Shaili Pankaj
Open to roles
← All work

5 · Lifecycle

Rebuilding the layer underneath, then the wins it made possible

Three months, one day of campaign freeze, and a measurement layer that could finally report revenue instead of opens.

Company
Doodle
Role
Chief Growth Officer
Dates
2023–2024, roughly three months for the migration
With
The CX and Lifecycle team, data, engineering

The problem

Doodle's Intercom workspace had years of accumulation in it. The event limit had capped out. Several events did the same job under different definitions. Nobody could tell what was genuinely firing. If you can't trust an event you can't trust a segment, and if you can't trust a segment you can't trust anything a campaign tells you. The team could report opens and clicks. That was the ceiling.

What we did

  • Rebuilt three layers rather than one: the events, the definitions, and the process for changing them.
  • Spent the first six weeks planning with data and engineering, with product and marketing as informed stakeholders.
  • Built and tested the core workflows in advance so they worked on switch day.
  • Handled the hard part deliberately: migrating customers who were part-way through legacy flows so they wouldn't restart at step one. The campaign freeze came to a single day.
  • Spent six weeks after the switch building the remaining flows and the continuous testing that made it a system rather than a migration.
  • Then went after the renewal boundary, which the new foundation had made addressable. A pre-renewal flow to customers we could now reliably identify as at risk, meaning auto-renew switched off or a dead card on file. And dunning sequences for involuntary churn, with payment-failure emails that explained what had actually gone wrong.

The call I had to make

The mid-flow customers were the whole risk. It would have been far simpler to cut over and let a few thousand people restart a sequence they were halfway through. We spent weeks planning that migration instead, because the alternative was making our migration into the customer's problem.

Impact

  • Measurement moved from opens and clicks to what a campaign did to the KPI it was meant to move, and then to revenue
  • One day of campaign freeze across a three-month migration
  • Pre-renewal flow: +2% on the renewal rate
  • Involuntary-churn recovery: ~1.5% of churned customers came back within ten days
  • Teams across the organisation could launch their own campaigns without working out how first

What it taught me

CRM spirals into something nobody can read without documentation, definitions and process. Get the foundation right or you can't tell what's working. That's the order I'd want to work in anywhere, because you can't run a serious lifecycle programme on data you don't trust.