New & activating
Weekly Growth review · ISO week undefined · undefined undefined · undefined · Generated undefined
Activation gets a new reseller to their first order. This page asks the next question: do we turn that into a habit? A first order is a trial; the win is the second, then the routine. Here we track newly-activated resellers from first order through to established regular — where they stall, how fast they come back, whether newer intakes are sticking, and what the subsidy we spend to habitualise them is actually buying. Base treats the inflow as one cohort; the channel split lives on Activation.
Fact · Cause · Action
AI-generated — stub. Activation is doing its job; turning first orders into a habit isn't. Only a third of newly-activated resellers become established regulars, a third never place a second order at all, and month-1 retention has slid 10 points across recent cohorts. The first order is won; the first month is being lost.
| Section | Fact | Cause | Proposed action |
|---|---|---|---|
| New cohort progression | Of 9 newly-activated resellers, 6 place a 2nd order, 4 a 3rd, and only 3 become established regulars — a third are one-and-done. | Nothing prompts or rewards the second order; a first order is often a trial that never gets followed up. | Add a post-first-order onboarding sequence and a 2nd-order incentive; instrument why one-and-done resellers drop out. |
| Month-1 retention | M1 retention has slid from 58% (Apr cohort) to 48% (Jun cohort) — newer new-customers stick less. | Recent intake is lower-intent — the paid activation dip on the Activation page is diluting cohort quality downstream. | Pair with the paid-activation fix; add a first-month check-in for at-risk new resellers before they lapse. |
| Activation → 2nd order velocity | The gap from first to second order has crept from 15 to 19 days, over the 14-day target. | No reorder prompt — resellers drift between orders with nothing pulling them back. | Trigger a reorder prompt around day 10; measure whether it pulls the gap back under target. |
| Subsidy effectiveness | Subsidy per new reseller is up 66% since April (£78 → £130), yet the 2nd-order rate has fallen 8pp (66% → 58%) — the cash lever is losing potency. | Diminishing returns on discount depth; the barrier to a second order looks like attention, not price — we may be paying for orders we'd get with a reminder. | Freeze subsidy depth and test the comms lever (a timed day-10 reminder) head-to-head before discounting harder. |
New cohort progression
This week. The yield is thin: 9 activated → 6 second orders → 4 third → just 3 established regulars. A third never come back after their first order, and only a third of the cohort turns into the routine ordering we actually want. Activation is the start line, not the finish. · AI stub
Catalog Error: Table with name base_new_progression does not exist!
Did you mean "pg_catalog.pg_constraint"?
LINE 2: ... SELECT * FROM (select stage, count, ln(count) as size from demo.base_new_progression order by stage_order
^Activation → 2nd order velocity
This week. The second order is taking longer to arrive — up from 15 to 19 days over the quarter, now past the 14-day target. · AI stub
Catalog Error: Table with name base_new_kpi does not exist!
Did you mean "pg_catalog.pg_namespace"?
LINE 11: from demo.base_new_kpi order by week_start
^Subsidy ladder — the cost to habitualise
This week. We front-load the discount and wean it off: 30% on the first order, 20% on the second, 10% on the third, full price thereafter. That's £120 of subsidy per reseller across the first three orders — on top of CAC — to convert a trial into a habit. · AI stub
Catalog Error: Table with name base_new_subsidy does not exist!
Did you mean "main.pragma_database_list"?
LINE 3: from demo.base_new_subsidy order by step_order
^Is the subsidy lever working?
This week. The two are pulling apart the wrong way. We're spending 66% more subsidy per new reseller than in April, and getting a lower 2nd-order rate for it (66% → 58%). The cash lever has stopped buying conversion. · AI stub
Catalog Error: Table with name base_new_subsidy_trend does not exist!
Did you mean "main.sqlite_master"?
LINE 5: from demo.base_new_subsidy_trend order by week_start
^Month-1 retention by cohort
This week. Newer cohorts are decaying faster. Month-1 retention has fallen from 58% (Apr) to 54% (May) to 48% (Jun) — each intake holds a little less well, and all three sit below the 60% target. This is the same story as progression, seen over time: we're activating resellers we're then not keeping. · AI stub
Catalog Error: Table with name base_new_cohorts does not exist!
Did you mean "information_schema.table_constraints"?
LINE 2: ...RIBE SELECT * FROM (select cohort, week_num, retained from demo.base_new_cohorts order by cohort, week_num
^