eventuley

Writing

Why event spreadsheets fail, in eight specific ways

· 7 min read

Every small events team builds the same spreadsheet. It starts as one sheet with a date and a headcount, and by the fourth event it has tabs for guests, costs, kit and a summary that nobody fully trusts. It is not a bad spreadsheet. It is a spreadsheet, doing something spreadsheets cannot do.

The marketing page for this product lists five ways it breaks. That is not the real number — it is how many fit on a page next to everything else. The workbook we read had eight, and the three that did not make the page are the ones that surprised us most.

What follows is that list. None of it is invented, none of it is a generic complaint about Excel, and none of it came from another article. It came from one working events team's operations workbook, read line by line, with permission. The file itself is not published and never will be — it contains real guests' names and contact details — so what you get here is the shape of each failure, which is the part that generalises anyway.

1. The numbers stop agreeing with each other

Business-level profit and the sum of per-event profit drift apart.

The cause is mundane: some costs belong to no single event. A website subscription, an accountant, a bulk order of napkins split across four dinners. They are real money and they have nowhere to sit, so they end up in a business-level total and not in any event's column — or in one event's column arbitrarily, which is worse.

Neither total is wrong on its own terms. That is precisely the problem. Nothing in the sheet surfaces the gap, so both numbers keep being quoted, and which one you happen to be looking at determines whether the month was good.

The cost is not the arithmetic. It is that after a while nobody quite believes either figure, and decisions quietly stop being made from the numbers at all.

2. A metric that is quietly wrong

The workbook had a "budget utilisation" figure. It divided spend by expected revenue.

Budget utilisation is spend against planned budget. Dividing by expected revenue produces a number that moves when you change your revenue forecast, which is not a thing budget utilisation should do. It is a ratio of two real quantities that answers no question anybody has.

And it looked completely fine. It was a percentage, it sat under a sensible label, it went up when spending went up. There is no cell colour for "this is measuring the wrong thing."

This is the failure mode worth being most afraid of, because a spreadsheet cannot help you here at all. A wrong formula and a right formula are the same kind of object.

3. A metric that was never finished

There was a cell labelled "most profitable event". It was empty.

Somebody designed the summary, laid out the labels, and did not get to the formula behind that one. Then the sheet was used for months. The label is still there, promising an answer, and the space beside it is blank.

Nobody is confused by an empty cell — you look at it, register that it is empty, and move on. That is the point. It costs nothing to ignore, so it is never fixed, and the question it was meant to answer simply stops being asked.

4. Money drifts

Amounts held as floating point accumulate error.

0.1 + 0.2 is not 0.3 in a spreadsheet any more than it is anywhere else. Across a few hundred rows of ticket prices, part-payments and split costs, this becomes visible: a total that is off by a few cents, then by more, with no single row you can point at as wrong.

The practical damage is the hour spent looking for it. The structural damage is that the sheet has now taught you it can be slightly wrong, and you cannot tell it when it is.

Money is not a number. It is an integer count of minor units plus a currency, and the currency decides how many minor units there are — the yen has none, the Kuwaiti dinar has three. A system that assumes two decimal places is already wrong for a customer it has not met yet.

5. The equipment sheet with no rows in it

A tab for equipment. Column headers. One formula. Zero rows.

Somebody knew that tracking what goes out to an event and what comes back matters — they designed the control. Then it turned out that keeping it current by hand, on a Saturday, while loading a van, is not a thing a person does. So it was never filled in, and it sat there as a monument to a good intention.

This is the one that generalises furthest. A control that depends on somebody remembering to update a sheet is not a control. It is a note asking for one.

The three that did not fit on the page:

6. People are modelled twice, badly

Guests get proper treatment: a row each, an identifier, contact details, a payment status.

The team are bare strings typed into a column. The same person appears as "Ali", "ali" and "Ali K." depending on who typed it and how much of a hurry they were in.

So you cannot filter to your own work. You cannot count how many tasks anybody has. You cannot ask who is doing too much, which is the question a small team most needs answered and least often asks in time.

The asymmetry is telling: the people you are selling to got a data model, and the people doing the work got a text field.

7. Structure that cannot be enforced

A due date reading ∞. A date mistyped four years into the future. Status cells left blank despite a validation list existing one menu away.

Every one of these is a person doing something reasonable under time pressure. ∞ means "no deadline" and is the clearest possible way to write it. The four-year typo is a fat finger nobody re-reads. The blank status is a task captured in five seconds because the alternative was forgetting it.

A spreadsheet can offer structure. It cannot require it, and the moment requiring it would slow somebody down at the wrong moment, it loses. Sorting by due date now silently puts a real deadline after a typo.

8. One table with two meanings

At some point a table starts holding a second kind of thing, because there is nowhere else to put it.

A column means one thing for some rows and something else for others. There is a convention. The convention lives in one person's head, and it is correct until they are on holiday.

This is the failure that makes every other failure harder to find, because now you cannot even write a formula that is unambiguously right.

What actually follows from this

Not "use better spreadsheets." Every one of these eight is something a spreadsheet cannot make impossible, and that is a property of the tool rather than of the person using it. A grid of cells has no opinion about what belongs in it.

Each of them, though, is something a data model can make impossible:

That is what eventuley is. It was built from this workbook, and the list above is what it was built against — which is also why the free plan is a real plan rather than a trial. If one event's worth of use does not show you the difference, more of it will not either.

Start with one event — one event, one seat, 100 contacts, no card.