Cascade's top-of-funnel was, by most standards, working. Paid and organic acquisition together were bringing in 40,000 sign-ups a month, and the marketing team had every reason to feel good about their numbers. The problem showed up two steps later: only 8% of those sign-ups ever reached what the team internally called "activation" — completing a first project and seeing a result.
Growth had been treating this as a funnel-optimisation problem — more A/B tests on the signup form, another onboarding email in the drip sequence. Eighteen months of that work had moved activation from 6% to 8%. Directionally right, but nowhere near enough to change the shape of the business, and nobody could say with confidence why the number was so low in the first place.
Finding the aha moment nobody had actually found
The team had a working assumption about what the "aha moment" was — the point in the product where a new user understood the value and became likely to stick around. It had never actually been tested against data. It was inherited folklore from an early user interview, repeated in enough onboarding decks that it had calcified into fact.
The first three weeks of the engagement were spent building a proper activation model: pulling event-level data for six months of sign-ups, and running a retention curve analysis against every meaningful action a new user could take in their first session. That analysis pointed to a different moment entirely than the one the team had been optimising toward — a specific combination of two actions, taken within the first ten minutes, that correlated with dramatically higher week-four retention. Users who hit that combination were retained at nearly four times the rate of users who didn't.
Twenty-two experiments, ten weeks
Once the real activation moment was identified, the work shifted from "improve the funnel" to "get more people to that specific moment, faster." That reframing changed what counted as a good idea. Some experiments were about removing friction before the moment — cutting an unnecessary setup step that was costing 12% of sign-ups before they ever got near the actions that mattered. Others were about making the moment itself more discoverable, since a meaningful share of users who could have reached it in their first session simply never found the path.
- Removed a mandatory setup step that had no bearing on activation but was quietly costing over a tenth of sign-ups before they reached anything meaningful.
- Redesigned the first-run experience to guide users toward the two actions that actually predicted retention, instead of a generic feature tour.
- Replaced the existing onboarding checklist — built around features, not outcomes — with a single guided path toward the real activation moment.
- Cut the average time from signup to first meaningful action from just under nineteen minutes to under six.
Not every experiment worked. About a third of the twenty-two moved the needle in the wrong direction or made no measurable difference — including one redesign the team was confident would be the big unlock, which turned out to add friction rather than remove it. That's a normal ratio for this kind of work, and it's exactly why the experiments were run as small, fast tests rather than one large redesign staked on a single bet.
Nobody at Cascade was wrong to focus on the funnel. They were focused on the wrong point in it. Once the data told us where the real value moment actually lived, most of the good ideas became obvious — the hard part was having the discipline to test the assumption in the first place.
The result
Over the ten-week engagement, activation climbed from 8% to just over 25% — a 220% relative increase. Because more of the same acquisition spend was now converting into genuinely retained users, blended customer acquisition cost dropped 35% without any change to the marketing budget or channel mix. Lifetime value rose 58%, driven by the much higher share of new users who now reached the moment that made them likely to stay.
+220%
Activation rate
-35%
CAC
+58%
LTV