Retention is usually treated as the thing you do after acquisition works, and that sequencing is backwards. Repeat purchase rate and churn determine lifetime contribution, which determines your maximum acceptable CAC, which determines whether you can win auctions at all. A brand at 22% repeat purchase rate can outbid one at 8% by roughly 20% on identical products, and over months of auction pressure that compounds into market share.
The arithmetic is unusually favorable. Reducing monthly churn from 5% to 3% raises average customer lifespan from 20 months to 33: a 65% lifetime value increase with no change to acquisition whatsoever. Almost no acquisition optimization produces a 65% efficiency gain; retention work routinely does.
Most of the practical work is unglamorous. Deliverability determines whether your emails are seen at all, and there is no fast version of recovering a damaged sending reputation. Activation determines whether new customers ever reach the value that makes them stay. Neither is a campaign, and both are usually where the largest untapped gain sits.
What decides outcomes here
Build retention before scaling acquisition
Scaling into a leaky bucket doesn't just waste spend. It locks you out of the auctions where the bid ceiling matters.
Most early churn is failed activation
Not dissatisfaction. Segment churn by whether the user reached your activation event; if the split is sharp, it's an onboarding problem and a cheaper one to fix.
Sending to unengaged subscribers harms the engaged
Engagement is a placement signal. A segment that hasn't opened in a year isn't neutral. It degrades delivery for everyone else.
Segment LTV by acquisition source
Discount-acquired customers churn measurably faster. A blended figure leads you to overinvest in the channels producing your worst cohorts.