Also known as: Retention Analysis, Cohort Reporting
Cohort Analysis is the practice of splitting referred users into groups (cohorts) that share a start event—usually the month they registered or first deposited—and tracking each group's behaviour, trading volume, and retention over time. It answers not "how many traders do I have?" but "how do the traders I acquired in a given period behave as they age?"
Instead of one blurred average, a cohort table lines up groups side by side against a common clock (month 1, month 2, month 3…). Because every cohort is measured from its own registration date rather than the calendar, you can compare January's traders at their month-3 mark against March's traders at their month-3 mark on equal footing. That isolates the effect of a change—a new landing page, a different traffic source, a bonus tweak—from ordinary seasonal noise.
For an IB earning lifetime revenue share, cohorts expose the shape of the money. Say your January cohort of 100 funded traders still shows 42 active in month 3, while March shows 61 active in month 3 after you added an onboarding email sequence—a 45% lift in month-3 retention. Trading volume follows the same curve: if the January cohort averaged $18 of spread revenue per trader in month 3 and March averaged $27, the difference compounds across the remaining lifetime of every trader in that group.
Cohort analysis is descriptive, not predictive of any individual account—it shows historical patterns you can act on, not a promise of future returns. Its power is diagnostic: it tells you when a group tends to go quiet, which acquisition channel produces durable traders versus one-and-done depositors, and whether a recent optimisation actually moved retention or just moved the headline number.
You assign each trader a cohort key—most often the year-month of registration or first funding—then measure a chosen metric (active accounts, lots traded, spread revenue) at each period offset from that key. Period 0 is the acquisition month; period 1 is one month later, and so on.
The result is a triangular table: rows are cohorts, columns are age offsets, and each cell is the metric for that cohort at that age. Reading down a column compares different cohorts at the same age; reading across a row shows a single cohort decaying (or holding) as it matures. Retention cohorts express cells as a percentage of period 0; revenue cohorts express them in USD per trader or as cumulative revenue toward LTV.
Because offsets are relative, the newest cohorts have fewer columns of data—March cannot yet have a month-6 number in April. Analysts read the mature diagonal for stable patterns and treat the newest cohorts as early signals only.
Pick the anchor event and grain—commonly the month of first funded deposit. Choose the metric you care about: active traders, lots, or net spread/commission revenue.
Tag each referred client with their cohort key and record their activity by calendar period so it can be re-expressed as an age offset from the anchor.
Recompute activity relative to each cohort's own start (month 0, 1, 2…) so groups acquired in different months line up against a shared clock.
Compare cohorts at the same age (down a column) to judge acquisition quality; follow one cohort across its row to see where retention or revenue drops off.
Where the curve bends—say activity halves between month 2 and month 3—schedule an intervention (education, a reload offer, a check-in) one period before it.
Why it matters for partnership: Cohort analysis reveals how long referred traders stay active, so IBs on lifetime RevShare can see which traffic sources produce durable clients and roughly when a group starts to churn—letting you time retention offers and defend real LTV.
An IB sending traffic to IC Markets groups funded clients by month. The January cohort held 40% of accounts active by month 3; after adding a two-week onboarding email series, the March cohort held 58% active at month 3 and averaged 30% more standard lots. On roughly $3.50 per-lot rebate, that retention lift materially raised the cumulative revenue the IB expects across each March trader's lifetime.
| Aspect | Cohort analysis | Aggregate reporting |
|---|---|---|
| Time base | Age from each group's signup | Calendar total for all users |
| Reveals | Retention curve and churn timing | Headline totals only |
| Blind spot it fixes | Old clients churning behind new signups | — (it creates that blind spot) |
| Best question | Which source ages well? | How big is it right now? |
Find the exact month your cohorts tend to go quiet, then schedule an educational touch or reload offer one month earlier—intervene before the drop-off, not after.
Watching only aggregate trading volume hides the truth that your older cohorts are churning fast while fresh signups mask the decline—so revenue looks stable right up until acquisition slows and it collapses.
A segment groups users by a shared attribute at any time (e.g. all UK traders). A cohort is anchored to a shared start event and tracked forward from it, which is what lets you measure retention over time.
Small cohorts swing wildly on a few accounts. Aim for at least 30–50 funded traders per cohort before reading percentages as signal rather than noise.
First funded deposit is usually more useful for IB revenue work, because unfunded registrations never generate spread or commission and can dilute the retention picture.
Many broker/CRM back offices expose per-client registration and volume data you can export and pivot into cohorts yourself, even without a dedicated analytics tool.
No. It describes historical patterns to guide decisions; it is not a forecast of any individual account and never a guarantee of future revenue.
Often they are not worse—they simply have fewer months of data, so their curve is incomplete. Compare cohorts only at ages where all of them have matured.
Trading educators leave money on the table with flat CPA deals. Here's how revenue share, hybrid models, and sub-IB structures pay for the retention your curriculum …