Why telling people what you want doesn't usually work

The first few years of building RORA were the most fun I've had at work.

I'd started it because accounting was broken at both ends, and we ended up selling fractional finance teams to startups.

Five of us, one room, and everything the company knew was in my head. I could tell you the state of every client, who was behind on what, and which of our people was having a rough week. Decisions took about ninety seconds, because I made all of them.

At five people that is the right way to run a company. It's fast, it's cheap, and writing any of it down would have slowed us down for no benefit at all.

The problem is that it has a shelf life, and nothing announces when you've reached the end of it.

Let's dive in.

Then it got hard

Capacity was the product. Our people were the thing the client actually bought, so every problem in the business arrived as a person problem.

A few years in I had two queues on my desk. Unhappy clients, and unhappy people. Neither of them ever emptied.

At first I thought they were two separate problems. After a while I realised they were the same one. The client was unhappy because the person on their account wasn't doing something the client thought was obvious. The person was unhappy because nobody had ever told them it was their job.

Both ends landed on me, because I was the only person in the building who could say what good looked like. I was doing that on Sunday nights, and I was doing it instead of the work I'd started the company to do.

I thought I'd told them

I had told them. In 1:1s, in all-hands, standing at somebody's desk on a Tuesday afternoon. I'd explained what a great account manager does maybe fifty times, and I'd explained it slightly differently every single time.

Nearly every founder I work with now is doing the same thing. They believe they have set a standard because they have said it out loud, repeatedly, to individuals, with real feeling. Saying it is not the same as writing it down, and it isn't close.

Try this in your own business. Ask three people in the same role, separately, to write down what good looks like in their job and what would get them promoted. You will get three different answers, and all three will be a degraded copy of something you said once, filtered through whatever that person was worried about that week.

Spoken standards decay. Written ones don't, and a written one is the only kind that somebody else can hold a colleague to without you in the room.

The skills matrix

We started with the client-facing roles, because that was where the pain was.

For every level, three things, written down: what you do, what you don't do yet, and what you have to be seen doing to move up.

The word that matters there is seen. Every criterion had to be something a manager could point at. Not "shows good commercial judgement", which means whatever the assessor had for breakfast, but "has told a client no, and the client stayed".

We attached money to it, and reviewed against it every three months instead of every year. Annual reviews are too rare to change anybody's behaviour and just frequent enough to feel like a verdict. On a quarterly cycle the conversation stopped being about how somebody felt they were doing and became about the line in the matrix they hadn't demonstrated yet.

Two things came out of that which I hadn't planned for.

Clients started getting the same experience whoever was on their account. In a service business that consistency is what turns a group of individuals into a product. What the client is really buying is the confidence that next month will be to the same standard as this one, whoever happens to be doing it.

And hiring got easier, because I finally knew what I was hiring for. You don't fill a role like that once. You fill it ten or twenty times over the life of the business, and every time you do it without a written definition you are gambling on your own mood in a forty minute interview.

The KPI cascade

That fixed the client-facing side. The HQ team was still coming to me, for a different reason. They didn't know what mattered most this quarter, so they asked the person who did.

So we built the other half. One company goal for the year, broken into a number for each function, and then into a number that individual people owned.

Every number had one person's name against it, not a team's. And if your name was on a number, you could make any decision that moved it without asking me first.

The other thing it did, which I hadn't expected, was make the HQ team feel like they were in the same business as the client teams. When you can see your number rolling up into the one on the wall, you stop being overhead.

Then I had to promote nearly everybody

Once the matrix existed I could see that a lot of people were already doing the job a level above the one they were paid for. I'd been getting that work for free, which is a polite way of saying I'd been underpaying people who trusted me. We moved most of the client-facing team up a level and the wage bill went up by [X]%. Best money I spent that year. Sam, Miriam and Charlie were already running the business between them, so Sam became CEO.

Every one of those promotions handed somebody a decision I used to make, and eventually there weren't many left. Nobody warns you how uncomfortable that is. Being needed is a large part of why the work felt good, and I'd engineered it away on purpose. It was also exactly what I'd set out to do, and it still took me the best part of a year to be glad about it.

What it turned out to be worth

Some time later we sold RORA to AAB, a Goldman Sachs backed group. No earn-out. I didn't stay on.

Buyers see one risk above all others in a founder-led business: the founder is the product. So they price it, holding back a chunk of the money and tying you in for two or three years to prove the thing still works without you. We didn't need one, because the business had already been running without me for over a year with somebody else's name on the CEO line. The proof was done before anybody asked for it.

What I'd tell a founder today

Three things I'd say to anyone sitting where I was then.

1. If it only exists in conversation, it isn't a system. Say it out loud as often as you like. Write it down once.

2. Unhappy clients and unhappy staff are usually the same problem in different clothes. Somebody is doing a job whose edges were never defined.

3. Write the levels in behaviour you can observe. If a criterion can't be evidenced by a manager who wasn't in the room, it will be scored on personality instead.

None of this is complicated. It is just work that never feels as urgent as the client email that came in twenty minutes ago, and that is exactly why it doesn't get done.

What I do now

I run Redway Ops. We work with founder-led businesses, usually between £1m and £10m, where the founder has become the bottleneck and hiring hasn't fixed it. Sometimes that means writing the matrix and the numbers with you. Sometimes it means finding the operator who owns them and managing them so that you don't have to.

I've put the map I wish I'd had into The Founder Extraction Ladder. Nine rungs, from a business that stops when you're ill to one that somebody else could own tomorrow. It's free, and it's blunt about which rung you're on rather than the one you'd like to be on.

If you read it and want to argue with me about where you sit, I'd enjoy that. Drop me a message.