The problem runs sideways (they asked almost eight thousand managers where execution breaks down, and it is not the axis everyone tightens)

Nobody broke a commitment

Twenty to ten on a Tuesday. The management committee of a packaging plant with six hundred employees, in Campana. On the screen, the month's eight dashboards: all eight green. On the table, printed out, the status of the project that was going to carry almost a third of the year's growth. Four months late.

“Did anyone miss a commitment?” the chairman asks.

Nobody raises a hand, and nobody is lying. Sales hit its target. Supply hit its target. IT hit its target. The only thing that was not delivered was the commitment none of the three had in writing: the one they had with each other.

There are a few seconds of silence. Then the operations director says the sentence that always gets said in that silence:

“We need to cascade the targets better.”

Nobody objects. They agree on a dashboard with more breakdown and a follow-up meeting every fortnight. The committee ends four minutes ahead of schedule, which is the only sign that everybody left reassured.

(The scene is a composite reconstruction. From here on, everything that carries a number has been measured and has a source.)

The reflex in that room does not belong to that room: it belongs to all of them. When execution stalls, the cascade gets tightened. More targets pushed down, more breakdown on the dashboard, more follow-up meetings, and if need be a committee to review the progress of the committees.

The myth is not “you have to align”. Aligning does need doing. The myth is smaller and a good deal more expensive: the idea that execution equals alignment, and that every execution problem therefore gets fixed by tightening the line that runs from the top down. That line exists, it can be measured, and there is a large survey that measured it. The problem is that it came back healthy.

The axis that does work

The cascade has one virtue it is almost never given credit for: it can be measured. And it has been.

In March 2015, three researchers published in Harvard Business Review a survey that did not ask for opinions about strategy but for something much more down to earth: whether the other person delivers on what they promise. It was answered by managers at large companies around the world, and it did not stop at the executive floor: it went all the way down to the front-line supervisor.

Start with the axis the Campana committee decided to tighten. Eighty-four per cent of managers say they can rely on their boss to deliver on what he promises, all or most of the time. The same 84% say exactly the same about their own direct reports.

This has to be said without hedging, because everything that follows rests on it. The target cascade works. The dashboard works. A year spent setting clear goals and cascading them properly was not time wasted, and this piece is not here to argue with that.

We have already written about the structure a company outgrows as it gets bigger: that is the vertical axis, the one on the org chart. This is the other one, the one that same growth breaks first and that no redesign fixes.

In an owner-run company there is a silent aggravating factor: the mechanism for coordinating across functions is the owner. It works remarkably well for as long as the conversations between areas still fit inside the owner's head, and it starts to fail when they no longer do. And it fails without warning, because it is the one mechanism in the company that leaves no record of anything.

Up and down, then, people keep their word. What is left is to ask sideways.

A reader with forty employees can close the tab here and be right about it: those are not his companies. But the mechanism being measured is not about size, it is about the number of crossings. A company of 6,000 people breaks the sideways axis with twenty functions; one of forty breaks it with three, and breaks it sooner, because it has neither the cross-functional committee nor the project office that the large one at least paid for. What follows does not describe how large companies fail: it describes how coordination fails when whoever used to do the coordinating can no longer keep up.

No more reliable than a supplier

The same managers were asked the same question turned ninety degrees: whether they can rely on colleagues in other functions and other units to deliver on what they promise. There the answer falls to 56%.

That number on its own says very little. What is uncomfortable is what it gets compared with inside the same survey: managers were also asked about partners at other companies —suppliers, distributors, contractors, people who invoice from the outside— and that trust came in at 59%. The manager down the corridor, paid out of the same payroll and carrying the same logo on his card, came in three points below the outsider. With self-reported perception, three points are a tie, and the tie is already the finding: sharing a logo buys no advantage at all.

There is a self-reported consequence pointing the same way: managers say they are three times more likely to miss a commitment because another function did not support them than because their own team failed.

The objection makes itself, and it is worth stating before your committee does: this is the same people passing judgment on themselves, and nobody answers that the weak link is his own team. True, and that is exactly why the weight does not sit in the blame that gets declared but in the comparison. The same managers, with the same bias in their own favour, rated their boss and their reports highly: the bias does not explain why sideways came out low.

And when the problem does show up, it does not get resolved where it was born either. Of every three conflicts between functions, two are handled badly: most get unblocked late and expensively, and the rest are closed quickly and badly or left to rot until somebody gives up.

That is the sideways axis, the one that appears on no line of the org chart. In the Argentine subsidiary of a multinational it also crosses a time zone, a currency and two bosses before it gets where it is going, and it still has no owner anywhere.

The four values from the same question, plotted on the two axes:

If the broken axis is that visible, the question left is why the Tuesday committee tightens the other one.

The alignment trap

The answer comes from the respondents themselves, and it cuts against this piece's own thesis. Asked what the biggest challenge to execution is, 40% answered lack of alignment; lack of coordination across functions came second, with 30%. The protagonists name first, and by ten points, the axis that already works. The committee's reflex therefore has a far more dignified explanation than laziness: it is the answer the organization gives when you ask it.

And it comes with armour nobody argues with: everyone walks into the meeting knowing that most strategies fail because of poor execution. That knowledge is the 70%, and the 70% has no study behind it: it comes from a 1999 magazine piece whose authors made clear they were estimating, and which then lost the word “estimate” travelling from citation to citation. The whole chain is in references 3, 4 and 5, and we have already taken apart the 70% of change efforts that fail. What matters for this piece is one thing only: a number with no origin does not lie, it simply does not say where the break is, and into that gap every organization drops the reflex it already had.

