The most important screen in your company is a spreadsheet

TL;DR — Off-the-shelf tools are built for the average of everyone. That is fine until one process is core to how your team works — then the generic tool becomes a tax: extra clicks, workarounds, and a spreadsheet holding the process together. What that tax costs when you count it, and what a purpose-built interface changes.

Tuesday, 08:55

The operations lead at a 16-person B2B services company starts every day the same way: she needs one number — how many pending partner reviews are older than 48 hours. The requests arrive in a ticketing tool. The real state of each review lives in a shared spreadsheet, because the tool’s statuses (open, pending, solved) don’t match the process (needs documents, in review, waiting on the partner, escalated, approved, rejected). Anything above a value threshold goes to email for sign-off.

So the number gets assembled: an export from the ticket tool, an export from the spreadsheet, a cross-check against yesterday’s approval threads. Forty minutes. Every morning.

The ticketing tool isn’t broken. It does what it was built to do — serve the average of everyone. The tax appears exactly when a process is core to how a team works and has to be bent into a shape made for everybody else.

Where the process actually lives

Walk through what this costs the team each month:

The click tax. Moving one item one step forward means updating the ticket, updating the spreadsheet row, and emailing the approver if the item is over threshold. About 90 seconds of pure navigation per item, 30 items a day: roughly 16 hours a month spent moving between tools, not reviewing.
The shadow spreadsheet. It exists because the tool can’t hold the real states — and it’s editable by anyone with the link, keeps no history, and disagrees with the ticket tool by Thursday. Someone spends 90 minutes every Friday reconciling the two, about 6 hours a month, and the quarterly review still catches mismatches.
The invisible queue. Approvals over the threshold live in email threads. Nobody can see what is waiting on whom — half of why the morning number must be assembled.
The seat tax. Two years ago the company bought the ticket tool’s workflow add-on at $39 a seat to force the process in. Fourteen reviewer seats, $546 a month — and the reviewers still work in the spreadsheet.

Add the lead’s 40 minutes a morning (about 13 hours a month) and the total is roughly 35 hours of coordination a month, plus $546 of licence for a module nobody uses. None of it is reviewing. All of it is the cost of a tool that almost fits.

What the right-sized interface looks like

Not a platform. One small app built around the process as it actually runs:

One queue, real states. The six states the process has, the fields reviewers use, nothing else. No settings pages nobody opens.
Approvals built in. Over the threshold, the item routes to the named approver: approve or reject with a note, one click, logged and visible. The email thread stops being the system.
The number on a screen. Pending items, ages, what is waiting on whom — live, computed from the data, zero assembly.
Sign-in with Google. The accounts everyone already has. No new passwords, no user-management project.
Connected to the systems of record. It reads and writes the ticket tool and the CRM directly, so it never disagrees with them. The spreadsheet retires.
No per-seat licence. It runs on the company’s own infrastructure, and the codebase and documentation are handed over — theirs to keep, change, or hand to anyone.

Note what this does not do: replace the tools they already pay for. The ticket tool keeps doing tickets. The internal tool sits alongside it and holds the one process. The add-on seats go.

The numbers, before and after

Morning number: 40 minutes of assembly → on a screen, current.
Friday reconciliation: 90 minutes → zero. There is one record, not two.
Navigation per item: ~90 seconds → ~10 seconds. One queue, one click.
Process overhead: ~35 hours a month → about 3 hours of genuine exception handling.
Workflow add-on: cancelled, $546 a month back.

The arithmetic: 32 hours a month returned at a conservative $40 an hour blended rate is about $15,400 a year, plus $6,500 of reclaimed seats — roughly $22,000 annually, for a tool that took three weeks to build. And the lead starts the day looking at the answer instead of building it.

One process, not a platform

Scoping is an hour with whoever owns the process: the states, the decisions, the thresholds, the one number that matters. The first version is deliberately small — one process, shipped in weeks. Then measure: is the queue moving faster, is the spreadsheet dead, does the right screen answer the morning question?

If it does, the next candidate is usually already waiting — the approval flow, the onboarding checklist, the dashboard the founder checks. Same approach, one process at a time.

The test is simple. If a core process in your company runs through a generic tool plus a spreadsheet beside it, the spreadsheet is the requirement document: it is telling you what the interface should have been all along.

If a review queue, an approval flow, or the one number that matters is assembled by hand every morning, that is the process worth building the interface for. We build small internal tools that fit the process exactly — on your data, your rules, and your codebase: Internal tools & dashboards.