Day-0 Lies. Day-7 Decides.
Day-0 conversion is loud.
Day-7 retention is honest.
I saw this in a B2C health funnel with paid traffic and weekly cohort readouts. Day-0 looked strong. Day-7 told the truth.
I track quality with one chain across weekly windows.
click -> onboard_finish -> purchase_initiated -> purchase_completed
Then I validate post-purchase quality before calling any change a win.
Day-0 Is Easy to Inflate
Aggressive offers can lift first payment.
Urgency copy can lift first payment.
Friction cuts can lift first payment.
None of that guarantees user fit.
In health funnels, wrong-fit users churn fast.
You see it one week later.
Health and fitness has brutal retention by default.
People buy on the emotion of a new start.
Same pattern as a one-year gym membership with one week of visits.
The funnel sees hope on Day-0. The product sees reality by Day-7.
We push hard to retain: emails, reminders, bonuses, urgency loops.
Most still drop.
The users who stay are the ones getting real value.
Another failure mode is payments, not product.
Recurring charges fail more often than teams expect.
Decline rate is quite high.
For banks, this category is high risk merchant traffic.
The spike is worse on ramp-up pricing, for example from $5 to $50.
Some users do not churn by choice. The bank declines the renewal.
Weekly Plans Break First
The fastest failure signal in this funnel is weekly subscriptions.
Users take the weekly plan, then cancel or refund almost immediately.
That pattern shows up before longer plans look weak.
When weekly churn spikes, acquisition quality is already off.
The Check I Use
I read acquisition by cohort.
- day_0_purchase_completed_rate
- day_1_activation_depth
- day_7_active_rate
- day_7_refund_rate
- day_7_cancel_rate
If Day-0 rises while Day-7 falls, quality degraded.
No debate needed.
Weekly Decision Rule
I do not ship by Day-0 alone.
I use a delayed gate.
- keep the change if day_0_purchase_completed_rate is up and day_7_active_rate holds
- kill the change if day_0_purchase_completed_rate is up and day_7_active_rate drops
- kill the change if refund or cancel pressure rises out of range
This removes mood-driven decisions.
A Simple Rule
Never ship acquisition changes with Day-0 alone.
Ship with a delayed verdict window.
For this funnel, that window is seven days.
Speed matters. False speed hurts more.
What Changed in Practice
We moved decision cadence from daily wins to weekly truth.
Short-term excitement dropped.
Bad experiments died faster.
Good experiments survived longer.
Trade-Off
Your feedback loop feels slower.
Your mistakes get cheaper.
I take cheaper mistakes every week.
Takeaway
Day-0 tells who clicked buy.
Day-7 tells who stays.
Use the second one to judge acquisition quality.