First-Party Purchase Data
Order history a retailer collects directly from its own customers, as opposed to data bought from a broker or inferred from tracking across other sites. Reorder prediction runs on this and needs nothing else.
Glossary of Reorder, Replenishment and Recurring Revenue Terms
There's a gap in e-commerce strategies Shopify Merchants don't have vocabulary to describe. Terms for behavior between one-time purchases and subscriptions. Between buy once and disappear consumers and committed subscribers. Most recurring revenue opportunities live in that gap and are earned with timely reminders that nudge shoppers toward repurchasing when they're most likely to.
The problem's that the going terminology for that space has been borrowed from subscriptions, email marketing, and retention analytics, leaving retailers without a shared language for what happens when a customer finishes their last bag of gourmet coffee, or realizes their favorite ferret's kibble bin's almost empty.
Reorder timing and replenishment reminder nudges aren't subscription concepts or traditional marketing concepts. They're something else. So. We built this glossary to help give that something else a vocabulary. If you want to understand the metrics behind recurring reorder strategies, you'll find actionable definitions here.
Goods that get used up and need replacing on a predictable rhythm, from supplements and pet food to skincare and coffee. Depletion is what creates the reorder, so how fast a product is consumed matters more than which category it sits in.
Speed at which one customer works through a given quantity of product. Two shoppers buying the same tub can empty it three weeks apart, which is why store-wide averages mislead.
Full loop from one order to the next for an individual shopper, covering purchase, use, depletion and repurchase. Knowing where someone sits in that loop is what makes timing possible.
Order history a retailer collects directly from its own customers, as opposed to data bought from a broker or inferred from tracking across other sites. Reorder prediction runs on this and needs nothing else.
How often someone buys across a set period. Averaged across a store it reads as a health metric. Measured per customer it becomes the input that decides when outreach should land.
Elapsed days between one order and the next for the same shopper and the same product. Track it long enough and the number stops being history and starts being forecast.
Ceiling on how many messages one customer can receive in a given window, set so reminders stay useful rather than turning into noise. Caps usually run per channel, since tolerance for SMS is far lower than for email.
Sending the same reorder reminder through whichever channel a customer responds to, across SMS, email, on-site and push. Channel preference varies far more than most retailers assume.
Checkout path that rebuilds a shopper's previous order in a single tap. Every step removed between intent and purchase raises the share of reminders that turn into revenue.
Range of days when a customer is most likely to repurchase, calculated from their own history rather than a store-wide default. It stays a window rather than a date because behavior drifts.
Software that reads first-party purchase data, predicts when each shopper is ready to buy again, and reaches out at that moment. Different from a scheduler, because nothing is fixed in advance.
Practice of deciding when to prompt a repurchase. Fixed intervals ignore whether the shopper is actually ready. Adaptive timing recalculates with every new order that arrives.
Message prompting a customer to reorder something they are running low on. Effective versions read as useful rather than promotional, because the timing does most of the persuading.
Bold's term for a well-timed reminder sent through whichever channel a shopper actually reads. Same job as a replenishment reminder, with prediction rather than a schedule behind it.
Choosing the hour a message goes out based on when each recipient has opened or bought before. Repeat order timing settles which day. Send-time optimization settles the hour inside it, and the best hour shifts by channel.
Describes the revenue from customers buying again, whether or not a subscription is involved. Broader than subscription MRR, which counts only the committed minority.
Portion of customers who buy the same product again rather than any product again. Narrower than repeat purchase rate and more honest about real replenishment demand.
Share of customers who buy more than once in a period. Usually the fastest read on whether a store has a retention problem or an acquisition one.
Percentage of customers still buying at the end of a period who were buying at the start of it. Small movements here compound heavily across a year.
Share of orders in a period placed by people who had ordered before. Reads on the order side, where repeat purchase rate reads on the customer side.
Speed at which subscribers cancel. Splits into voluntary cancellation and involuntary loss from failed payments, and the two need entirely different fixes.
Shopper who has moved well past their own expected reorder point without buying. Defining lapse against individual history rather than a store-wide cutoff catches people earlier.
Consent models governing who a store may message and on which channel. Requirements shift by channel and jurisdiction, with SMS and EU privacy rules usually the strictest.
Outreach aimed at customers who have stopped buying. Works best when it fires on a missed personal reorder window instead of an arbitrary ninety days of silence.
Vendor charging a share of the revenue produced rather than a monthly fee. Risk sits with the vendor, and the retailer pays nothing in a period that earns nothing.
Retail paradigms based on consumers buying repeatedly. Subscriptions are one version. Prompted reordering is another, and the prompts reach customers subscriptions can't convert.
Growing reluctance among shoppers to take on another recurring commitment. Shows up as weak subscribe-and-save uptake and early cancellation rather than as complaints.
Setup runs about three minutes. First predictions land inside a day.