It is Sunday night. The dishes are done, the workshop lights are off, and a laptop is open on the kitchen table with a coffee that went cold an hour ago. The monthly close has just finished. Revenue is up eighteen percent year over year. Order count is up. Average order value is up. By every leading metric on the dashboard, the business is winning.
The bank account, somehow, did not get the memo. Neither did the credit card balance. The accountant's email from last Tuesday — the one that opened with Could we talk before the quarter closes? — is still flagged red in the inbox. And underneath the eighteen-percent revenue line, in a smaller font, is the number that actually matters: net margin. It is down. Quietly, steadily, over the last six months.
This is the most common conversation in a maker business. Revenue grows. Margin shrinks. Nobody can say exactly when it started or what changed. The cause is rarely one thing. It is almost always a stack of small leaks, each too small to notice on its own, that together turn a healthy business into a tired one.
The short version. Shrinking margins on growing revenue is a leak problem, not a sales problem. Six symptoms cover the overwhelming majority of cases: silent supplier drift, packaging creep, shipping spillover, custom-order time bleed, marketplace fee inflation, and discount habituation. Each has a specific report or calculation that confirms it. Run the diagnostic in order — most leaks turn up in the first three symptoms — and fix the leak, not the symptom.
The rest of this post is the diagnostic. For each symptom: how to spot it, the specific number that confirms it, and what to do once you have found it. Bring a calculator and twelve months of receipts. The work is not glamorous, but it pays better than another Saturday market.
Why margins shrink quietly
A maker business has roughly thirty line items between the moment a customer clicks buy and the moment the deposit clears. Raw materials. Packaging. Postage. Marketplace listing fees. Marketplace transaction fees. Payment processing. Offsite ad commissions. Returns. Reshipments. Each of those line items has its own owner, its own pricing schedule, and its own quiet way of drifting upward over time. None of them ever sends a letter that says we are now eating four percent of your margin.
This is the central problem. The leaks are diffuse, and the dashboards most sellers use show revenue, not margin. A growing top line can mask a shrinking middle line for twelve to eighteen months — long enough that by the time the bank account flashes a warning, the business has lived inside the leak for a full tax year.
The six symptoms below are listed in rough order of frequency. The first three are responsible for most of the diagnoses that close before reaching the others. Work through them in order. Stop when one fits.
Symptom 1: Silent supplier price drift
What you notice. Material costs feel higher than they used to but you cannot point to a specific increase. You do not remember a supplier sending a price-change announcement. The recipe sheet is the same. The invoices, when you look at them, are also the same — just with slightly different totals.
What is actually happening. Supplier prices drift. Sometimes the announced way: a coffee importer raises green bean prices 7% in March, sends an email, you nod, and forget. More often the quiet way: the unit price stays the same but the case size dropped from 50 to 48 bars; the shipping line on the invoice grew by .80; the "fuel surcharge" appeared without an announcement; the minimum order moved from to . Each individual change is small enough to ignore. The compound effect over twelve months is the entire leak.
The number that confirms it. Pull the last twelve months of purchase orders for your top three suppliers. For each supplier, calculate the average effective cost per unit of your most-purchased SKU. Effective means landed: unit price plus shipping plus surcharges, divided by units received. Compare month-twelve to month-one. A 6-8% drift over a year on essential oils, wax, flour, or fabric is normal. A 12-15% drift is the leak.
Consider a composite scenario. A candlemaker scaling from market sales into wholesale checks her wax-flake invoices and finds the cost-per-pound has climbed from .12 to .47 over fourteen months. The supplier never raised the per-pound price. They quietly added a "fuel and handling" line that grew with every order — .50 in January, .25 by July, .00 by November. Nobody had opened a supplier invoice with a calculator in hand once during that period. The price did not change. The total did. The diagnostic missed it because the diagnostic was not running.
The fix. Three moves, in order. First, calculate the new effective unit cost for each of your top five inputs and update your cost-per-product math. Most cost sheets are running on numbers that were accurate eighteen months ago. Second, decide whether to absorb, pass through, or substitute. Absorb if the drift is small (under 5%) and your margin can take it. Pass through if the drift is meaningful (5-12%) and your retail prices have not moved in a year. Substitute if the drift is severe (over 12%) and an alternate supplier is available, which requires having scored two or three alternates per incumbent before the day you actually need them. Third, set a calendar reminder for ninety days from now to repeat the exercise. Supplier drift is not a one-time fix; it is a quarterly habit.
Symptom 2: Packaging creep
What you notice. The cost of a finished, ready-to-ship product is higher than the recipe sheet says it should be. The materials math checks out. Labor math checks out. There is a gap of to .50 per unit that you cannot quite account for.
