Outgrown Excel? How to move your business off spreadsheets
A spreadsheet stops being a tool and becomes infrastructure the day a broken formula can cost you money. The fix is almost never an enterprise platform. It is one small app that does exactly what the load-bearing sheet does — shared data, logins, validation, history — built in three to five weeks, while Excel keeps doing the ad-hoc analysis it is genuinely good at.
This post covers how to tell which of your sheets crossed the line, what replaces them, and how to migrate without a big-bang rebuild.
Five signs a sheet crossed the line
Nobody decides to run a company on Excel. It happens one column at a time, and it works — right up until it doesn't. The line between "useful tool" and "accidental infrastructure" is surprisingly crisp:
1. More than one person edits it. Excel has no real concept of two people, which is why you have Q3_pipeline_FINAL_v7_SM_edits.xlsx.
2. The filename has a version number. Version numbers in filenames are a database migration system run by hand. You already know which file this describes.
3. One person understands the formulas. When the sheet's author is on holiday, part of your business is on holiday. That is key-person risk stored in cell notation.
4. Someone re-keys the same data twice. If a number is typed into the sheet and then again into the accounting system or the CRM, you are paying a person to be an integration — the most error-prone integration there is.
5. An error has already cost you. A mis-dragged formula, a stale copy sent to a client, a row sorted without its neighbours. If you have a story like this, the sheet has already presented its invoice.
Two or more of these on one sheet, and that sheet is infrastructure. Infrastructure deserves better than =VLOOKUP.
Why the platform you looked at didn't fit
The standard next move is to shop for software, and for the standard parts of your business it is the right move: accounting, payroll, and email are solved problems — buy them.
But the load-bearing sheet is usually load-bearing precisely because it encodes something specific about how your business runs: your pricing logic, your job-costing method, the way you schedule crews or track consignments. That specificity is why the platform demo always feels 80% right and why the remaining 20% never closes. You end up bending your process to fit someone else's data model, paying per seat for everyone who glances at it monthly. We wrote a fuller build-vs-buy framework, but the short version: standard process, buy; distinctive process, build.
What replaces the sheet
Less than you'd think. The replacement for a load-bearing spreadsheet is not "a system" — it is a small app with four properties the sheet can't have:
| The sheet's job | What replaces it |
|---|---|
| Everyone's copy of the truth | One shared database; everyone sees the same live numbers |
| Columns anyone can mangle | Forms with validation — wrong data is rejected at the door |
| Formulas one person understands | Calculations in code, tested, visible to whoever asks |
| No memory of who changed what | History on every record — who, what, when |
The mistake to avoid is scope. The app should do what the sheet does — not what a platform brochure does. Every feature beyond the sheet's actual job is cost and delay attached to a hypothesis.
The goal is not an Excel-free company. Ad-hoc analysis, one-off modelling, "let me just check something" — Excel is unbeatable there and should stay. You are only retiring the sheets where errors cost money. A good replacement app exports clean CSVs precisely so the analysts can keep analysing.
How the migration actually goes
Week 1: the sheet is mapped — every column, every formula, every human workaround (the cells that are yellow "because Dana knows what that means"). This archaeology is the real work; the sheet is the spec.
Weeks 2–3: the app is built and the historical data that matters is migrated in. Not all of it — most sheets carry years of rows nobody will query again. Bringing the last 12–24 months is usually right.
Weeks 3–5: the team runs both in parallel briefly, the app wins on the first Monday someone finds a number in three seconds that used to take a phone call, and the sheet is archived read-only. It never quite gets deleted. That's fine.
For a commit-by-commit account of a real one, see the build log.
What it costs
A focused workflow app of this shape runs $12,000–$30,000 including migration, with simple internal dashboards starting around $8,000 — the full bands are in our pricing breakdown, or you can get a range for your specific sheet in two minutes with the cost calculator. Against that one-time number, price the status quo honestly: the hours lost weekly, forever, plus the next formula error at the worst possible moment.
Frequently asked questions
How do I know my business has outgrown Excel?
Five signs: more than one person edits the same file, the filename has a version number in it, one person is the only one who understands the formulas, someone re-keys the same data into a second system, or a broken formula has ever cost you money or a customer. Two or more and the spreadsheet has quietly become infrastructure.
What should replace a business-critical spreadsheet?
Usually one small custom app that does exactly what the sheet does — shared data, logins, validation, and history — not an enterprise platform. Buy software where your process is standard (accounting, payroll). Build where the sheet encodes the way your specific business runs, because that is the part no platform will ever quite fit.
How much does it cost to replace a spreadsheet with an app?
A focused workflow app typically costs $12,000–$30,000 and takes three to five weeks, including migrating the data that matters. Simple internal dashboards start around $8,000. That is a one-time cost against hours currently lost every week, forever.
Do we have to move everything off spreadsheets at once?
No, and you shouldn’t. Replace the one or two load-bearing sheets where errors cost money; keep Excel for the ad-hoc analysis it is genuinely good at. Big-bang migrations fail because they try to move the analysis along with the infrastructure.
Which of your sheets crossed the line?
Bring the one with the version number in the filename. Thirty minutes, and we'll tell you what replacing it takes — or whether it should stay a spreadsheet.
Book a 30-minute call