Two numbers that tell you if a project will move

TLDR — Years ago, as an interim CIO, I inherited a technology budget process that had decayed into bureaucracy: annual set-piece reviews, budget clamour, whoever shouted loudest, and a year end scramble to spend what was left before it vanished. I fixed it with two crude numbers and one blunt rule. How many layers sit between the person who wants the thing and the person building it, how much of their time the sponsor actually spends with them, and money released in small chunks to the next useful milestone with the sponsor in the room. Those two numbers will tell you which of your projects are going to move. It worked at the time. Explaining why it worked has taken rather longer.

There is a particular meeting that anyone who has worked in a large organisation will recognise. It is the annual portfolio review, or the quarterly one if the place is really keen. Everybody arrives with a slide. Everybody wants budget. The projects that get funded are, too often, not the clearest or the most valuable, but the ones attached to the loudest voice or the most senior sponsor. Scopes are vague. Business cases are written to clear the bar rather than to be true. And then, reliably, comes the strangest part of the annual cycle: the year-end scramble, where money allocated to a big programme that never quite moved is flung at something, anything, in the last few weeks, because an unspent budget is treated as a worse sin than a wasted one.

I inherited exactly this as an interim CIO, some years ago now. Meanwhile, small, well defined, obviously useful bits of work sat unfunded, because they were too modest to fight their way onto the big slide. It was a machine that was almost perfectly designed to fund the wrong things slowly.

So I did something about it.

A word on what that was, before anyone accuses me of inconsistency. On a blog that spends a fair amount of its time warning people off buying methodologies, I went and built a small one of my own. Deliberately. With a formula in it. There is no contradiction. What I object to is the packaged noun and the all encompassing methodology, bought precisely so that nobody has to do the thinking. I am an enthusiast for the opposite: hand crafting your own methods, tailored to the outcome you actually want, chosen by a mindset that starts with customers and value.

And there is a third version of the same abdication which gets far less attention than it deserves. "We have always done it like this." Received wisdom is the same surrender as the bought methodology, only cheaper and much better disguised. Nobody sold it to you, so nobody has to defend it. It simply sits there, and the process, the protocol and the job titles quietly do the thinking that people ought to be doing themselves. That was what I had inherited, and that was what the whole exercise was built to dismantle. The job of a CIO is to get the most value out of a finite budget. The process I walked into was doing close to the opposite, so I built something that would. I would do it again tomorrow.

What I could not have articulated at the time was why it worked. I knew that it did. The reasoning was mostly instinct, sharpened by watching enough expensive things stall to recognise the pattern. It has taken the Mindset, Method and Momentum frame I write under now to put proper words to it, and the words are these. The method was never the point. It was an instrument, and what it measured were the two things a budget process is best at hiding.

Two things govern pace

I convinced myself that the speed at which good ideas turn into delivered things comes down to two questions, and not really to the budget at all.

The first is distance. How many people sit between the person with the clearest view of what is needed and the person who will actually build it? Not the senior manager who owns the slide, but the key user or designer who genuinely understands the need. And not that person's manager, but the developer or maker who will do the work. Every layer between those two is a place where intent leaks, gets reinterpreted, and slows down.

The second is attention. How much of their time does the person who wants the thing actually spend with the person building it? Not in steering committees. With them, on the thing.

Neither is really about money, and that is the whole point. Each stands in for something you cannot ask about directly. Distance stands in for whether the conditions for movement exist at all: a short, direct line between wanting and building, with nothing in the middle to dilute it, reinterpret it or slow it down. Attention stands in for something even harder to put on a form, which is whether the person who asked for this genuinely means it. You cannot write "do you actually care?" into a governance pack. You can, however, notice how many hours a week someone spends sitting with the people doing the work, and that answers the same question rather more honestly.

So I wrote it as a deliberately crude formula, because a crude number you can argue about beats a vague feeling you cannot.

At the good end, a sponsor sitting directly with the maker, no layers, all in: a hundred out of a hundred. At the bad end, the marketing manager with a brilliant idea who briefs a product manager who briefs a portfolio manager who talks to an account manager who appoints a project manager who assigns a developer, and who personally gives the whole thing a tenth of their time in reviews. Four or five layers, ten percent attention. A score of two. You can feel that project in your bones. It is probably still going!

The number was never meant to be precise. It is relative, not absolute: a way to lay your work side by side and decide where the attention, and the money, should actually go. What it does do is dodge the two ways these conversations normally fail. On one side is spurious accuracy, the score carried to two decimal places, the weighted model with fourteen inputs, dressed up as science and none the truer for the fancy decoration. On the other is the vague opinion, the "I just feel this one is struggling", which may well be right but cannot be tested or compared. A crude number sits deliberately between the two. Nobody can pretend it is precise, so nobody defends it as gospel. But it is specific enough to disagree with, and that debate turned out to be where the real value lay.

So what is a good score?

You need a line somewhere, or the number is just decoration. I set the working baseline at twenty.

Twenty sounds low against a theoretical hundred, but in reality it turned out to be tough. That was deliberate, and it is worth seeing what a score of twenty actually costs. With the sponsor sitting directly with the maker and no layers at all, twenty means a fifth of their working time. Put one person in between and you need forty percent to hold the same score. Two layers, sixty. Three layers, eighty. Four layers and the sponsor would have to be working on nothing else whatsoever, which nobody is going to do and, in the real world, nobody should.

 "Beyond about three layers you cannot buy your way back with attention. You have to shorten the line."

