Where they started
SaaS Analytics collected plenty of data and had no good way to show it. Their product exposed a solid API; the customer-facing layer was an analyst, a spreadsheet template, and a fortnightly deadline.
Omar, their founder, put the cost at about ten hours a week across the team. The bigger issue was consistency. Three analysts meant three interpretations of the same chart, and clients noticed when a metric they'd been tracking quietly changed shape between reports.
They'd looked at off-the-shelf BI tools and stalled on the same thing every time: the seat pricing made client access unaffordable, and the branding was someone else's.
What we built
We spent the first week with the analysts rather than the founder, watching them assemble a real report. That's where the requirements were. Half the manual work turned out to be reconciling two data sources that disagreed on timezone handling — a bug, not a feature request.
The dashboard is a Next.js app talking to their existing API, with a caching layer so a fifteen-panel view doesn't fire fifteen cold queries. Charts are Recharts with a custom theme; we tried a heavier viz library first and dropped it when the bundle cost outweighed what it added.
Saved views were the feature that made it stick. An analyst configures a view once per client, and the client gets a read-only link with their own logo on it. Scheduled exports go out as PDF and CSV on whatever cadence each account wants, which retired the fortnightly ritual entirely.
