The rule
The reason is ad spend. Tuesday’s ads bought Tuesday’s buyers; the only honest way to ask “did Tuesday’s spend pay for itself?” is to hold Tuesday’s spend against everything those buyers turn out to be worth — including the refund that arrives in three weeks. Booking money on the day it moved would flatter bad days and punish good ones.
What this does to recent days
Recent days change retroactively. A day gets better as its trials convert and worse as its refunds land. Practically:- A day is mature once its trial window has fully passed. Before that, low conversions and lots of pending trials are the norm, not a signal.
- A fresh day’s ROI is a floor, not a verdict — re-read it after the trial window.
- Refunds can drag a day down weeks later. Aggregate to weeks before judging a small funnel; a single refund can turn a thin day negative.
- Trials whose conversion charge is failing and being retried are counted as not converted from the first failure. Retry recoveries are a nice surprise later, never something Opti forecasts on.
What’s not in daily funnel revenue
Renewals. The second, third, fourth payment of a subscription belongs to the retention view (plan by plan, cycle by cycle), not to the day’s acquisition economics. Mixing them in would make acquisition ROI look better every month a customer stays — the number would stop meaning “did this day’s ads work”. Ask for either explicitly: “net revenue per day for this funnel, last 30 days” (acquisition) vs “what is a buyer on the 6-month plan worth so far?” (lifetime, incl. renewals).Vocabulary Opti uses
Meta backdates too — differently. Ads Manager books a purchase to the day of the ad click or impression that led to it, in your ad account’s timezone; Opti books it to the checkout day. Same idea, different clock — one reason the two dashboards disagree at day boundaries. See Facebook numbers vs real sales.