That is the genuinely useful thing the arithmetic tells you, and I did not spot it at first. Beyond about three layers you cannot buy your way back with attention. There is no amount of sponsor commitment that rescues a project sitting five people away from the work. You have to shorten the line instead. The number stops being a score and starts being an instruction.

Then the confronting part. On the first pass across a live portfolio of more than fifty projects, a handful cleared twenty. A handful, out of fifty!

I want to be careful about what that did and did not mean. It did not mean the other forty-odd were worthless, or that the people on them were idle. Most were perfectly sensible pieces of work being done by capable people. It meant something less comfortable: that the organisation had, without ever deciding to, arranged almost all of its work so that the people who wanted things sat a long way from the people building them, and visited occasionally. Nobody chose that. It had simply accumulated.

What I actually did, and how the circle closed

I stopped running the set-piece monthly portfolio reviews, with their pre-meetings and their theatre and fiefdom, and replaced them with something smaller and much more frequent. We met weekly. Money was released in small amounts, only enough to reach the next point that would genuinely tell us more. And there was one non-negotiable rule that made me quite unpopular: to get the next chunk of budget, the sponsor had to be in the room, in person, alongside the person actually doing the work. Not a delegate. Not a slide. Them.

It sounds like an administrative tweak. It was really a filter. A sponsor who would not give an hour a week to sit with the key deliverer was telling you, clearly, that the thing did not matter enough to fund, and you found that out in week one instead of quarter three.

It also did not stay where it started. Once people could see the number, the obvious question was why we only asked it of things we had already paid for. So it ended up running at three points. At initiation, where a new project had to make its case with the two numbers alongside everything else, said out loud, in front of the person who would have to live up to them. A surprising amount of weak work simply withdrew at that point, which saved everyone a year. And yes, plenty of projects had multiple people doing the work, but then the debate became one about structure and clarity, which was also useful. At each milestone review, the same two numbers again before the next chunk of money moved, which is where it bit hardest, because a score that had quietly slipped since kick-off told you the sponsor's attention had wandered long before any status report admitted it. And periodically across the standing portfolio, the pass that started all this, to see what had drifted.

Existing work reviewed. Funding confirmed, reduced or stopped on what that review found. New work challenged on the same basis before it started, and challenged again at every gate. The same question asked at every point where money moves.

The polished versions, and why I would still build my own

I was not the only one to notice that the year-end scramble is real. When researchers looked at government IT spending, they found money obligated in the final week of the budget year ran at around five times the normal rate, and that projects procured in that last-week rush were between two and six times more likely to be rated low quality. That way of budgeting does not merely waste money. It reliably buys worse things.

And the deeper idea, that annual project budgets are the enemy of pace, is no longer a fringe opinion. The big firms preach it now. Fund teams, not projects. Give product teams perpetual funding rather than locking a number to a fixed scope a year in advance. Release money against learning rather than against a Gantt chart. Deloitte and Thoughtworks and half the consulting world will sell you a very polished version, and I should say plainly that the thinking in it is often good and a great deal better than the annual guess.

But watch what happens on the way from the idea to the invoice. A way of behaving gets a name, then a diagram, then a maturity model, then a price. And you are back where this blog usually finds you: buying the noun instead of doing the verb, purchasing a way of working so that nobody has to do the harder thinking underneath it. The irony is fairly rich. A body of thought about releasing money against learning, sold as a fixed scope programme.

They also leave out the two variables. Almost all of the modern advice is about the cadence of the money: how often you release it, against what. Mine was about the distance to the work and the presence of the person who wanted it, which is to say it was never a funding mechanism at all. It was a way of measuring whether anyone meant it. The cash was never the scarce currency. Attention was, and is.

What has genuinely changed is not the consulting. It is that building your own is now far easier than it was. What took me a few evenings, a spreadsheet and a long argument with myself would now take an afternoon with an AI that will happily help you shape a crude metric, stress it, and tell you where it breaks. The tools for hand-crafting a method have never been cheaper or closer to hand, which ought to make the packaged version a harder sell.

Ought to. But the same warning applies as everywhere else. AI will not fix your mindset. Something that will help you build any method at all will just as cheerfully help you build a useless one, beautifully. What decided this was knowing what was worth measuring and why, and being willing to be unpopular about it. The clarity had to come first. Nothing supplies that for you.

And that was the point all along

So, back to where I started. The fiddly little number was never really about budgeting. T, the sponsor's time, is a proxy for mindset: clarity, commitment, whether the thing genuinely matters to the person who asked for it. N, the layers, is a proxy for momentum: whether there is a short enough line between wanting and building for anything to move at all. The method was only ever the instrument. Mindset and momentum were what it was pointed at. And notice what is not in there: I did not care in the slightest which project method any of those teams used, and I never asked. That was theirs to choose.

It also only works if the fixation stays on value delivered, not on protocol, job titles and the 'busy work' that props up fiefdoms. Do that, and a quiet by-product appears: the great people actually putting in the thought and the effort become visible through the bureaucratic fog, which is usually exactly where they had been buried.

One last thing, on the name. I called it the engagement quotient at the time. A former colleague, a few years later, used it in a keynote speech and called it the 'Worthington quotient', which I don't much like, partly out of taste and partly because branding your own little metric with your own name is exactly the sort of noun selling I moan about elsewhere. So it has no name, and I never thought it needed one. But perhaps, having resurrected it in the context of mindset, method and momentum, it could have a name around that. In the meantime it is simply a question you can ask about any piece of work that is moving too slowly. How close is your money to the work, and is the person who wanted it actually in the room, physically or metaphorically? Have you got a better name for it? How about the PACE index? Let me know.