Why this exists · Progression

The short version

We run a software services company. A few years ago the job changed underneath our own people, and nobody told them. We tried to fix it with outcome-based goals and found that telling somebody to own an outcome is not an instruction. So we went and looked at who was actually moving and who was stuck, and found the difference was never knowledge. It was how much ambiguity arrived with the work, and whether it got absorbed or handed back.

That is the whole idea. Five contracts describe how much ambiguity a person can absorb. Four muscles describe how they absorb it. We now run it on ourselves across every function, and we are opening the vocabulary because it is not worth anything if it only works in one company.

01 The problem

The job changed underneath our people, and nobody told them.

This is easiest to see in one person's week.

A designer on one of our client teams used to do something clear. A brief came in. She produced the design. She handed it over. She was good at it, she was recognised for it, and the brief was always somebody else's problem to write.

Then the shape of client work changed. Now the same person is expected to understand that client's whole business, form her own view of what is actually wrong, produce a design, and be able to defend why this design produces that business result. Nobody handed her the brief, because there isn't one.

She was not underperforming. She was still being excellent at the job she was hired to do. It was just no longer the job.

We saw this in engineering first, which is why this work started with a question about forward-deployed engineers. Then we found the same thing in design, then in delivery, then in marketing, then in our own operations. The specifics differed completely. The shape was identical every time: the amount of ambiguity arriving with the work went up, and the language we used to describe seniority had not moved at all.

Meanwhile the floor was rising from the other direction. The work that used to be how a person proved themselves, take a clear spec and return it correct, is work that capable AI now does quickly and cheaply. That is not a prediction. It is the current state of the bottom of the ladder, and it means the old proof of competence is getting less informative every quarter.

02 What we tried first

We set outcome-based goals. Then we watched them not work.

The obvious fix was to stop measuring activity and start measuring outcomes. So we did that. Everybody got outcome-shaped objectives.

It half worked, and the half that failed was instructive. People agreed with the objectives. People wanted to hit them. And a great many of them genuinely did not know what to do differently on Monday morning.

We kept saying the goals come from outcomes. Nobody could tell you how to actually do that.

"Own the outcome" is a description of a destination, not an instruction. It tells a person where they are not, and nothing about the next step. The gap was never motivation or effort. The gap was that nobody could state what the next rung actually required in terms concrete enough to act on before Friday.

The second thing we tried was training. Mostly on AI tools, because that was the loudest answer available. Tool fluency turns out to be table stakes with a half-life of about a quarter, and it does not answer the real question, which is what a person is for once the tool can do the task. We are glad we did it. It was not the thing.

03 What we found

The difference was never knowledge.

So we stopped theorising and looked at actual people. Same skill on paper, same effort, same years, wildly different trajectories. We wanted to know what separated them.

It was not knowledge, and it was not talent in any sense we could measure. It was this: how much ambiguity arrived attached to the work, and whether the person absorbed it or handed it back.

Two people are given the same project. One asks what exactly is wanted, gets told, and delivers it correctly. The other notices the ask is slightly wrong, decides what should actually be built, and returns something nobody specified but everybody needed. Both did good work. Only one of them absorbed anything.

Once we could see it, we could see it everywhere, including in our own promotion conversations. Every genuinely difficult promotion decision we had ever had was an argument about this, conducted without the words for it. Every easy one was a case where a person had obviously absorbed a category of ambiguity they had not been asked to.

This is not a clever insight and we do not claim it as one. People have been circling the idea that seniority is about ambiguity for decades. What did not exist was a version written down plainly enough that a manager and a team member could use it in a conversation on a Tuesday, without a form, and both walk out agreeing on what happens next.

04 The design decisions

Four choices that made it usable instead of true.

Plenty of frameworks are correct and unused. These are the four decisions that were about use, not accuracy.

Promises, not levels

Levels already existed everywhere and meant nothing. Titles inflate over time, differ wildly between companies, and are frequently a negotiation outcome rather than a description. So we wrote the rungs as promises instead: what can you credibly say about the kind of work somebody can hand you. A promise does not inflate. You can either make it or you cannot, and you know which.

Four muscles, not forty competencies

We could have written a competency matrix. Competency matrices get read once a year, at review time, by somebody filling in a form, and never again. The test we set was: could a manager use this in a one-to-one, from memory, with no document open. Four is what survived that test.

They are also not four separate skills. They are four phases of the single loop everyone runs every week: something ambiguous arrives, you turn it into intent, you get it done through more than your own hands, you decide what was worth doing, and you move people you do not control.

