What to Sell: Choosing a Cottage Food Product Line That Works at the Market

Most people who start a cottage food business can already make twenty things well. The market does not need twenty things from you. It needs three or four that you can make every week, that hold up on a folding table in August, that your state lets you sell, and that the vendor two tents down is not already selling. Choosing those few products is the first real business decision you make, and it shapes the labels, the pricing, and the Friday-night production schedule that follow.

This guide is five filters. Run every candidate through them in order and the list gets short on its own.

Filter 1: What your state will let you sell

The Harvard Law School Food Law and Policy Clinic's fifty-state review puts it plainly: states generally limit cottage food production to foods with a low risk of causing foodborne illness, the ones that fall outside the FDA Food Code's definition of TCS because they do not need time and temperature control for safety. The clinic also names the less obvious members of the risky group: low-sugar jams, cooked vegetables, and low-acidity pickles and salsas all count as TCS. Some states write only a definition (Utah's law, in the clinic's 2018 survey, simply said cottage foods cannot be potentially hazardous); most publish a list. So the first step is dull but decisive: pull up your state's list and read both halves of it, the allowed side and the ineligible side.

The ineligible column is where new sellers get surprised. Pumpkin pie is a fall staple at any ordinary bakery and is off the table under California's approved list and Colorado's, which also bars sweet potato and cream pies, sauces, and buttercream made with butter (ghee or vegetable oil versions pass there; California allows butter buttercream). Wisconsin has no general cottage food statute. Its home-canned foods exemption covers pickles, salsas, sauerkraut, and fruit jams at pH 4.6 or lower, sold only at farmers markets, flea markets, and community events; lemon curd, pesto, and pickled eggs need a license regardless of pH. Baked goods are sold there under a 2017 court order, not a statute, and candy and other unbaked goods still need a license, as the Legislative Council's 2025 brief explains.

Texas went the other direction. Since SB 541 took effect on September 1, 2025, a Texas cottage food operation may make any food except meat, poultry, seafood, ice and frozen desserts, low-acid canned goods, raw milk, and CBD or THC products. Refrigerated foods are allowed, but selling them means registering with DSHS, printing safe handling and temperature instructions on the label, and keeping a riskier product safe on a hot table. Our Texas guide has the details.

Caps shape lineups too. Under Colorado's current Act, blueberry, banana, and chocolate chip muffins are three products with three separate $10,000 allowances, so range pays. HB26-1033, signed June 4, 2026, replaces that on January 1, 2027 with a single $150,000 net revenue cap, annual registration, and refrigerated foods after extra training, with one catch: a producer may sell only one type of refrigerated product, in up to five variations. Florida's $250,000 gross cap with no wholesale means every dollar of the cap is a direct sale, so margin per dollar sold is the number to watch. Our explainer on sales caps covers how limits are counted.

Filter 2: Shelf life and the drive to market

Once a product is legal, ask whether it will still be good on Saturday afternoon after being made Thursday night and riding in a hot car. "Shelf-stable" comes down to two numbers. Bacteria need available water to grow, measured as water activity on a scale from 0 to 1, and the FDA draws its line at 0.85. Acid does a similar job. Kansas State's water activity fact sheet says keeping pH below 4.6 or water activity below 0.85 is what makes a food safe to hold without refrigeration; Clostridium botulinum needs a water activity of about 0.93 to grow at all. Minnesota's cottage food law requires 0.85 or lower, or pH 4.6 or lower, and many state lists use the same lines.

Kansas State's typical values, in the table below, explain why some products are easy and some are not. Crisp cookies and crackers sit so far below the line that spoilage is a texture problem, not a safety one; sugar keeps jams comfortably under it; sponge cakes sit above it.

Moist quick breads are the classic edge case. University of Minnesota Extension tested a banana bread and got individual readings of 0.852 and 0.855, over the limit, even though the five-sample average was 0.8196. A recipe that lives at the line can pass one week and fail the next with a riper banana or a shorter bake, and Extension warns that any change to ingredients or bake time can move the number. It found a usable meter for under $100 and still recommends commercial lab verification for anything close.

The practical rule: pick products that are stable by a wide margin, treat anything at the line as a product with a testing budget attached, and read food safety fundamentals for home producers before you sell any of it. Then run the trunk test: pack a batch the way you will sell it, drive it to the market and back, and open the boxes. Bars, loaves, crackers, granola, and jars arrive the way they left. Frosted cupcakes, decorated sugar cookies, and meringues often do not, and a product that needs a second person to carry it has a hidden labor cost.

Filter 3: Can you make it forty Saturdays in a row?

A product that draws a crowd once and cannot be reproduced is a demo, not a product. The question is whether you can make it well every week, in a home oven with two racks, after a day job, at a standard a regular could buy blind.

The number that answers this is margin per labor hour, and it is why two products with the same sticker price can be very different businesses. A hypothetical pair: a batch of jam yields 12 half-pint jars, takes two hours including jarring and labeling, sells at $8 a jar, and costs $30 in fruit, sugar, jars, and labels, so $96 in revenue, $66 over cost, $33 an hour before overhead. Two dozen decorated sugar cookies also bring in $96 at $4 each and cost only $15, but decorating pushes the work to five hours: $81 over cost, about $16 an hour. Same revenue, half the hourly return. Your numbers will differ; run the comparison with your own receipts and a timer before you fall in love with a product. The full costing method, including overhead and the market fee, is in how to price baked goods and homemade foods.

