checked twenty-nine products across grocery and specialty retail point of sale, bakery production software and farm-direct commerce platforms on August 5, 2026, and nine publish a price you can read without speaking to anybody. Markt POS lists $49 a month. IT Retail lists $49 a month. Rain lists $99 a month. FlexiBake lists $145, $165 and $175. Barn2Door publishes the fullest card in the category: Entrepreneur at $119 a month, Business at $159 and Scale at $299 on its annual view, with one-time setup fees of $399, $499 and $599, a fixed processing rate of 2.9 percent plus 30 cents on every tier, point of sale at $59 per device, and add-ons at $39 a month for its marketing toolkit and $499 for a logo. GrazeCart lists Starter at $89 a month billed annually. Notch lists $65 and $300. Stocksmith, which has rebranded from Craftybase since we last looked, publishes a full ladder from $20 to $199 a month. Square publishes its whole rate card as usual. As a planning figure, a Baltimore shop running a specialty point of sale, a farm-commerce or e-commerce storefront and scale integration is realistically spending $2,500 to $6,000 a year, plus Maryland's 3 percent tax on information technology services. On a $1.2 million shop that is about two-tenths of one percent of sales, which is why we spend most of this article on the other numbers.

How much meat can a butcher shop sell to restaurants without USDA inspection?

In calendar year 2026, $109,600 of meat and meat food products and $76,100 of poultry and poultry products, and both figures are ceilings on sales to consumers other than household consumers. The exemption itself is at 9 CFR 303.1(d), which excuses operations traditionally and usually conducted at retail stores from federal inspection of the preparation of products. Section 303.1(d)(2)(iii)(b) sets two tests that must both hold: at least 75 percent of the dollar value of your total product sales must be to household consumers, and the total dollar value of sales to consumers other than household consumers must not exceed the dollar limitation set by the Administrator. That limitation is adjusted for inflation and published in the Federal Register. The 2026 adjustment appeared at 91 FR 22789 on April 28, 2026 and is applicable from May 28, 2026. The two tests interact in a way most shops have never worked out. Below about $438,400 of annual product sales the 75 percent rule binds first. Above it the fixed dollar cap binds, and because the cap is a dollar figure rather than a percentage, the larger your shop grows the smaller a share of your business your wholesale side is permitted to be. At $1.05 million of product sales the ceiling is 10.4 percent of sales; at $2 million it is 5.5 percent.

Can an exempt retail butcher shop sell its own smoked or cured products to restaurants?

On the face of the regulation, no — and this is the detail that catches good shops out. 9 CFR 303.1(d)(2)(i) lists the five operations a retail store may perform under the exemption: (a) cutting, slicing and trimming into retail cuts and freezing them, (b) grinding and freezing, (c) curing, cooking, smoking, rendering or refining of livestock fat or other preparation of products, (d) breaking bulk shipments, and (e) wrapping or rewrapping. Section 303.1(d)(2)(iii)(e) allows all five when you are preparing product for household consumers. Section 303.1(d)(2)(iii)(f) then limits preparation for sale to other than household consumers to paragraphs (a), (b), (d) and (e). Paragraph (c) is the one that is left out. Read plainly, that means the same batch of house bacon is an ordinary retail product when it goes over the counter to a neighbor and falls outside the exemption when it goes out of the back door to the restaurant next door. The distinguishing fact is not the product, the process or the day — it is a single attribute of the buyer, which is precisely the sort of thing a point of sale system is not asked to record. We are describing what the regulation says rather than giving legal advice, and it is worth confirming your own position with FSIS or with counsel before you build a wholesale line.

Does Maryland have its own state meat inspection program?

No, and that absence shapes the whole trade here. Twenty-seven states run an at-least-equal-to state meat and poultry inspection program under a cooperative agreement with FSIS, covering roughly 1,900 establishments, all of them small or very small, and supported by about $50 million a year of federal funding. Maryland is not one of them; commercial meat processing in this state falls under federal jurisdiction only. For a Baltimore shop, that removes the middle rung of the ladder. There is no Maryland-inspected tier to graduate into — the choice is the federal retail exemption or full federal inspection, with nothing between them. It also permanently closes off the Cooperative Interstate Shipment program, which lets a state-inspected plant ship across state lines under specific conditions and is limited to plants located in states that run their own inspection program. FSIS has CIS agreements with Indiana, Iowa, Maine, Missouri, Montana, North Dakota, Ohio, South Dakota, Vermont and Wisconsin, and added Georgia on July 27, 2026. A Maryland processor cannot join, because Maryland removed the rung the ladder starts on.

