A good year is a poor diagnostic

There is a particular kind of meeting that happens after a strong quarter, and we want to describe it accurately, because most of us have chaired one.

The numbers are up. The chair is pleased. Somebody presents a slide with an arrow on it. There is a short discussion of what drove the performance, in which several different explanations are offered, all of them plausible, none of them tested. Then the meeting moves on to the operational items, and the operational items are where the difficulties live, and by then there are eleven minutes left.

The thing to notice is that nobody lied and nobody was lazy. It is simply that a good number closes conversations. It is very hard to sustain a hard question in a room that has just been told it is winning, and the person who tries usually finds that the cost of pressing is higher than the cost of letting it go. So they let it go. We have let it go.

This is what we mean when we say a good year is a poor diagnostic. Growth is a generous host. It covers process debt, because there is money to hire around the broken process. It covers concentration risk, because the concentration is currently paying. It covers products that are winning for reasons nobody has actually checked. And it covers people problems, because in a growing business the frustrated ones can be promoted sideways into the new structure rather than talked to.

None of that is hidden. It is all in the business. It is simply not visible in the number everybody is looking at, and the number everybody is looking at is going up.

The case we want to start with is one that most people listening will have shopped in.

Nakumatt was, for a time, the most impressive retail story in East Africa. It grew from a modest operation into a chain of dozens of stores across Kenya, Uganda, Tanzania and Rwanda. It was a genuine regional champion. For a lot of us it was simply where you went, and there was real pride in the fact that it was ours.

By January two thousand and eighteen it was under administration. In January two thousand and twenty a court ordered its liquidation. Suppliers went unpaid. Landlords went unpaid. Staff lost jobs. A number of smaller businesses that had built themselves around supplying Nakumatt went down with it.

The version of this story we find least useful is the one about a single villain. The more useful reading is about the relationship between expansion and control, and it is a reading that applies to any organisation growing quickly, including several that people listening will be running.

Nakumatt expanded fast, and it funded that expansion in significant part on supplier credit and short-term borrowing. That is not unusual in retail. Retail runs on the gap between when you sell the goods and when you pay for them, and that gap is a legitimate source of working capital. It becomes something else when new stores are being opened partly to generate the cash flow that services the obligations from the last round of stores. At that point the growth is no longer producing surplus. It is producing the appearance of health, and the appearance is what everyone, including the leadership, is reading.

And here is the part that matters for this episode. Right up until quite near the end, the top line looked good. New stores. New countries. More customers through the door. Every one of those is a number that goes in the right direction on a board pack, and every one of them was true.

The controls were the thing that had not grown. Stock management, supplier terms, the discipline of knowing which stores actually made money after fully loaded costs, the governance to say no to the next opening. Those are the unglamorous functions, and in a fast-growing business they are chronically under-resourced, because the person asking for two more people in stock control is competing against the person asking for the capital to open in a new city, and the second request is far more exciting.

We would put it this way. Nakumatt did not fail because it grew. It failed because the control environment did not grow at the same rate, and the growth was loud enough that nobody had to hear the difference.

The second case makes a related point from the opposite direction, and it is important because everything in it was measured.

Wells Fargo was, for years, the bank other banks were told to imitate. Its distinguishing achievement was cross-selling: getting each customer to hold more products. Checking account, savings, credit card, and so on. The internal ambition was famously summarised as eight products per household. The metric was tracked relentlessly, reported publicly, and celebrated.

The metric was met. And it was met, in a great many cases, by staff opening accounts that customers had not asked for and did not know about.

In September two thousand and sixteen, regulators fined the bank one hundred and eighty-five million dollars. The number of accounts eventually identified as potentially unauthorised ran into the millions. The chief executive resigned the following month. The reputational damage lasted years and cost far more than the fine.

What we want to draw out is not that some employees behaved badly. It is the mechanism. Set an aggressive target, tie people's livelihoods to it, monitor it closely, and do not build an equally serious mechanism to ask how the target is being hit. You have not created performance. You have created pressure, and pressure finds the cheapest available route.

The dashboard was green throughout. That is the whole point. Every quarter the metric said the strategy was working. The information that would have told them otherwise, which was in complaints, in staff turnover, in what branch managers knew perfectly well, did not travel upward, because everything that travelled upward was cross-sell numbers.

We think there is a straightforward question a leadership team can ask itself here, and it is uncomfortable in a useful way. For each of our headline metrics: what is the dishonest way to hit this. Not because we assume our people are dishonest. Because if there is a cheap way to hit the number that does not create the value the number is supposed to represent, then over a long enough period, at sufficient scale, under sufficient pressure, somebody will find it. That is not cynicism about people. It is arithmetic.

We have talked about process and about metrics. The costliest thing growth conceals is neither.

In an organisation that is expanding, there is always somewhere to put a problem. A manager who is technically capable and corrosive to work with does not have to be confronted, because there is a new region, or a new function, or a project of sufficient importance to justify a sideways move that everybody privately understands. The conversation that should have happened does not happen. It is deferred, at no apparent cost, and the deferral is invisible because the person has been promoted.

Then growth slows. And now there is nowhere to put anybody, and every one of those deferred conversations is still outstanding, except that the person is now more senior, has a longer service record, and has been promoted twice since the behaviour was first noticed. Which means the organisation has, in effect, documented its approval.

We have each watched this happen and several of us have been the manager who deferred. It is not weakness. In a growing business the sideways move genuinely is the lower-friction option this quarter, and this quarter is what you are being asked about.