What is actually happening. Packaging creep is the slow accumulation of unboxing details that each seemed harmless when added. The kraft mailer with the gold sticker. The branded tissue paper. The hand-written thank-you card. The custom dust bag. The polymailer inside the rigid box. The eco-foam peanuts. The little jar of bath salts as a "thank you" gift for orders over . Individually, each costs between fifteen cents and a dollar. Collectively, on a candle, they can add to of unmodeled cost — and almost none of it is in the original costing spreadsheet.
The number that confirms it. Pick one bestselling product. Lay every component of one finished unit on the table — the candle, the wick sticker, the bottom sticker, the warning label, the tissue, the rigid box, the ribbon, the thank-you card, the shipping mailer, the void fill, the address label. Cost each one individually. Add them up. Compare the total to the packaging line in your current cost sheet. If the lay-out total exceeds the cost-sheet number by more than 15%, the leak is here.
The fix. Two questions, in order. First: which of these components does the customer actually mention? Customer feedback (DMs, repeat orders, reviews) will name the candle and the scent. It almost never names the tissue. Anything no customer has mentioned in two years is a candidate for the trim list. Second: which components scale with the product price? A .50 ribbon is fine on a candle and absurd on a tea light. Match the packaging tier to the product tier. The brutal truth is that packaging is the easiest leak to plug because it does not require renegotiation, repricing, or saying no to anyone — just a quieter mailer.
Symptom 3: Shipping spillover
What you notice. Shipping is supposed to be a pass-through cost. The customer pays it; you charge it on the receipt; the carrier takes it. Somehow, every month, the total shipping you paid is higher than the total shipping you collected from customers — by a few hundred dollars, then a few thousand.
What is actually happening. Shipping spillover comes from three places, in roughly this order:
- Free-shipping promises. A free-shipping threshold at on a product line where the average order is means you eat the shipping on most orders. The math feels good in the conversion column and bleeds in the COGS column.
- Dimensional weight surprises. USPS, UPS, and FedEx all charge by the greater of actual weight or dimensional weight (length × width × height ÷ a divisor). A light-but-bulky product like a knit scarf in a 12×9×4 mailer is billed at dim weight, not the half-pound it actually weighs. If your shipping calculator is set on actual weight, every shipment underbills the customer and overcharges you. The USPS dimensional weight rules and divisors are published in the USPS Domestic Mail Manual — pull the divisor for your service class and recompute.
- Class drift. A product that was Priority Mail Small Flat Rate two years ago crept past the flat-rate dimensions when the packaging got fancier (see Symptom 2) and is now being shipped via Priority Cubic or Ground Advantage at three times the cost. The shipping label printer does not flag this; it just prints the new label and charges the higher rate.
The number that confirms it. Pull one month of shipping. Two columns: shipping collected from customer, shipping paid to carrier. Sum each column. The gap is the spillover. If it is more than 5-8% of the carrier total, dig deeper: are the highest-loss orders the free-shipping ones, the dim-weight ones, or the class-drift ones? The reason for the spillover dictates the fix.
The fix. Three options, one each:
- Free shipping is a marketing tool, not a fulfillment policy. Raise the threshold until the average order size plus the average shipping cost still clears your target margin. The right number is usually higher than your gut.
- Dim weight is a packaging problem. Trim box size. Or move from Priority to Ground Advantage where dim weight rules are friendlier on bulky-light items.
- Class drift is a product redesign problem. If the packaging push out of flat-rate dimensions saved fifty cents per unit on the box but added three dollars per unit on shipping, reverse the change.
If sales run through more than one channel, run the in-versus-out split per channel as well as in aggregate. The leak is usually concentrated in one storefront, not spread evenly across all of them.
Symptom 4: Custom-order time bleed
What you notice. Custom orders feel good — bigger ticket, happy customer, hand-written thank-you note — and yet, on weeks dominated by custom work, the profit-and-loss looks worse than weeks dominated by standard inventory sales. The accounting does not match the feeling.
What is actually happening. Custom orders quietly eat margin in two ways:
- Labor inflation. A standard scented soap bar takes four minutes to make from pour to label. A custom-color, custom-fragrance soap bar for a wedding favor order takes fourteen minutes — and almost none of that delta shows up in the price. The price went up 30%. The labor went up 250%. The math does not survive that gap.
- Hidden communications cost. Custom orders come with email threads, mockup approvals, shipping address corrections, and the occasional "can you make one more in pink?" exchange. None of that time appears on the invoice. None of it is billable. It is also not zero — for a wedding favor order, it can easily be ninety minutes of unpriced labor before the first soap is poured.
The number that confirms it. Pick three recent custom orders and three recent standard orders of similar revenue. For each, log the total hands-on minutes (production, labeling, packaging) and the unbilled minutes (emails, calls, mockups, revisions). Multiply hands-on minutes by your target labor rate; multiply unbilled minutes by the same rate. Add both to the COGS. Recompute margin per order. If the custom-order margin is more than 10 percentage points below the standard-order margin, the leak is real.
