Fewer errors, fewer tickets, more people using what they'd paid for
Errors, ticket volume, and feature adoption all moved in the same quarter — evidence that the redesign fixed the actual bottleneck, not just its symptoms.
Finding the real problem before designing the fix
Support tickets don't explain themselves. Before touching a single screen, I needed to know whether Engagespot had a UI problem or something the interface alone couldn't fix — and whether solving it was worth the engineering time it would take.
A focused generative research sprint — workshops and interviews with internal teams and power users, mapped against how they actually set up, sent, and monitored notifications.
To surface real pain points and their cost in time, errors, and drop-offs — not a wishlist of feature requests.
Stakeholder interviews across developers, PM, and support; a six-competitor scan; and a full pass through a quarter of support tickets and error logs.
2–3 engineers (backend and frontend), 1–2 support and CS reps, and 2 customer developers already integrated with Engagespot.
I ran discovery workshops with product, engineering, and support, then sat down individually with the customers most likely to churn. In parallel, I read every ticket tagged setup or integration from the previous quarter. The pattern showed up fast: friction wasn't spread evenly across the product — it clustered almost entirely in the first ten minutes after signup.
From developers
- No clear "start here" moment after signup
- Setup broke most often at the API key / HMAC step
- No way to see, at a glance, which channels were live or where to add a provider
- Templates were repetitive and demanded more dev time than they should have
From product & support
- Setup, channels, and templates were the top three drivers of support tickets
- PMs and CS staff found the interface too technical to self-serve
- No single place to see what was sent, what failed, and why
- Friction was UX and information architecture, not missing functionality — with real appetite for guided onboarding
I benchmarked Engagespot against six notification infrastructure products — SuprSend, Novu, Knock, Courier, Raven, and MagicBell — across 15 dimensions, from onboarding and accessibility to dashboard customization and documentation quality.
- Dark mode, theming, and multi-channel support had become baseline expectations, not differentiators
- Infrastructure-heavy tools optimized for routing and workflows; interface-heavy tools optimized for inbox UX — almost none did both well
- Most competitors read as either enterprise-heavy or too technical for a non-developer to self-serve — Engagespot's opening
Dev John
PM Priya
Sole UX designer, embedded end to end
The only designer on a team of founders, engineers, a PM, and support. My scope covered:
- Discovery — ticket analysis, interviews, workshops
- Design — IA, flows, UI, design system
- Validation — usability testing, iteration, instrumentation
What I was designing around
- An engineer-led UI with inconsistent patterns — a full rebuild wasn't an option on this timeline
- Discoverability, not missing functionality, was driving most of the support load
- Competitors already shipped dark mode and multi-channel support — the redesign had to clear that bar, not just fix onboarding
Three calls that shaped the redesign
Break setup into a guided four-step flow with a single primary action at each step — fewer choices, faster decisions — replacing one dense settings page.
Regroup channels into "Configured" and "Not added," matching how users already sorted them mentally, instead of one undifferentiated list.
Prioritize setup and discoverability over new features. The research pointed to friction, not a capability gap.
How the redesign came together
Three phases, in the order the project actually unfolded — click through to follow the arc from research to a validated design.
Ticket data, error logs, and interviews across developers, PMs, and support reps all pointed the same direction: setup broke most often at the API key/HMAC step, there was no clear "start here" after signup, and templates ate up more dev time than they should have. A six-competitor scan confirmed the bar Engagespot had to clear — dark mode, theming, and multi-channel support were no longer differentiators, they were table stakes.
Replace the wall-of-text settings page with a guided flow that always surfaces one obvious next action.
Rebuilt Getting Started as a four-step flow, and regrouped Channels into Configured and Not added for at-a-glance health.
Usability testing validated the direction — participants called out the step-by-step flow and the Configured/Not added grouping as clear wins, and described the darker UI as modern and comfortable to work in for long sessions. What didn't land, like the template-canvas interaction, fed straight into the next iteration.
What this project changed about how I work
The fix was never a feature gap — it was discoverability and sequencing. The biggest single lever turned out to be reducing how many decisions a user had to make at each step.
Ticket and error-log data made a far stronger case for prioritization than opinions in the room — worth building that habit into the start of every project, not just this one.
A design system pays off twice: once in this redesign's consistency, and again in how much faster the parallel Design System project was able to ship afterward.