The average US consumer unit — roughly a household — spent $978 on personal care products and services in 2024 (Bureau of Labor Statistics, Consumer Expenditures — 2024 (opens in new tab), released 19 December 2025). That figure bundles haircuts and manicures in with the jars and bottles, so the product-only slice is smaller — but a single $26 jar of body butter is still a rounding error inside it.
So the customer who bought that jar in March and never came back did not stop buying skincare. They bought it somewhere else — or, more often than makers expect, they are still working through the jar and got counted as gone about eight weeks too early.
Skincare should be the easiest category in the handmade economy to earn a second sale in. The product runs out. There is a date, knowable in advance, when the customer needs more. Many small skincare shops never work out what that date is, and then read the silence before it as rejection.
Your one-and-done number is wrong before you start
Before diagnosing anything, subtract the buyers who were never candidates for a second purchase.
Every skincare shop's customer list is padded with people who bought exactly once by design. Gift buyers — a set bought in December for someone else, by someone who does not use the product and never intended to. Market walk-ups who live three states away. The friend-of-a-friend who was being supportive. A wholesale buyer's sample order. None of these people lapsed. They completed the only transaction that was ever on offer, and they are sitting inside a "one-and-done rate" making it look like a retention problem.
This matters more than it sounds, because the padding is not evenly spread. A shop that does 40% of its annual volume at holiday markets and gift sets will tend to carry a worse-looking repeat rate than a shop selling the same products online year-round — not because it retains customers worse, but because a larger share of its buyers were shopping for someone else. Comparing those two shops' headline numbers tells you nothing.
The correction is unglamorous: separate buyers who bought for themselves from buyers who bought for someone else, and stop counting the second group as churn. For a single season's worth of markets, this is often possible from memory. After that it needs to be recorded at the point of sale, because nobody remembers in April who was buying a gift in November.
What remains after the subtraction is the real denominator — people who used the product, ran out, and made a decision. Those are the ones worth diagnosing, and there are three ways they go quiet.
The reorder clock is a number you already own
A jar has a use-up window. Container size, divided by how much comes out per use, times how often it gets used. That is not a marketing metric, it is arithmetic, and the person best placed to run it is the one who formulated the product.
Nobody else will. In the US there are no laws or regulations that require cosmetics to have specific shelf lives or expiration dates on their labels, and the FDA "considers determining a product's shelf life to be part of the manufacturer's responsibility" (FDA, Shelf Life and Expiration Dating of Cosmetics (opens in new tab)). Shelf life and use-up rate are different questions, but they share a source: the maker is the only party who knows either one.
Run it for each format you sell. The numbers below are illustrative — substitute your own container sizes and your own sense of how customers actually use the product:
| Format | Container | Rough use per application | Typical frequency | Weeks of supply |
|---|---|---|---|---|
| Facial oil | 30 ml | 4–5 drops | Twice daily | ~9 |
| Whipped body butter | 8 oz | A heaped teaspoon | 3× weekly | ~15 |
| Cold-process bar soap | 4.5 oz | Not applicable | Daily shower | ~4 |
| Lip balm | 0.15 oz | Not applicable | Several times daily | ~4 |
Four products, one shop, and a spread from about a month to nearly four. A single "customers who haven't bought in 90 days have churned" rule is wrong for every row in that table — too impatient for the body butter, far too relaxed for the lip balm. The bar-soap buyer who has been silent for 90 days made a decision two months ago and nobody noticed.
The estimate is only a starting point. Once you have a handful of repeat buyers, the median number of days between their orders is the measured version of the same figure — and it is one of the headline numbers on Ardent Seller's customer retention report.
The useful version of this number is per customer, not per product. Someone who buys a bar of soap and a facial oil together is on two clocks at once, and the shorter one is the one that expires first.
Three shops, same one-and-done rate
The three shops below are composites — invented to make the arithmetic concrete, not customers of anyone. Each has roughly the same headline repeat rate. Each has a different problem, and the fix for one would do nothing for the other two.
Case study: the shop that panicked at 90 days
Nadia sells one hero product: an 8 oz whipped body butter at $26, mostly online, mostly to people buying for themselves. She had read that 90 days of silence means a lost customer, and by that measure she was losing almost everybody.
Her jar lasts most people somewhere between three and four months. So her 90-day mark lands before the average customer has finished the jar. Every quarter she was writing off a cohort that had not yet reached the moment of decision, then discounting to win back people who were not going anywhere. Some of them took the discount, which taught her regular buyers to wait for one.
The diagnosis was not a retention failure. It was a measurement window shorter than the product's own life. Her real question — how many customers reach month five without reordering — she had never asked, because she had already declared them gone in month three.
Nadia's fix is not a win-back campaign. It is to move the line: stop measuring silence at 90 days and start measuring it at the week the jar actually runs out.
Case study: the shop with a real lapse and no reminder in place
Teo sells a 30 ml facial oil. His committed customers reorder every nine or ten weeks, which matches the arithmetic almost exactly. A meaningful group of them drifts to five or six months and then reappears, usually after seeing a post.