One ladder, including for managers

Management is not a parallel track on this ladder. Managing is Leverage and Influence pointed at people-outcomes rather than at things. We made this choice because two-track systems quietly rank one track above the other, and then people take management jobs they do not want in order to progress.

Corroboration, not self-assessment

Every framework we looked at placed people either by self-assessment or by a manager's impression. Both are unreliable, and unreliable in known directions. So the placement refuses to accept your opinion about yourself. The question is who could describe what changed, cold, with nobody prompting them. It is harsher, and it is the only version that survives contact with a real promotion conversation.

05 What we refused to build

The refusals say more than the features do.

Each of these was proposed, seriously, and turned down. They are written into how the thing works, not into a values page.

Never a number. A person's standing is a contract they can make, expressed as evidence-backed judgment. Not a score, not a rating, not a percentage. The moment it becomes a number, people optimise the number.
Never a leaderboard. No ranking of people against each other, anywhere, for any reason. Careers are not a tournament and treating them as one produces exactly the behaviour you would expect.
No machine verdict. Software can organise the evidence. It does not decide whether somebody moved up. Humans decide promotions, and a system that returns a verdict lets humans stop being accountable for a judgment that is theirs.
No self-certification. Nothing you assert about yourself places you, ever. The floor is two corroborated pieces of evidence, and you cannot be your own corroboration.
Reasoning frozen before the result. For anything involving a judgment call, the thinking is recorded before the outcome is known, so the call gets graded on the reasoning rather than on luck.

Why any of this matters to you. A career framework is mostly a set of incentives wearing a costume. If you want to know what one will actually do to a company, ignore what it measures and look at what it refused to measure.

06 The honest commercial answer

What we get out of giving this away.

You should assume a company publishing a career framework wants something. We do. Here it is, without the polish.

What is free, permanently

The whole framework. Every contract, every muscle, every test, every way people fake it.
The placement exercise, unlimited, no email, nothing stored.
The community, the open sessions, and the roles that get shared in it.

What we sell

Rooms. Small, human-led, on your own live work, one muscle at a time.
We sell that and not the vocabulary because facilitator time is the one input that genuinely costs something and cannot be copied.
The framework works whether or not you ever pay us a rupee. If it did not, it would not be worth publishing.

And the part most companies leave out

We are a services company, and we do sometimes hire people we met here. That is a real reason this community exists and it would be silly to pretend otherwise. We would rather write it on this page than have you work it out later and wonder what else we did not mention.

What that does not mean: nobody is sold to in a direct message, membership is not a recruitment funnel with a price on it, and being in a paid room has nothing to do with being hired. Those are the lines we hold, and they are checkable, because you are in the room and you would notice.

07 What we do not claim

The limits, written down while nobody is asking.

We did not invent this. The idea that seniority tracks ambiguity is old. Our contribution is a plain-language form that two people can use in a real conversation without a document.
We do not award you a band. Bands move in rooms inside your own company, decided by people who have watched you work over time. Nothing we do can change that, and anybody selling you a level is selling you a certificate.
This will not get you promoted on its own. What it can do is make sure that when a room reads you, there is something real to read, and that you are not surprised by what it says.
We do not know where the floor settles. Capable agents now do a growing share of the first two rungs. Our honest read is that the entry point to a career is moving up the ladder rather than the ladder changing shape, and we are not certain of that.
We have not resolved deep craft. Read literally, the fourth rung requires that other people's ceiling went up, which appears to require teaching or systems. That would imply a world-class specialist working largely alone cannot pass the third rung. We do not believe that is right and we do not yet have a clean formulation of the alternative.

Both of those last two are printed on the contracts page as well. A framework that cannot say what it does not know is a sales document.

08 Where it is now

We run it on ourselves. That is the only reason to trust it at all.

Every function at our own company is now described by these five contracts and these four muscles. Engineering, design, delivery, marketing, people, finance. Not as a poster. As the actual language of growth conversations.

That is not offered as proof that it is correct. It is offered as proof that we are exposed to being wrong about it, which is a different and more useful thing. If the framework produces bad outcomes, we get those outcomes first and we get them in a company we have to keep running.

The vocabulary is deliberately company-agnostic, and that is not generosity. A career framework that only works inside the company that wrote it is a compliance system. The test of whether this is real is whether you can take it to an employer that has never heard of us and still have it describe your week accurately.

That is what we would like you to go and check.

Place yourself Read the framework

We did not set out to build a framework. We set out to answer one question we could not answer.