Lifecycle / Behavioral Systems / Fintech
Building the behavioral trigger layer
The activation flow moved every user through the same time-based sequence. The real question was whether it could respond to what each user actually did, and convert the ones who stalled after the hardest step.
The read
Intelligent Investing launched as a unified app, but new signups got no structured communication after creating an account. The first version of the activation Canvas fixed that with a time-based sequence. Day 0, Day 2, Day 7, and so on.
But a time-based flow has a blind spot. It treats every user the same regardless of behavior. The data showed the real gap. Users who crossed the hardest barrier, connecting their bank, often stalled before the action that actually generated revenue. For self-directed users that was the first trade. These were funded-idle users. They trusted the platform enough to link a bank, then went quiet. A scheduled email on Day 10 does nothing for someone whose behavior already changed on Day 4.
The decision
Build a behavioral layer on top of the time-based Canvas. The system needed to detect a specific behavioral state, bank connected but no first trade, and respond to it directly with a targeted push at the moment the user stalled, not on a fixed calendar day.
This split activation into two distinct problems. Trust activation, getting the bank connected, and product activation, getting the first transaction. The behavioral trigger owned the second one.
The metric I chose
Funded-idle to first-trade conversion rate. This was the cleanest possible test of whether behavioral triggering worked, because the baseline was zero. There was no mechanism for converting funded-idle users before this. Any movement was attributable directly to the trigger layer, not to other variables in the funnel.
I chose it over a blended activation number on purpose. A blended number would have hidden whether the behavioral layer specifically did anything. Isolating funded-idle conversion made the system prove itself.
The build
Built the behavioral branch inside the Braze Canvas using product events, not time delays. Wallet funded and first-trade-placed events defined entry and exit. Added a first-trade nudge push triggered by the funded-idle state, deep-linked straight to the trading view. Set suppression so a user exited the moment the first trade fired, and capped activation pushes at one per week so the behavioral layer never turned into noise.
Partnered with the product team to confirm the event instrumentation was firing correctly, since the whole system depended on accurate behavioral signals.
0% → ~19%
Funded-idle to first-trade, 2-month
~22%
First-trade nudge interaction rate
Funded-idle to first-trade conversion went from 0% to ~19% within two months of launch, and reached ~24% by the end-of-year read. The first-trade nudge push held a ~22% interaction rate. The behavioral layer proved that responding to what users did, rather than how long they had been in the funnel, was what moved the revenue behavior. It also validated the next step: branching the whole Canvas by product intent.
Tech
Braze, multi-channel Canvas architecture
Product event instrumentation, behavioral triggers
Push deep-linking, suppression and frequency logic
Miguel N. Monzones
Vancouver, BC, Canada
