By James Shaw, COO
Every operation starts the same way. Someone builds a spreadsheet, and it has a few hundred rows in it, then a few thousand, and you and your team have to be embedded in it every day just to read the thing. You know which column means what. You know someone's row is wrong because someone's row is always wrong. Then someone gets ambitious and puts charts on top of it. Trend lines, some color coding, a summary tab up front. Now there's two things to faff with instead of one. And then, and this is the part nobody warns you about, you stop looking at either one. Not because anyone got lazy. There's just too much of it, and none of it is telling you anything you didn't already half know.
I'm not an accountant, right, so I came at this from the wrong direction on purpose. I don't sit down every morning wanting to admire our revenue chart. I want to know the one thing that's broken before it costs me money for another week. Those are different questions, and most reporting tools are only built to answer the first one.
I went through that whole progression building the reporting for our own company, and somewhere around the third pass I gave up on the idea that showing people more data was the answer. Because you don't actually need to know all of those numbers. You just need to know the ones that are wrong.
I tested that on our own expenses first, because I'm not going to build something for a client that I haven't tried on myself. I pulled two years of our credit card transactions in, every vendor, every recurring charge, all of it, and I built a rules engine on top of it. Sounds fancy. It isn't. It just flags anything that doesn't look like what it looked like last month, the way the smoke alarm in your kitchen says nothing all day and then actually gets your attention. I ran it, and what came out the other end wasn't a report at all. I've just imported two years of credit card transactions and ended up with a just dashboard with five lines on it that says you just need to look at this one. It went up by 40% this month. Five lines. One of them mattered.
Two years of transactions is not a small amount of data, and the honest version of this story is that the first pass looked like nothing at all, because I'd built it wrong, and had to go back and fix the rules before it would tell me anything I could trust. That part never makes it into the pitch. But once it was working, the difference was obvious. The spreadsheet tells you everything. The five lines tell you what to do today.
Then we ran the same idea across the whole company. Not five lines anymore, obviously, we're bigger than a credit card statement, but the shape of it is identical. Every day there's a list, and it's short, and it's ranked, so whatever actually needs a human today sits at the top. At the last all-hands I pulled up the real list, unfiltered, and told the room: these are the 10 things that someone should look at today. There's a million numbers every day, so this is just a way of us searching through the fog and seeing exactly what we need to do today.
Nobody enjoys landing on that list, by the way. I'm not going to pretend they do. It points out real mistakes, wrong rate, default role picked instead of the correct one, on an hourly basis, and I said as much to our own accountant when she saw it: people don't like being on the warning sheet, but it's a great tool. And she's like, the reason it's a great tool is it teaches them how to do it right, and then they drop off the warning sheet. That's the whole point of building it that way. Someone picks the default role because they're moving fast, not because they're trying to get one past anyone. The system tells them once, they fix it, and the same mistake stops showing up.
The honest end of this is that we've never gotten our own list to zero. The goal is an empty warnings page, nothing on it, everything clean. We've been running this at Inventive every day since we built it, and we have not had a single day without something on it. Fix the thing at the top and something else moves up to take its place. That sounded like it should be depressing the first time I noticed the pattern, so it was, for about a week. Then I got used to it. The goal is still the empty page. We just haven't seen it yet.
So no, I'm not going to build you a dashboard with all of your numbers on it, tempting as that sounds in a sales meeting. I'll build you the five lines instead, and on most days four of them won't need you at all. If you want the full spreadsheet too, fine, it's still there underneath, page after page of it, in case you or I ever need to go trace a warning back to where it started. Most people never open it again once they've seen the short list.
If you want to know what your own warning sheet would look like — what's the first thing you'd like the answer to? Start the conversation.