Those customers had not chosen a competitor. They ran out on a busy week, meant to reorder, and the thought did not survive the month. Nothing about the product failed. What failed was that the only reminder in the system was Teo happening to post at the right moment, which is a coin toss run once a week.
This is the one failure mode that is genuinely fixable by asking. It is also the one that looks identical to every other kind of silence from the outside — which is exactly why it goes unaddressed.
For Teo, the answer is a reminder at week nine rather than a loyalty program. One scheduled message at the point the bottle empties does the work a discount is currently being asked to do.
Case study: the shop whose bundle quietly failed
Bea sells single bars and a four-piece starter set. Her single-bar buyers come back at a healthy rate. Her set buyers barely come back at all, and for two years she read that as normal — a set is a bigger purchase, so surely people need longer.
Splitting the two groups told a different story. The set contained a clay mask that a noticeable share of customers found harsh. Nobody complained; they simply stopped. The mask took the other three products down with it, because a customer who had a bad experience with one item in a set does not go back and work out which item it was. They write off the brand.
A set is a bet that every item lands. When one does not, the failure is silent and it is attributed to the whole shop.
Bea's next step is a comparison, not a change: set buyers against single-item buyers, before touching the recipe or the price. A gap there is a product signal, and it names which bet is not paying.
The three facts that separate them
All three shops could see the same thing — customers who bought once and stopped. Telling them apart takes three facts about each departing customer, and none of them require a survey:
- Did they buy for themselves? Separates real lapses from buyers who were never on a clock. Record it at the point of sale, in the moment, because it is unrecoverable later.
- What did they buy? The product sets the clock. Without it, "90 days quiet" means four different things across a four-product catalog.
- Did they ever come back for anything? A customer with two or more orders has revealed their own rhythm. A customer with one has not, and the two groups need completely different treatment.
Those three facts are what turn a customer list into a diagnosis. They are also how Ardent Seller's retention report reads one: it measures each person against their own history rather than against a fixed global rule.
That is why its went-quiet list is narrower than it first looks:
- It only ever contains customers with two or more orders — people whose typical gap between orders is already known.
- It flags one only when the silence exceeds twice that customer's own average gap.
- A 60-day floor sits underneath that, so a fast-moving product does not raise an alarm every fortnight.
Nadia's body-butter buyers would never appear at 90 days. Bea's set buyers, having only ever ordered once, would not appear at all, because they are a different problem with a different fix.
Two more headline figures are worth watching beside the median gap: repeat rate and one-and-done share. One caveat is worth naming: sales recorded without a customer attached, including unattributed marketplace orders, are excluded from the math entirely. A shop that sells mainly through a channel that hands over no customer identity has a structural blind spot here, and no report can close it. (The platform comparison covers what each sales channel does and does not hand back.)
Start the clock this week
None of this requires a loyalty program or an email platform. It takes three moves, in the same order as the facts above:
- Mark the gifts. Go back through the last six months of sales and flag the ones bought for someone else. Those buyers were never on a clock, and leaving them in the denominator is what makes the one-and-done rate unreadable.
- Run the arithmetic on your top-selling product — container size, use per application, uses per week — and write the number down. That is the earliest date a reorder could reasonably happen, and every retention judgment you make before it is noise.
- Split your list by order count. Anyone with two or more orders has shown you their own rhythm and can be judged against it. Anyone with exactly one has not, and belongs in a separate pile until they buy again.
If you sell sets, add a fourth check: the set-versus-single-item comparison from the bundle case study above.
The shops that get repeat business in this category are rarely the ones with the best loyalty scheme. They are the ones who know, for each product they make, roughly when it runs out — and who ask at that moment instead of ninety days early or five months late.
Ardent Seller records the customer on the sale, tracks each one's own reorder rhythm, and surfaces the repeat buyers who have gone quiet past their own typical gap — free on every plan.
Related reading
- Why subscription box customers cancel in the first 90 days — the same diagnosis where a renewal date exists to read: four causes of subscription churn and the fingerprint each leaves in the renewal ledger.
- Email marketing for a handmade business — the fix for the reminder-gap case above: what to send on the weeks when nothing is launching, so the reorder reminder is not a coin toss.
- Pricing handmade skincare — before you work out how often a customer reorders, work out what the jar earns you when they do.
- Which online selling platform handmade sellers should use — the channel decision that determines whether you get a customer record at all.
Free resources
Two free downloads from the Ardent Workshop library that pair with the diagnosis above:
- Customer Service Response Starter — templates for the reorder-time message this post says nobody sends, without it reading as a sales pitch.
- Should I Raise My Prices? Decision Tool — useful once you know your repeat rate, because a shop with genuine repeat buyers has far more pricing room than one selling to strangers every time.
This article is provided for educational purposes only and does not constitute legal, regulatory, cosmetic-safety, or health advice. Cosmetic labeling, shelf-life determination, and product-safety requirements vary by jurisdiction and product, and change frequently. Consult a qualified cosmetic regulatory consultant, your state health department, or an attorney before making compliance or safety decisions.