Increase trial → paid conversion via display

HYPOTHESIS

Step one, deliberately blunt: if we surround a trialist with our ads during their trial weeks, do more of them stick and convert to paid? We're not yet asking which channel does the work, YouTube, Meta or Display, or what the perfect frequency is. First we prove exposure moves stickiness at all. If it does, the follow-up tests isolate the channel and tune the cost. If it doesn't, we kill the spend and bank €360k a year. Either answer is a win.

// split at the edge, one hash, no cookies needed
if (murmur3(trial.id) % 2 === 0) {  // 50% → ARM A
  fbq('init', '742900481...');          // Meta pixel
  fbq('track', 'StartTrial');
  gtag('event', 'trial_start', {      // Google tag
    send_to: 'AW-8823104/trial',
    user_data: { email: sha256(email) }
  });
  audiences.add('trial-assurance-30d');
}

else {  // 50% → ARM B · HOLDOUT

  // nothing fires.
  // no pixels, no audiences, no ads.

  // this silence is the control group.

}

91% probability the lift is real · 3,120 trialists enrolled · read at week 6 or ≥95%, whichever first. Still running. If it confirms: next tests isolate the channel (YouTube vs Meta vs Display) and walk CPMs down. If it nulls: the spend dies and €360k/yr is freed.

The profitability line is simple: the lift is worth +€7.1k of new MRR every month, so if the campaign runs for less than that, it's in profit. As a working example we'd budget ~€3k/month for the exposure layer: comfortably under the line on month-one revenue alone, before counting the months those customers stay. And this is the starting point, not the ceiling. Frequency, creative and channel mix all get optimised from here. Illustrative numbers, real mechanism.