Is food taxable in Maryland at a butcher shop, bakery or specialty market?

Mostly not, but the boundary runs through the middle of your own counter rather than around the outside of your business. Tax-General Section 11-206(c)(1) exempts food sold for consumption off the premises by a food vendor who operates a substantial grocery or market business at the same location, and Section 11-206(a)(6) defines that as a business at which at least 10 percent of all sales of food are sales of grocery or market food items. Almost any butcher, baker or cheesemonger clears that threshold comfortably. What the exemption does not cover is set out at Section 11-206(a)(4), which defines food for immediate consumption as food from a salad, soup or dessert bar; party platters; heated food; sandwiches suitable for immediate consumption; and ice cream and other frozen desserts sold in containers of less than one pint. Read that against the Comptroller's own item list and the results are genuinely odd. Cookies are exempt and a cookie platter is taxable. Fruit is exempt and a fruit platter is taxable. Popped popcorn is exempt and caramel-coated popcorn is taxable. Nuts are exempt and honey-roasted nuts are taxable. Coffee is exempt and heated coffee is taxable. A pint of ice cream is exempt and a half pint is taxable. Section 11-206(f) then exempts crabs outright, in their own subsection, with no immediate-consumption qualifier attached — so a bushel of steamed crabs sold to go is treated differently from steamed shrimp sitting in the same case. None of this is expressible as a tax category on a product. It is a property of the transaction.

Why can't Shopify or Square sell products by weight online?

Because a normal e-commerce checkout is built on the assumption that the price is known before the customer pays, and a whole rib roast does not have a price until it has been cut and put on a scale. Shopify's own staff have confirmed in its community forums that there is no native support for selling by measurement or weight in the admin or in POS; the ecosystem answer is third-party apps, and the common workarounds are to charge an overage of ten percent or so and refund the difference after fulfillment, or to authorize a card and capture the final amount later. Both work. Both also mean your order record and your accounting record disagree with each other until somebody reconciles them by hand. The platforms built for this trade solve it directly: Barn2Door describes taking an initial deposit, entering the final weight when the order is closed, and automatically deducting the deposit from the final charge, and calls the mechanism patented. That is the correct shape. It is also a two-stage payment, which is a materially different piece of software from a shopping cart, and it explains why the ability to sell a ribeye online is a specialist feature in 2026 rather than a checkbox.

How many butcher shops, bakeries and specialty food stores are there in Baltimore?

We counted from the Census Bureau's 2023 County Business Patterns county file rather than repeating a trade figure. Baltimore City has 13 meat markets with paid employees, which is the largest count of any jurisdiction in Maryland — ahead of Montgomery's 11, Prince George's 7 and Baltimore County's 6, out of 49 statewide. It also has 21 retail bakeries and baked goods stores, 7 fish and seafood markets, 9 confectionery and nut stores, 10 other specialty food stores and 4 fruit and vegetable markets: 64 specialty food establishments in total, against Montgomery's 105 and Baltimore County's 68. The average city meat market employs 3.7 people, which makes this the smallest-premises trade we have surveyed in this series. The contrast worth sitting with is in groceries. Baltimore City has 231 establishments under NAICS 445110, more than any other Maryland jurisdiction including Montgomery's 212, but only 3,897 employees against Montgomery's 10,811 — 16.9 employees per store against 51.0. The city leads the state in the number of grocery stores and has the smallest average store in the region, which is the statistical shape of a corner-store food system. County Business Patterns counts only establishments with paid employees, so every family shop running without payroll is missing from all of these figures.

Is it worth building custom software for a butcher shop, bakery or specialty food store?

Not to replace your point of sale, and not to replace your accounting. Square, Markt POS, IT Retail, Rain and the bakery production systems solve real problems, and a shop that is happy with its till should keep it. The case for a build is narrow and specific. The first piece is the storefront that can actually sell what you actually cut: variable weight with a deposit and a final charge, whole and half animal shares with cut sheets, pickup windows tied to when the van leaves, and Maryland's tax boundary applied per transaction rather than per product, all on your own merchant account. The second is the ledger that keeps your exemption: every sale classified as household or non-household, running against the two federal ceilings, showing headroom in dollars and weeks rather than being reconstructed from invoices in a panic. The third is the yield and shrink record that joins what you bought to what you sold, because on a $1.2 million shop one point of shrink is $12,000 a year and roughly five times the entire software bill. Our Prototype Sprint is $3,500 and is usually the right first step. Online stores start at $6,000 and full operations systems at $12,000.