Most replenishment reminders fire on a fixed interval. Thirty days after purchase, everyone who bought the 60-capsule bottle gets the same email.
Some of those customers take one capsule a day and have a month of product left. Some take three and ran out on day twenty. The reminder is early for one group and late for the other, and the brand reads the mediocre click rate as a copy problem.
The information needed to fix this is usually already collected. Quizzes ask about usage. The answers land in the email platform. Almost nobody wires them to the send time.
Replenishment timing needs a rate and a quantity. Everything else is refinement.
The rate is how fast the customer goes through the product. Servings per day, cups per week, scoops per feeding. The quantity is what they bought: bottle count, bag weight, pack size. Your order data already has the quantity. The quiz is the only place you get the rate.
Asking for it is less intrusive than it sounds, because usage questions do double duty. “How many cups a day do you drink?” is a recommendation input before it is a timing input, so it earns its place in the quiz regardless. The question design principles in quiz question psychology apply here: keep the answer set to four or five ranges rather than asking for a number, and put it in the middle of the quiz rather than at the start.
One caution about placement. Usage questions sit better after the customer has already invested a few answers. Opening with one reads like a sales qualification.
The arithmetic is not complicated, which is part of why it gets skipped.
A supplement customer buys a 60-capsule bottle. The quiz says they take two a day. That is 30 days of product. The reminder should land somewhere around day 24, giving them a week to reorder before they run out.
The reason to think in windows rather than dates is that the rate is self-reported and self-reports drift. Someone who says “daily” means five days a week. Build in the slack: a window of roughly the last 20% of the estimated supply, rather than a single day.
For a brand running Sensez, there is a shortcut worth knowing. The subscription engine includes frequency logic that takes the usage answer from the quiz and sets the delivery interval directly, choosing between 30, 45, and 60 day cadences based on what the customer said. That covers most consumable categories cleanly. It is a fixed set of intervals rather than an arbitrary per-customer day count, so if your product genuinely needs a 38 day cycle you are rounding to 45 and adjusting the copy rather than the schedule.
The rate is self-reported and self-reports drift. Someone who says “daily” means five days a week.
Three messages, and a decision at the end.
The lapse message earns its place because the answer changes what you do next, and because a customer who tells you they switched is worth more to you than one who silently stops opening.
The calculation is clean in some verticals and does not apply at all in others.
Supplements are the cleanest case on the list. Serving size and daily count are both known, the pack size is fixed, and the customer’s own answer gives you the rate. If you are going to build one replenishment flow, build it here first.
Pet food needs one more field. Daily feeding amount and bag weight give you the window, but multi-pet households break the math badly, and a two-dog home burns through a bag in half the time. Ask about pet count, and take feeding quantities from the manufacturer guidance on the bag rather than stating your own.
Coffee runs on cups per day against bag size. Grind preference adds a wrinkle worth respecting: pre-ground customers tolerate a slightly late reminder, whole bean customers who care about freshness do not.
CBD depends heavily on format. A tincture and a gummy pack deplete on completely different curves, and a topical is closer to unpredictable. Segment by format before applying any timing logic.
Wine does not work on a consumption rate at all. Nobody drinks two glasses a night at a predictable pace and if they do, that is not a fact to build a marketing calendar on. Wine replenishment runs on a cadence question instead: how often do you want to hear from us, monthly or quarterly. Let the customer set the interval and stop calculating.
Home goods and outdoor gear are not replenishment categories. A duvet cover and a rain shell do not run out. Both of these should be running something different: a seasonal trigger for outdoor gear, driven by the activity and season fields, and a complementary-product trigger for home goods, driven by room and project stage. The mechanics look similar, the input is not consumption.
People overestimate their usage. Not by much, but consistently, and in a predictable direction. Someone who says they take a supplement daily is describing their intention.
So the rule is: reorder behaviour beats stated rate. Once a customer has two actual reorders, the interval between them is better data than anything the quiz collected, and the flow should switch to it. The quiz answer is the best estimate you have for a customer with no purchase history, which is exactly the moment you most need one.
This is also the argument against treating quiz data as permanent. A check-in quiz six months in will catch the customer who dropped from three capsules to one, and what to do with quiz data after the first sale covers the broader version of the same problem.
If the flow converts customers to subscriptions, and it should, then the retention work moves to the subscription portal.
Ordergroove’s platform data is the clearest evidence available on what matters there. Subscribers who can skip an order stay 135% longer, and subscribers who can swap a product stay 71% longer. The most common reason people cancel is having too much product, which is the same problem bad replenishment timing creates, arriving through a different door.
For context on what normal looks like, Recharge analysed more than 15,000 subscription merchants and found average subscriber retention of 45% at six months and 33% at twelve. Replenishment categories generally sit above that line and curation categories below it, so read your own numbers against your category rather than the blended average.
Worth being clear about the division of labour: Sensez sets the frequency and drives the subscription sign-up, and the skip, pause, and swap controls live in your subscription platform. Recharge is one of the 30-plus integrations, but the portal behaviour is on that side of the line.
Reorder rate at the window is the number to watch, not open rate. The question is whether more customers reordered inside their estimated supply period than before you changed the timing, and that comparison needs a cohort, not a campaign report.
Two supporting metrics:
The metric that misleads is total email revenue, because a well-timed reminder often replaces revenue that a badly-timed one would eventually have captured anyway. Quiz analytics decoded goes further into attribution for quiz-driven revenue.
A duvet cover does not run out. If your product is not consumed, you are not building a replenishment flow, you are building a seasonal one.
Take your single best-selling consumable SKU. Find the usage question in your existing quiz, or add one. Calculate the window for that one product, build the three messages, and run it for a quarter against your current fixed-interval flow.
One product, one flow, one comparison. If it works, the pattern extends to the rest of the catalogue without further argument. If you would rather not retrofit it, our team builds the quiz logic and the timing together.
