RechargeSMS Onboarding

We didn't have a formal analytics funnel for onboarding. So I pulled the data myself, writing SQL queries against the database to piece together where merchants were dropping off. I could see what state each merchant was in: whether they'd completed onboarding, whether they'd ever enabled the tool, if it was currently on or off, and how many messages they'd sent. That gave me a rough but usable picture of the drop-off points.

The data told me where merchants were falling out, but not why. Support ticket analysis added another layer: merchants weren't confused by the product. They were hesitant.

Mapping the full onboarding journey was eye-opening as it revealed that the flow was too long, which amplified anxiety. And the UI gave merchants nothing back: no progress indicators, no success states, no signal that mistakes could be corrected.

💡 Onboarding wasn't a complexity problem. Rather, it was a confidence problem.

Shortening the flow would reduce exposure to anxiety, and adding feedback loops would give merchants the assurance to keep going. Both levers needed to work together.

With the confidence reframe in hand, I worked with product design to explore solutions. We had to be ruthless about scope and hone in on the essentials merchants needed to complete in order to send SMS.

We mapped every step in the existing flow against one filter: does the product break without this? That cut the list quickly.

• ✅ Store import stayed, but we rearchitected the timing. Instead of making merchants wait, we kicked off the Shopify product import in the background while they started onboarding. By the time they reached the import page, it was already done or in progress.
• ✂️ Swap Groups and other enhancement features got deferred. They make the product better and promote upsells, but they aren't required to launch.
• ✂️ A Shopify configuration step got removed entirely. It required merchants to leave the app mid-onboarding, go into their Shopify settings, and configure opt-in. It was one of the biggest drop-off points in the funnel, so we found a way to eliminate it from the flow.
• ✂️ Broken tooltips got scrapped instead of fixed. They were more difficult for engineering to retain, and fixing them meant investing in a pattern we didn't want to keep.

Because onboarding touches every part of the product experience, this wasn't a project I could run in a silo. I wrote the PRD, defined the user stories and acceptance criteria, and prioritized the tickets. But the real work was keeping five teams aligned: design, engineering, product marketing, support, and data analytics.

The data side was scrappy from the start (I'd already been pulling numbers directly from the database during discovery), and we started building out better instrumentation alongside the project.

I partnered closely with product design to iterate on solutions, worked with engineering to scope feasibility and extensibility, and looped in product marketing and support so the rollout wouldn't catch anyone off guard. Then I ran QA acceptance testing before launch.