Still running the firm on spreadsheets? What that actually costs, and what replaces them

A new administrator starts at an established firm. Good business, solid clients, years of steady work. By Wednesday of the first week the picture is clear: the entire operation lives in spreadsheets and a shared drive. Client records in one workbook, jobs in another, invoicing in a third that mostly agrees with the first two. The processes live in nobody's documentation and one long-serving colleague's memory.

We hear a version of this story constantly, and it is worth saying plainly: it is normal. Most firms between five and fifty people run this way, because spreadsheets are how sensible people solve problems quickly. Each file was the right decision on the day it was created. The trouble is what they add up to.

What the spreadsheet sprawl actually costs

The costs rarely appear as a line item, which is why the sprawl survives so long. They appear as friction, and the friction has recognisable shapes:

• Two versions of the truth. The figure in the meeting differs from the figure in the file, and twenty minutes go on working out which is right. Multiply by every meeting.
• The month-end assembly. Someone spends days stitching numbers together from separate files to find out how the business performed weeks ago. By the time you can see the picture, it is history.
• The person who is the system. One colleague knows which file is current, where the exceptions live and why row 214 is highlighted. The firm's operating model is their memory. Their holiday is an outage; their resignation is a crisis.
• Silent breakage. A formula gets overwritten, a column gets sorted without its neighbours, and the error compounds quietly until an invoice or a report surfaces it in front of a client.
• The onboarding tax. New starters take weeks to learn the folklore: which file, which tab, which colour means what. None of it is written down, because it grew rather than being designed.

None of these kill a firm on their own. Together they set a ceiling: the business can only grow as fast as its most overloaded spreadsheet and its most indispensable person.

In defence of the spreadsheet

The instinct to blame Excel is wrong, and firms that hear "you must digitise everything" from a consultant are right to bristle. A spreadsheet is a superb calculator and a fine place to think. What it is not is a shared database. It has no idea who changed what, no way to stop two people editing the same row, no concept of an approval, and no memory of last Tuesday. Used as the firm's system of record, it is doing a job it was never designed for, and doing it as well as anything could that lacks those foundations.

So the question is not "how do we get better at spreadsheets". It is "what should the system of record be, now that we have one worth recording".

What replaces them

Three routes, honestly stated. An off-the-shelf package for your sector is the right answer when your process is genuinely standard: accounting is the obvious example, and nobody should commission bespoke bookkeeping. A low-code platform suits firms with someone technical enough to own it, and works until the workflows outgrow what the platform allows. And a purpose-built system is the right answer when the way you run cases, jobs or clients is part of why people choose you, because that is exactly what a package will flatten.

We wrote a fuller guide to that decision: off-the-shelf or built for you. Whichever route fits, the destination has the same shape: one record for each case or job that everything else references, figures that are live rather than assembled, paperwork generated from the record rather than retyped into it, and a history of who did what. When that lands, the month-end assembly simply stops existing; the answer to "how are we doing" is a screen, not a project.

For what that looks like in a real firm rather than a diagram, walk through the operations platform we built for a UK probate property services business, or read the longer piece on that sector. The same six elements sit under every system we build: Data, Interface, Intelligence, Automation, Output and Governance.

How to start without a big bang

The firms that get this right do not announce a transformation programme. They pick the single most painful flow, usually whichever spreadsheet is edited by the most people, and move that first. The rest follows in order of pain. A good build imports your existing files rather than asking anyone to retype history, and runs alongside the old way until trust is earned. The spreadsheets do not vanish; they return to the analysis work they were always good at.

Common questions