Two other capacity questions belong here. Oven throughput: four trays through a home oven for one market's inventory eats your Friday; loaves and jars do not. And tolerance: some people find jarring 60 jars restful and decorating 60 cookies miserable, and others are the reverse.

Filter 4: What your market already has

Baked goods are the one category on nearly every state's list; in the Harvard clinic's survey, the narrowest law, Oklahoma's, allowed little else. Cookies, brownies, and cupcakes tend to be the default first product for a new cottage seller, and the result at many markets is a bakery row. We could not find a reliable national count of what cottage producers sell by category, so treat that as an observation to check, not a statistic. The check takes one Saturday: walk the market before you apply and write down every vendor selling in your candidate category. Walk a second market too. Colorado State's extension staff note that different markets attract different demographics, and a product that is redundant downtown on Saturday may be the only one of its kind at a Wednesday neighborhood market ten miles away.

The market manager is doing the same count from the other side. The Farmers Market Coalition's manager FAQ advises watching what happens when a new vendor is admitted for a product already sold, and reconsidering the number allowed next year if the existing vendor's sales drop and do not recover; some markets set quotas or limit producers in a category, particularly non-farm vendors. In a full category, your application competes with people who already have a spot. In an empty one, it fills a gap.

Penn State's Food for Profit series asks the question you need to answer for the fourth cookie table: "Is my product significantly different from those currently on the market?" Different can mean a form nobody else offers (a savory scone next to sweet ones), an ingredient story (the fruit came from the farm two tents over), a dietary niche, or a price point. It cannot mean "mine taste better," because every table believes that. Penn State also asks whether people would eat the product every day or only on special occasions; an everyday product builds regulars, an occasional one needs a bigger crowd.

Under-served categories show up in the state lists themselves. California's approved list includes crackers, flatbreads, pizzelles, baklava, churros, fruit-only empanadas and tamales, granola, baking and soup mixes, spice rubs, nut butters, and fruit-infused vinegars; Colorado's adds tortillas, freeze-dried produce, and dry teas. None of these is "more profitable" than cookies, and we would be suspicious of anyone who says one is. They are simply less likely to have three competitors already, and several travel and keep better than a cupcake.

Filter 5: What you will sell in February

Seasonality shapes both ingredients and demand. USDA's Seasonal Produce Guide lists berries, cherries, peaches, and plums for summer and apples, pears, pomegranates, and pumpkin for fall, with the caveat that produce in your area will vary with growing conditions and weather. Strawberry jam is a June product in most of the country; apple butter is an October one. Both are good products. Neither is a year-round hero.

So plan the year, not the week. If your market runs May through October, a lineup anchored on fresh fruit fits it. If you sell at a winter market or through pre-orders, at least one product needs ingredients you can buy in January: grains, nuts, dried fruit, chocolate, citrus. Holiday demand is its own season, and a gift-able product that keeps for weeks earns more in December than a loaf that must be eaten by Sunday. Our guide to planning a year of products turns this into a calendar.

The lineup: one hero and a few supporting players

Run twenty ideas through those five filters and you will usually have four or five left. That is the right size. Each product you add needs its own cost sheet, label, packaging, and inventory guess, and each one is another thing that can go unsold at 1 p.m. A small lineup with clear roles beats a long one.

The hero should be the product that scored best on every filter, especially labor margin and stability, because it will carry most of your volume; our guides to selling sourdough and selling jams, jellies, and preserves each work through one such hero. The supporting items make the table easy to buy from at every price and give regulars something new to try. Where a per-product cap applies, as in Colorado through 2026, flavor variations of the hero add allowances without adding a whole new process.

Test with small batches before you print 500 labels

Nothing in this guide replaces selling. Colorado State's extension guidance recommends starting with small batches because very few owners get things right the first time, and the market tells you what you got wrong faster than any spreadsheet. A workable test:

1. Pick three finalists, no more, and make a small batch of each, sized so that selling out by noon is possible.
2. Price them from cost, not from what feels safe.
3. Sell all three for four consecutive markets. Each week, write down what sold out and when, what needed marking down at close, and who came back.
4. Offer samples where the rules allow. Texas now lets cottage operations sample at any location; other states and markets set their own rules, so ask the manager first.
5. Keep the product that sells through or draws repeat buyers. Drop the one you had to discount two weeks running, however much you like making it.

Four weeks of real sales will tell you more than any survey of friends, who will always say yes, and they produce the first honest numbers for your cost sheet.

What to do next

Print your state's allowed and ineligible lists and strike every candidate they rule out; if a product is not clearly on the list, ask the agency in writing, and start with what counts as cottage food if the categories are new to you. Then make each finalist once, time it, run the trunk test, walk your market, and pick a hero, two or three supporting products, and a seasonal rotator for the four-week test.

If you have not registered, applied, or read your state's labeling rules yet, the cornerstone guide to starting a cottage food business covers those steps in order. The product line comes first because everything after it is easier when you know what you are selling.