The problem is what happens next, and the sequence repeats. Execution stalls. The leadership team tightens alignment: more metrics, more breakdown, more follow-up meetings. That squeeze drifts, without anybody deciding it, into micromanagement, and micromanagement chokes off precisely the two things no dashboard ever asks for: trying something different, and picking up the phone to talk to your counterpart in the other function. With less of those two, sideways coordination gets worse and execution stalls further. Since nobody understands why it keeps failing, they tighten again.

The authors gave that self-devouring loop a name: the alignment trap, a spiral in which more alignment produces worse results. A manager too focused on aligning ends up, in their words, “getting better and better at answering the wrong question”. With one caveat: the spiral is the mechanism they describe, not a sequence the survey followed over time; what it measures is states, not laps.

The full loop, station by station:

If the cure makes the disease worse, what helps is not tightening better: it is writing down the one thing the spiral never touches. And that does not cost a new system. It costs one sheet of paper.

The commitment nobody wrote down

Telling one axis from the other does not take a diagnostic: it takes looking at three things that happen on any ordinary Tuesday. Deliveries arrive late without anybody having missed a target of their own. Conflicts between functions escalate to the committee instead of being settled between peers. And there is at least one coordination committee you could suspend for a month without anybody missing it.

Against the objection that always turns up here —that the last thing the company needs is more bureaucracy— there is a figure that turns it around: more than half of managers ask for more structure to coordinate across functions, twice as many as ask for it for management by objectives. The demand already exists, and it comes from below.

And there is another figure that corrects the easy reading of the first. More than 80% of companies already have at least one formal system for coordinating across functions —a cross-functional committee, service agreements, a project office— and only 20% of managers believe that system works well. The mechanism has already been paid for; what it lacks is teeth.

Here the local context is not decoration, and it is the most concrete recommendation in the whole piece. In a stable economy, a corridor agreement between Sales and Supply survives a few months on inertia; here it evaporates the week the exchange rate moves, and it does not even get recorded as a miss, because it never became a commitment in the first place. With an outside supplier none of that is up for debate: there is a purchase order, a date and a consequence. That is why, in volatile contexts, a commitment between functions has to be written down with the same formality as the target cascaded from the plan. It does not add bureaucracy: it compensates for the memory the context keeps wiping out.

The fields that commitment needs in order to actually exist:

One objection remains from whoever is not yet convinced, and it is a fair one: tightening alignment at least does no harm. That has been measured too, and not by the same people.

What it costs to align something that does not work yet

In 2007, eight years before that survey and on a different global sample, four consultants published in MIT Sloan Management Review a study of more than 500 business and technology executives. With two questions they built a grid: how well aligned the IT function was with the strategy, and how well that function actually worked.

The quadrant that gives the whole thing its name is the highly aligned and barely effective one: one in nine. Aligning a function that does not work yet cost them twice over —in what they overspent and in what they failed to grow. And the full grid shows something worse: how far it sits from the quadrant opposite, the companies that first made sure the function worked.

The same piece of work, in two different orders:

Source: Shpilberg, Berez, Puryear and Shah, MIT Sloan Management Review, 49(1), 2007 (ref. 2); same study, same three years, four quadrants. It is a study of IT, with quadrant placement self-reported by the respondent and correlational in nature: it works as a verified mirror, not as proof of the general case. The order is not an implementation detail: it is the distance between growing above the average and growing below it.

The point stands anyway. Aligning is not free: aligning before the capability works is the most expensive way to align. Two pieces of research separated by eight years, by two journals and by two objects of study ended up describing the same trap, and what does not explain itself is why the reflex both of them contradict has never once been questioned in a room.

The beliefs that hold the reflex up

The same piece of work lists five beliefs about execution and confronts them, one by one, with what the respondents answered. None of them is an oversight: all five are what the trade teaches. Two —communication equals understanding and execution should be driven from the top— rest on data this piece does not use, and are left out. The other three explain the room in Campana:

All three are learned on the job. Only one of them gets fixed with a sheet of paper and two names. Source: Sull, Homkes and Sull, Harvard Business Review, 93(3), 2015 (ref. 1).

The third explains why the sideways axis is still broken after years of knowing about it: the system does precisely what it was asked to do, which is watch your own number. So somebody else's commitment is the first thing dropped when something has to be dropped, and it gets dropped at no cost. If you want to know whether your company is in that row, you do not need a survey: in your last round of promotions, how much did a track record of collaboration weigh against the person's own numbers?

Questions for Monday

Start with the one that is not a question. If getting started requires setting up a system, nobody gets started.

The Campana committee agreed on a dashboard with more breakdown and a meeting every fortnight. Both will work, and that is exactly the problem: they reinforce the axis that was already being delivered on. At the next committee the eight dashboards will be green again, with more rows.

The operations director's sentence —we need to cascade the targets better— is not false. It is precise: it describes exactly the work of the axis that is already healthy. What nobody said that morning is that between Sales and Supply there was a commitment with a date on it, and that the date lived in the heads of two people and nowhere else. When the exchange rate moved, it evaporated without a trace. It does not show up as a miss because it never became a commitment.

One more uncomfortable question remains. If the broken axis is the sideways one, and the dashboard reinforces the other, where does a company's strategy actually live?

Not in the three-year plan, nor at the offsite, nor in the slide that was approved in December. It lives in the budget: in which unit gets the money and which one has to give it back. Part 2 opens that file and starts with the figure that orders everything else: the best predictor of this year's budget is not your strategy. It is last year's budget.