Every retention tool reports the revenue it “drove.” What was actually observed is smaller than that word: a customer was the subject of a decision, and some time later she placed an order. The first happened before the second. Whether it produced the second is a different claim.
This matters most for exactly the customers retention work is aimed at. Someone who buys the same serum every six weeks was likely to buy it again. A reminder sent in week five is followed by a purchase almost regardless of what it said.
Three rules that keep the number honest
Altvary records an order against a decision under three rules.
- A window. The order has to be placed within 30 days of the decision being handed off. Without a window, every future purchase is a win for something.
- One order, one decision. An order credits at most one decision, the most recent. Otherwise the same sale is counted by every play that touched the customer.
- Net of refunds. What is credited is the order less what was refunded. A returned order did not earn what its total says.
The result is called influenced revenue, and the name is chosen on purpose. It says a purchase followed. It does not say the decision caused it.
Count what didn’t happen too
A ledger that only shows the wins is a brochure. Next to each decision that was followed by a purchase there are decisions still inside their window, and decisions whose window closed with no order. Seeing all three is the difference between measuring and promising.
What would it take to say “caused”?
A comparison. The only way to know what a message added is to know what happens without it, among customers who are otherwise alike. Until you have that, the careful reading of any retention figure, from any tool, is “these purchases followed” and not “these purchases were created.”
It is a less exciting number. It is also one you can still believe a year from now.