Nobody tells you that getting good at something is mostly boring. The story we are sold is a steady diagonal line: effort in, improvement out, roughly in proportion, week after week. Real capability does not build like that — not in a person and not in a business.
It builds in short bursts separated by long stretches where nothing visible happens at all. Those stretches are where most initiatives get killed, most skills get abandoned, and most owners conclude that a perfectly good idea did not work.
What does the model say?
George Leonard set this out in his book Mastery, published in 1991, and the shape of it has held up better than most models of its age.
Learning goes: a brief spurt of visible improvement, then a long plateau slightly above where you started, then another brief spurt, then another long plateau. The plateau is not a stall. It is where the new capability is being consolidated into something you can do reliably rather than something you managed once. You cannot skip it, and you cannot shorten it much by trying harder.
Leonard also named the three ways people avoid mastery, and every one of them is recognisable in business.
The Dabbler
Loves the first spurt. The new system, the new market, the new adviser. When the plateau arrives, so does the enthusiasm for something else — and the setup cost of the last thing is written off without anybody calling it a loss.
The Obsessive
Refuses to accept the plateau and attacks it with more hours, more spend and more pressure. This usually breaks something or someone, and the wreckage gets blamed on the idea rather than on the pace.
The Hacker
Reaches the first plateau, decides it is good enough, and stays there for fifteen years. Of the three, the Hacker is the most common in established businesses and the hardest to spot, because from the outside it looks like stability.
Why does this matter to an owner?
Because you are the person who decides when to stop funding something, and you almost always make that decision from inside a plateau.
Everything worth building in a business has this shape. A new service line. A new market. A salesperson learning a complex product. A management team learning to run a meeting without you. The system nobody uses properly until month five. In every case the first two months produce visible movement, and then the graph goes flat while the actual capability is being laid down underneath it.
If you read “flat” as “failed”, you will systematically kill your best initiatives at exactly the point where the expensive part has been paid for and the return has not yet arrived. Do that four or five times and you will have spent a great deal of money to acquire no capability at all, which is the most expensive way to run a business there is.
A worked example
Illustrative figures, but the shape is real. A facilities management firm turning over £2m adds a new service line: planned maintenance contracts, rather than the reactive call-out work that has always been the bulk of the business.
Months one and two are the spurt. Three contracts signed, worth about £48,000 a year in recurring revenue. There is a launch, a mailshot, some genuine excitement, and the owner starts telling people this is the future of the business.
Months three to eight are the plateau. Two more contracts in six months. Recurring revenue crawls from £48,000 to £76,000. The mailshot has been done. The easy contacts have been called. The owner concludes the market does not really want planned maintenance, and it becomes the thing nobody mentions in the management meeting.
Underneath the flat line, this is what actually moved. The team learned to scope a site properly, so quoting time fell from two days to four hours. Cost to serve on the first three contracts dropped from 71 per cent of contract value to 58 per cent as the engineers stopped improvising. All three original contracts renewed. The proposal document went through four versions and stopped confusing people.
The Dabbler kills it in month seven, writes off roughly £26,000 of setup cost, and starts something new. What actually gets destroyed is not the £26,000 — it is the capability that had just finished being built.
The alternative is to hold the rhythm through the flat stretch: same weekly slot, same review, same standard. Months nine to twelve are the second spurt, and now the business converts faster because quoting is quicker, margins are better and there are three referenceable clients. Fourteen contracts by the end of year one is a plausible outcome — and none of it was available in month seven.
The test that keeps this honest
Some plateaus really are dead ends. A model that tells you to persist with everything forever is worse than useless, because it turns stubbornness into a virtue.
The distinguishing question is whether anything is improving underneath the headline number. Cost to serve, conversion rate, quoting time, delivery time, renewal rate, referrals, complaints. Pick two or three second-order measures at the start, before you need them.
If the headline is flat and the second-order numbers are moving, you are on a plateau and you should hold. If the headline is flat and nothing underneath it has moved in six months, that is not a plateau, that is a dead end, and persisting is just an expensive way of avoiding a decision. Knowing which one you are looking at is the whole skill.
How to use this in the next week
- List every initiative you have started in the last two years. Mark each one running, embedded or abandoned. Total up what the abandoned ones cost. That number is usually the most persuasive thing you will read this month.
- Take the one you are closest to killing and name two second-order measures. Then go and find out whether they have moved. Decide afterwards, not before.
- Set a no-decision window in advance. When you start something, write down the date before which you will not review whether to continue. It stops good things being killed in a bad week, which is when almost all of them are killed.
- Put practice in the diary, not exposure. Sitting through another course is exposure. Doing the thing badly, with feedback, on a schedule, is practice. Only one of them moves you off a plateau.
- Name which of the three you are. Dabbler, Obsessive or Hacker. Ask two people who work with you before you answer, because your own answer will be wrong in a predictable direction.
- Get someone external to hold the plateau with you. Not for motivation — for the specific job of asking, every month, what has moved underneath, so the decision to hold or kill is made against evidence rather than mood.
The mistake most owners make
They read the plateau as market feedback. It is almost never market feedback. It is the ordinary shape of getting good at something, arriving on schedule, and being misread because it feels like failure and failure is the only category available for a flat line.
The second mistake belongs to the Hacker, and it is quieter and more expensive. The business reaches a level of competence that works, and stops. Pricing set four years ago. A sales approach that was good enough once. A team that is fine. Nothing is going wrong, which is exactly why nobody addresses it, and the market moves past you at a pace you do not notice until a competitor is doing your job better.
The third is expecting the plateau to be short. It is usually longer than the spurt by a factor of three or four. Plan for six weeks, take seven months, and you will abandon it — not for want of persistence, but because you budgeted the wrong amount of patience.
Questions to ask yourself
- Out of 10, how good are you at doing the same thing well for six months while nothing visible happens?
- What have you abandoned at around month four, more than once? What is the pattern telling you about you rather than about the ideas?
- Where in your business are you competent and coasting — nothing going wrong, and nothing improving either?
- What is on a plateau right now that you would genuinely regret killing in two years’ time?
This is one of the most commercially useful ideas in Personal Coaching. Most owners do not have an ideas problem or an effort problem. They have a persistence problem with very specific timing to it, and it shows up around month four, every time.