Drop a client’s order history in. Affinsy mines what sells together, ranks test hypotheses by expected revenue, and hands you the audience for each one: Klaviyo, CSV, API, or MCP.
Treat Affinsy like a coworker on the account: hand over a client’s orders, delegate the mining, review the filed work. The analysis an analyst would bill weeks for is back in minutes.
CSV upload or read-only API key. No app install on the client's store, no dev review, no IT ticket. You can have a dataset flowing before the kickoff call ends.
It mines the orders and comes back in minutes with ranked test hypotheses: expected lift, audience size, confidence band. Ask a follow-up in plain language and the assistant runs the next analysis and builds the audience.
Every run is filed into the client's spaces, so the work never leaves with whoever ran it. Re-run before every call with the same settings, open the analysis from two months ago in one click, and the next QBR is already half-written.
Every workspace comes with a Merchandising Engineer, an assistant that knows the client’s orders, products and analyses. It answers from the data, starts a new analysis when the question needs one, and builds the customer lists to act on.
Every number comes from the workspace's own orders and completed analyses. If something isn't in the data, it says so. It doesn't guess.
Ask a question that needs new work and it starts the Market Basket or RFM run itself, files it into the right space, and links the result when it's ready.
Turn a cross-sell rule into the exact customers to contact — sized, filtered by segment, and downloadable as a CSV for your email tool.
The Mine, Hypothesize, and Activate panels above are the interactive surface. Underneath, every run mines your order history into rules, segments, movers, and product diagnostics, and files the result in its space. When the client asks what changed since last quarter, that quarter’s run is still there, exactly as it was mined: open it next to the new one or drop it in the deck. Nothing to re-run, nothing to reconstruct.
One bridge from captured revenue to what's still on the table: the share flowing through co-purchase patterns, cross-sell winnable right now, and replenishment coming due. Every dollar of upside is sized on customers you can actually reach.
Open on the live report
Every run also grades the customer base with RFM: Champions to Hibernating, revenue mapped per segment, and a ranked share table ready for the deck. Win-back and VIP audiences export exactly like basket audiences.
Rules whose lift shifted by 10%+ versus the previous 30-day window, split into risers and fallers. Feature the rising ones before the trend matures, investigate the falling ones before retiring them — the temporal signal you don't get from a single-snapshot report.
Open on the live report
Every catalog item sorted by orders, with the number of rules it appears in, what it substitutes for, and a week-by-week trend sparkline. Useful for the moment a client asks why their best-seller stopped lifting the basket.
Open on the live reportInvite analysts and account managers with their own logins. They open the same dataset, the same MBA and RFM reports and the same spaces. The AI connection is wired once, for the workspace, so a client engagement survives any single laptop.
How CRO and retention agencies (and the merchants we work with) typically use Affinsy. Quick, honest answers.
Still have a question? Email usDrop a client’s order history in. Affinsy returns the rules, the hypotheses, and the audience for each one.
7-day free trial · Refreshed in under two minutes
For CRO and retention agencies delivering across up to 10 client stores.
7-day free trial · Cancel anytime
See a sample auditMore than 10 client stores or higher data volumes? Tell us about your book, we size larger plans case by case.