The fix. Either price custom work at the rate it actually requires (typical handmade brands need a 40-60% custom premium, not the 10-15% most sellers apply), or stop taking custom orders below a minimum size that justifies the unbilled communications cost. A common rule of thumb in the small-batch food and soap world is a 24-unit minimum on custom orders — small enough to be friendly, large enough to spread the email overhead. The Home Baker's Order Tracker covers a similar problem for cake bakers; the principle generalizes.
Symptom 5: Marketplace fee inflation
What you notice. Etsy, Shopify, Faire, Amazon Handmade — whichever marketplace is the largest revenue source — is taking a bigger share than it used to. Listings have not changed. Pricing has not changed. The deposit per sold has.
What is actually happening. Marketplaces add fee categories more often than they raise the headline rate. Etsy is the most visible example. The transaction fee (currently 6.5%) is the line most sellers track. Underneath it: a 3% + .25 payment processing fee, a regulatory operating fee that varies by country, Offsite Ads commissions at 15% for sellers below ,000 in trailing-twelve-month sales and 12% for sellers above that threshold, the listing fee of .20 per listing, listing renewals when an item sells, and currency conversion charges if the buyer pays in a different currency than the listing. Etsy publishes the full fee schedule on its fees and payments page — pull it and reconcile against your last three monthly statements.
The math compounds because the fee categories are layered, not bundled. A candle sold via Offsite Ads incurs the transaction fee on the sale price, the processing fee on the sale price plus shipping, the regulatory fee on a different base, and a 12-15% Offsite Ads cut on the sale price plus shipping. The all-in fee on that one order can land anywhere between 13% and 28% depending on which categories trigger. A seller assuming "Etsy takes 6.5%" is off by a factor of two to four.
The number that confirms it. Pull a recent Etsy statement (or the equivalent for your channel). Calculate total fees as a percent of gross sales. Compare to the same month last year and to the same month two years ago. A drift from 12% to 17% over twenty-four months is the leak.
The fix. Three moves, in order:
- Reprice listings to be fee-aware. Fee-aware pricing means knowing that on a candle, Etsy may take . The retail price needs to be set against the post-fee number, not the pre-fee one. Build the all-in fee line into the cost sheet directly so future repricing is mechanical, not investigative.
- Reconsider Offsite Ads opt-in status. Sellers under in annual Etsy sales can opt out of Offsite Ads. Sellers over cannot. If you are near the threshold, the Offsite Ads cut materially changes the breakeven on every listing it touches.
- Add a second channel that does not stack fees. A Shopify storefront, a direct-checkout site, a craft fair appearance, or a wholesale customer all have lower percentage fees than a fee-stacked marketplace. The right answer is rarely leave the marketplace — it is usually stop relying on a single marketplace to carry the margin.
Symptom 6: Discount habituation
What you notice. Every promotion seems to "work" — sales go up during a sale and drop when it ends. The dashboard reads like a series of small wins. The bank account does not.
What is actually happening. Discounts have two costs. The first is obvious: every percent off the listing price is a percent off the margin, dollar for dollar. The second is invisible: customers learn the discount cadence and begin timing their purchases to it. The first 20%-off-Memorial-Day sale gives you a real lift over baseline. The fourth 20%-off-Memorial-Day sale, after three years of running it, is mostly converting customers who would have bought anyway. The discount is now a transfer of margin from the business to the customer, dressed as marketing.
Discount habituation is hardest to spot because it does not show up in any single sale's accounting. It shows up in the slow rise of promotional sales as a percent of total sales over years. A maker who sold 12% of units at a discount in 2022, 19% in 2023, 27% in 2024, and 36% in 2025 has a discount-habituation problem regardless of how strong revenue looks year over year. The product's effective price — the average dollars actually received per unit — has dropped even when the listing price has not.
The number that confirms it. Take any twelve months of sales and split them into two buckets: full-price units sold and discounted units sold. Calculate the effective price per unit across all sales (total revenue / total units). Compare to the listing price. The gap between effective and listed price is the cost of the discount habit. A 4-5% gap is normal. An 11-15% gap is the leak.
The fix. Cold turkey is rarely the answer, because the loyal customer base has been trained on the cadence. The cleaner fix is to change the cadence. Replace four annual broad-discount events with two narrower ones — clearance for end-of-season inventory, a tiered loyalty offer for repeat customers — and protect the listing price on everything else. Run the volume-versus-margin math on the change before announcing it: cutting two of the four annual sales costs you some discount-driven volume; the question is whether the recovered margin on the protected sales clears the loss.
The monthly margin check: a four-step decision framework
The six-symptom diagnostic is heavy. The good news: it only runs when a lighter, faster check flashes a signal. Run the four steps below at every month-end. The full check takes about ten minutes once the inputs are wired up.
