Benchmarking push opt-in without vanity
An opt-in spike after a celebrity campaign is a weather report. Treating it as your durable consent baseline will punish the next quiet month and train the team to chase another famous face.
Push permission is not engagement. It is a gate. Once the gate is open, frequency, hour, and content decide whether the user still wants the gate open next month. We have seen Thai consumer apps report “72% opt-in” after a campaign, then quietly lose a fifth of those permissions over eight weeks as evening blasts piled up.
In Push Timing & Silent Hours we ask students to plot opt-in remaining, not opt-in acquired. The remaining curve is the one that belongs next to a Mobile Engagement Benchmarking table. Acquired is a campaign KPI. Remaining is a product policy outcome.
Silent hours are local
Bangkok commute peaks and dinner hours are not the same as a US “do not disturb after 21:00” default. A 21:40 blast during a BTS delay can look clever in a vendor case study and feel like harassment on the platform. We mark send calendars against local motion: office dismissal, dinner, late-night transit. If your tool only ships US quiet hours, you still have to write the local policy yourself.
A tap is not a contract
Consent that was collected under a one-time incentive should decay faster in your model than consent collected during a calm onboarding. That sounds obvious until a dashboard averages them. Split the cohorts. Publish the limitation: “opt-in after incentive is not comparable to organic permission.” If legal asks for a cleaner story, give them the split, not a blended vanity rate.
Frequency caps belong in product policy, not in a vendor default you never read. Engagement Benchmark Studio only sketches this in Module 05; the two-week workshop is where the calendar actually gets written.