Which Creator and Affiliate Channels Bring You the Best Customers
To find which creator and affiliate channels bring the best customers, build a cohort table that tracks retention and revenue by acquisition source over time. A partner whose audience buys once and churns is worth far less than one whose audience stays and expands, even if both look identical on a first-purchase report.
Setting Up the Cohort Table
Group customers by the month they signed up and by the partner or channel that referred them, then track retention and revenue for each group over the following months. The table has months down one side and acquisition source across the top, with each cell showing either a retention percentage or a cumulative revenue figure for that cohort at that point in its life. This takes a working attribution setup, tagging every signup with its source at the point of acquisition, since reconstructing it later from memory or a general marketing report will not hold up.
A Worked Example of What the Table Reveals
Imagine two partners each sent you a hundred new customers in the same month. Partner A's cohort has forty of those hundred still active six months later, each spending close to their original plan amount. Partner B's cohort has seventy still active at six months, but their average spend has dropped because most downgraded to a cheaper plan shortly after signing up. On a first-purchase report the two partners looked similar. On the cohort table, Partner A's smaller, more retained group is worth more to the business over time than Partner B's larger but downgrading one.
Why First-Purchase Revenue Alone Hides This Difference
A report that only shows revenue at the moment of signup treats every new customer as equally valuable, which is the exact assumption a cohort table exists to challenge. Two partners can generate identical first-month revenue and represent very different long-term value once churn and expansion or downgrade patterns show up over the following months. If partner payouts or priority are still being decided on first-purchase numbers alone, the program is very likely over-rewarding the wrong partners without anyone noticing until the annual numbers come in soft.
How Far Out You Need to Track Before Trusting the Comparison
Track long enough to capture at least one full renewal or repurchase cycle for your business, since a cohort that looks strong at one month can still collapse by month three once an initial promotional period ends. Comparing partners on a cohort window that is too short just moves the same first-purchase blind spot a little further down the timeline instead of actually fixing it. Once you have a full cycle of data, revisit partner tiers and terms using the cohort numbers rather than the original signup report.
A useful rule of thumb: treat a weak cohort as a pattern only when it repeats. One weak cohort can come from a promotion, a seasonal spike or a single viral post that attracted casual buyers. A partner whose cohorts retain worse than your program average across two or three periods in a row points to something structural about the audience or the promise the partner is making. Adjust terms only after you see that repeat, and note the reason in the partner record so the next review starts from what you already learned instead of from scratch.
Turning the Table Into a Decision, Not Just a Report
A cohort table that nobody acts on is just a more detailed version of the same problem. Once the pattern is clear, adjust commission structure, priority support or co-marketing budget toward the partners whose cohorts actually hold up over time, and have a direct conversation with partners whose traffic looks strong short term but weak on retention, since that conversation might surface a fixable issue, like a mismatched audience expectation the partner's content is setting up front.
The full process, in order:
- Tag every signup with its source at the point of acquisition, so cohorts can be built without reconstructing history later.
- Group customers by signup month and referring partner, and track retention and revenue for each group over the following months.
- Wait for at least one full renewal or repurchase cycle before comparing partners against each other.
- Move commission, priority support or co-marketing budget toward partners whose cohorts hold up over time.
- Talk to partners with strong short-term but weak retention, since a mismatched audience expectation may be fixable.
Watching for a Channel-Level Pattern, Not Just a Partner-Level One
Sometimes the pattern is not about one specific partner being weaker, but about an entire channel type consistently under-delivering on long-term value, for example, coupon-site traffic converting at a decent rate but retaining poorly across nearly every partner in that category. If several partners of the same type show the same cohort weakness, the fix is a channel-level decision, adjusting how that whole category is compensated or targeted, rather than trying to solve it one partner conversation at a time.
What Good Looks Like
A working cohort view tags every customer with their acquisition source at signup, tracks retention and revenue by cohort over at least one full renewal cycle, and feeds that back into partner terms.
Building The Capability (5-Stage Skill Ladder)
How to Get Started
Frequently Asked Questions
How many months of cohort data do I need before trusting a comparison between partners?
At least one full renewal or repurchase cycle for your business, which varies by product. A subscription business needs several months to see past an initial promotional period; a one-time purchase business needs enough time to see whether referred customers come back for a second order.
What if two partners send similar volume but very different cohort quality?
Treat them differently in your program, even if their raw signup numbers look the same. Better terms, priority support or a bigger co-marketing budget should follow the partner whose cohort actually retains and expands, not the one with the bigger initial number.
Can a partner's cohort quality change over time?
Yes, especially if their content or audience shifts. Recheck cohort quality on a regular schedule rather than assuming a partner who sent great customers a year ago still does, since their audience or content approach may have changed since then.
About the numbers
This guide doesn't quote a sourced benchmark. Figures in it are estimates or general guidance, so check them against your own numbers.
Related Guides
Choosing the Right Cohorts for LTV and Expansion Analysis
A decision guide for picking cohort definitions when analyzing customer lifetime value and expansion revenue, and the mistakes that make the numbers misleading.
The Clauses That Actually Matter in a Creator Contract
A checklist of the contract clauses that most often cause problems in creator and influencer partnerships, and the mistakes to avoid when negotiating them.
Reading Cohort LTV Data Without Fooling Yourself
How to build cohorts that actually reveal expansion opportunity, avoid the most common cohort analysis mistakes, and act on what the data shows.
What a Good Affiliate Landing Page Conversion Rate Looks Like
How to set a realistic conversion rate benchmark for creator and affiliate landing pages, and where to look first when a page underperforms.
Building a Cohort Table That Actually Shows Where LTV Comes From
A step-by-step walkthrough of building a customer cohort table by hand, so you can see which signup groups expand and which quietly shrink.
A Runbook for Growing LTV Cohort by Cohort
A step-by-step way to group customers into cohorts by signup period and run targeted expansion plays against each one instead of one blanket campaign.