From seventeen tabs to one dashboard.
The spreadsheet that ran our business, the system that replaced it, and why the same pattern fits yours.
Every business has one: the spreadsheet that runs the place. Ours had FINAL in the filename, twice, and a tab called DO NOT DELETE that nobody could explain. This is the story of how we replaced it, because it is the same story we now sell.
Where we started
Harewood Digital runs a portfolio of web products. Each one generates money through Stripe or PayPal, costs money in hosting and services, raises support tickets, and needs its uptime and speed watching. Individually, every one of those has a perfectly good dashboard. Collectively, they had nothing, so the whole picture lived in a spreadsheet fed by hand.
The spreadsheet was not the problem. The feeding was. Exports on the first of the month, copy, paste, reconcile, notice a number that looks wrong, trace it, fix it, lose confidence in the rest. Call it a day of work, done grudgingly, twelve times a year.
What we built instead
One dashboard, one screen, fed by the systems themselves:
- Revenue pulled from Stripe and PayPal daily and reconciled automatically, so a missing payment surfaces in hours rather than at month end
- Costs prefilled from the recurring bills, with one-off invoices dropped in as they arrive
- Site health checked every five minutes, with page speed tracked against competitors
- Support read by AI, which drafts a reply for a human to approve and flags anything unusual
- The monthly summary drafted by AI from the numbers above, then given a final human read
What actually changed
The honest answer is not "we saved a day a month", although we did. It is that decisions stopped waiting for the spreadsheet. Which product deserves the next hour of attention, whether an ad campaign is paying for itself, whether a cost has quietly crept: those questions now take a glance, not an afternoon of assembly first.
The part that applies to you
Nothing in this build is exotic. Every piece is a pattern: a system with an API on one side, a place the numbers should appear on the other, and AI doing the reading, matching and drafting in between. Your version might connect Xero and your job sheets rather than Stripe and support tickets, but the shape is identical.
That is the point of a pilot: pick the workflow that costs you the most hours, build the pattern, measure it, and only then decide whether to go further.
Tell us about your spreadsheet →
More from the blog: five jobs you should not give to AI, and how to stop retyping between Shopify and Xero.