Step 1 — Pull three numbers for this month and the same month last year.
- Total revenue (top line, before fees)
- Total COGS (materials + packaging + direct labor + payment processing + marketplace fees + shipping cost)
- Gross margin percent:
(Revenue - COGS) / Revenue
Step 2 — Compare year over year.
- Revenue change: this month vs. same month last year
- Gross margin change: this month vs. same month last year, in percentage points, not in dollars
A revenue change without a corresponding margin change is fine. A revenue increase alongside a margin decrease is the signal.
Step 3 — If gross margin dropped by more than 2 percentage points, route to the diagnostic.
The thresholds below are rules of thumb, not laws. Adjust to your business if you have better historical data:
- Margin dropped 0-2 points: Normal monthly noise. Log it. Move on.
- Margin dropped 2-4 points: Run Symptoms 1, 2, and 3 only (supplier drift, packaging creep, shipping spillover). These three cover the majority of moderate leaks.
- Margin dropped 4+ points or has been dropping for three months in a row: Run all six symptoms in order. Stop when one fits.
- Margin dropped 6+ points or revenue is also dropping: This is no longer a margin leak; it is a business model question. Time to look at channel mix, product line, and pricing strategy as a whole, not at the leak symptoms.
Step 4 — Fix one leak and re-measure next month.
The instinct after a bad diagnostic is to fix everything at once. Resist it. Fix one leak. Let the next month's numbers tell you whether the fix landed. If margin recovers two points, the fix worked. If it does not, the diagnosis was wrong or the leak was somewhere else. Single-variable changes are the only way to know what is moving the number.
The End-of-Month Closeout Checklist is built around this four-step rhythm — print one, keep it in the binder, and the monthly check turns into a habit instead of an emergency.
Where the tooling actually helps
Most of this diagnostic is doable in a spreadsheet, with twelve months of invoices, and a quiet Sunday. The bottleneck is not the math; it is the data. Most maker businesses lose three to four hours per month just assembling the inputs: pulling Etsy CSVs, copying Shopify exports, hunting through email for supplier invoices, hand-totalling shipping receipts.
Ardent Seller closes that gap by keeping the inputs together: per-supplier purchase history that surfaces effective unit cost over time, per-product packaging and direct-labor cost lines that ride with each SKU, marketplace fee tracking that splits out each fee category by month, and shipping in-versus-out reporting on the same screen. The four-step monthly margin check, in a system that already has the data, takes about three minutes — short enough that running it never feels like a chore, which is the only reason it actually gets run.
Tooling does not replace the diagnosis. It removes the excuse for skipping the diagnosis. The leaks are always there. The question is whether the rhythm of looking for them is.
Run the diagnostic once. Then run it again next quarter.
Shrinking margins are not a failure of effort. Effort is the part most maker businesses have in surplus. Shrinking margins are a failure of attention — specifically, attention paid to revenue at the expense of the half-dozen smaller numbers that compound into margin.
The diagnostic above is one Sunday's work. The monthly check is ten minutes. Together they are the difference between finding a leak in month two of a slow drift and finding it in month sixteen after the credit card balance has caught fire.
If the math feels heavy because the inputs are scattered across four spreadsheets, three CSV exports, and a shoebox of paper invoices, that is its own diagnosis. Start a free Ardent Seller trial and keep the supplier history, packaging cost lines, marketplace fee breakdowns, and monthly close in one place. The diagnostic stays the same. The data stops fighting you.
Related reading
- COGS Explained: A Maker Business Primer — The foundational COGS framework underneath the diagnostic above; useful if "what counts as COGS" is part of the confusion.
- Margin vs. Markup: The Pricing Math Mistake That Costs You Money — Half of margin diagnoses fall apart because the spreadsheet was computing markup when it claimed to be computing margin. Worth a re-read before running the diagnostic.
- Against the "Just Charge More" School of Pricing Advice — A companion piece on why the reflex to raise prices is sometimes wrong, and what to check first.
Free resources
A few free downloads from the Ardent Workshop library that pair well with this post:
- End-of-Month Closeout Checklist — Turns the four-step monthly margin check into a printable habit, alongside sales reconciliation, expense review, and a P&L glance.
- Should I Raise My Prices? Decision Tool — Once a leak is identified and a price hike is on the table, this runs the +10%, +20%, +30% volume-versus-margin math so the move is informed instead of hopeful.
- Product Pricing Calculator — A working Excel pricing model where the updated supplier costs, packaging trims, and platform-fee math all plug in and a defensible retail price drops out.
This article is provided for educational purposes only and does not constitute financial, tax, or business advice. Cost structures, pricing examples, and margin figures are illustrative and will vary by your specific circumstances. Consult a qualified accountant or small-business advisor before making financial decisions based on this content.