Replenishment reminders
before they run out, not after
Consumables sell again on a rhythm. Purands learns each customer's rhythm per product and reminds them a few days before the last one is gone.
The problem
A customer finishes a jar or a bag and buys whatever is nearest. The reorder is lost to convenience, not to preference.
A monthly newsletter cannot catch that moment. Every customer runs out on a different day, and every product lasts a different length of time.
What triggers it, and what it reads
The trigger is a predicted run-out date, modelled per customer and per product rather than as one blanket interval.
- How long each product has lasted this customer before
- Order frequency and quantity per SKU
- Life stage and context, where the category has one
- Purchase intent and churn risk, scored daily
- Stock and catalogue state, so nothing unavailable is suggested
Brands running this skill report repurchase moving from 12% to 35%, and per-SKU reminders lifting click-through from 6% to 29%.
What gets sent
Examples only, written for this page. Real messages are generated per customer from their own profile and your catalogue.
- WhatsApp: Hi Ravi, Buddy's Royal Canin Medium Adult bag is almost finished. Based on his feeding schedule you will run out in 5 days.
- Email: Names the product, and offers a one-tap reorder of the exact size bought last time.
- SMS: Running low on your Niacinamide 10% serum? Reorder in one tap: [link]
- Push: Timed to the predicted run-out date and deep-linked to that product.
- Web assistant: Recognises a returning customer and offers the repeat order before they search for it.
How the agent picks the moment and the channel
Consumption modelling learns how long each product lasts for each customer, so a heavy user and a light user are reminded on different days.
The reminder lands before the run-out, not on it. A skincare customer on a 60-day serum is nudged around day 50.
Channel is chosen from where that customer replies, and the send lands in their peak engagement window.
If the customer has already bought again, or has heard from you too often this week, the reminder is not sent.
Setting it up
Connect the store and let order history sync. The model needs repeat purchases to learn from, so accuracy improves over the first cycles.
Per-SKU reminders need a clean catalogue: sizes and variants as separate products, so a 3kg bag is not modelled as a 1kg one.
The other retention jobs
- Abandoned cart recovery: one decision per shopper, not one schedule.
- Win-back campaigns: a reason to return, not a blanket discount.
- Back-in-stock alerts: the moment it lands, to the people still waiting.
- Review requests: asked at the moment the answer is yes.
- Order updates: the messages customers actually open.
All six run on one engine: Agentic Skills, reading Profiling.
Replenishment reminders
See this running on your own data
A 30 minute demo on your catalogue and your customers, with the decisions the agent would make shown one by one.