The same dynamic works in reverse, and it is worse. Strong people leave growing organisations quietly, and the growth conceals it. One good person resigning in a year when headcount rose by two hundred does not register anywhere. Nobody does an exit analysis on a business that is hiring. But the people who leave a growing company are not usually leaving for money. They are leaving because they can see something that the numbers are covering, and they have concluded that saying so is not worth the cost.

That is the single most reliable early indicator we know of, and almost nobody tracks it, because it requires asking a question the answer to which is unwelcome. Not what is our attrition rate. Attrition rates in a growing business are meaningless. Ask: of the people who left in the last two years, how many would we have wanted to keep, and what did each of them say on their way out. Read the actual words. Then ask why the same sentence appears three times.

One more thing growth covers, and it is the one that kills fastest.

Every business we know of has a concentration somewhere. A handful of customers who are most of the revenue. One channel that brings most of the volume. A regulatory arrangement, a single supplier relationship, one person who understands a critical system. Sometimes a single contract that was won years ago and quietly renews.

While the concentration is paying, it does not read as risk. It reads as a strength, and it is usually described that way in board papers. Our strong relationship with this client. Our dominant position in this channel.

It is a strength. It is also the thing that determines how long you have if it moves, and the number of leadership teams who can state that figure precisely is small. The question is not are we concentrated. The answer is yes, everybody is. The question is: if this went away in ninety days, what would we actually do, and who would we tell first.

We would suggest running that as an exercise once a year, in writing, in a good year. Not as a risk register entry, which is where these go to be filed and never read. As a real half day with the actual leadership team, on the assumption that the thing has already happened. What people find is rarely the loss itself. It is that four other things they had thought of as independent are all downstream of the same relationship.

Now the counter-example, and we have chosen one from this region deliberately, because the international cases can feel like they belong to a different kind of company.

Equity began as a building society in Kenya in nineteen eighty-four. By the early nineteen nineties it was in serious difficulty. The supervisory assessment was, in plain terms, that it was technically insolvent. That is not a slow decline. That is a business that has run out of room.

What happened next, under leadership including James Mwangi, is a case study in reading a market that everybody else had written off. The conventional view of low-income customers in Kenya at the time was that they were not bankable. Balances too small, transactions too many, cost to serve too high, no collateral. The economics did not work.

Equity's move was to conclude that the economics did not work under the existing cost structure and the existing product design, which is a different statement, and a much more useful one. So they changed the structure. They removed minimum balance requirements. They rebuilt the branch model and the account economics around small, frequent transactions. They treated customers who had been dismissed as marginal as the actual core of the business.

The result was a bank that grew to serve millions of customers, and became one of the largest institutions on the continent by customer numbers.

There are two things in that story that belong in this episode specifically.

The first is that the diagnosis was correct because the situation was undeniable. When a business is technically insolvent, nobody in the room is arguing that things are fine. The comfort has already been removed. And this is the awkward truth underneath everything we do: crisis produces clarity. Our whole argument is that you have to find a way to manufacture that clarity while you are still comfortable, because the clarity is what is useful, and the crisis is merely how most organisations obtain it.

The second is that the growth which followed was growth of a specific kind. It came from a redesigned model, and the leadership could say precisely why each new customer was profitable. That is what distinguishes real growth from Nakumatt's growth. Not the rate. The fact that you can explain the mechanism, and that the explanation survives being checked.

Before the questions, one reframe that we have found changes how leadership teams treat a strong result.

A good year is not a verdict. It is a budget.

It buys you the ability to do the thing that is uncomfortable and correct, at the only moment when doing it is cheap. The write-off you have been deferring. The restructure of the team that everybody has worked around for two years. The product line that is small, loved by three people internally, and losing money. The system migration that will be twice as expensive and four times as painful in three years.

Every one of those is easier to do in a year when the headline number can absorb it. In a bad year, the same decision is read by staff as panic, by the market as distress, and by your board as an admission. The action is identical. Only the interpretation has changed, and the interpretation determines what it costs you.

So the discipline we would argue for is this. When the year is good, spend part of it deliberately on something that will not show a return inside the cycle. Treat the surplus as the price of the decision you have been putting off, rather than as evidence that you were right to put it off.

Most organisations do the opposite. A good year is banked as confirmation, and the difficult item moves to next year's list, where it will be harder and more expensive and read as weakness.

So we want to end with something practical, because the diagnosis is useless without the practice.

Here are three questions we think belong on the agenda when the numbers are good. Not in the crisis. When they are good, because that is the only time you can ask them cheaply.

The first. Which of our results this quarter can we explain mechanically, and which are we attributing. There is a real difference between we grew because we opened four branches in this region and here is the per-branch contribution, and we grew because of our strong brand and customer focus. The second is not an explanation. It is a decoration. Insist on the first, and you will find that some of your performance is genuinely not understood, and the unexplained part is where your surprise is going to come from.

The second. What are we currently protecting that we would not build today. This one is quiet and it does the most work. Every business carries things that exist because they once worked. Some of them are still right. Some of them are being defended because they are ours, or because a respected colleague built them, or because removing them would be an admission. Naming that list is the beginning of choosing.

The third. For each headline metric, what is the cheap way to hit it. Then go and check whether anyone is. Not as an accusation. As maintenance.

And one more thing which is not a question but a practice. Find the person in your organisation who was right about something eighteen months ago and was overruled. They exist. They are almost always still there, and they have usually stopped raising things. What they know now is the most valuable unread report in the building.

A good quarter is not evidence that the strategy is working. It is evidence that the strategy was working, in the conditions that applied, for reasons you may or may not have correctly identified. That is worth something. It is not worth what most boards spend it on, which is the confidence not to ask.