Your best coach might actually be a problem

I've met some exceptional coaches in businesses that were absolutely terrible at coaching. That's less contradictory than it sounds. One manager has thoughtful, challenging conversations with their people every week. They prepare properly, ask good questions, know when to push and when to shut up, and somehow manage to make somebody think harder without making the conversation feel like an interrogation. Twenty metres away, another manager conducts what is technically called a coaching session by opening the CRM and asking why three deals haven't moved.

Both appear on the organisation chart as managers. Both may have attended the same leadership programme and both probably have a recurring one-to-one in Outlook. Yet the experience of being managed by them, and consequently the opportunity to develop while doing the job, can be completely different.

Most businesses know this. Ask people who the genuinely good managers are and you rarely need to wait long for an answer. Their reputations travel, people want to work for them and senior leaders point to them as examples of the culture they want. We tend to treat their existence as evidence that something is working.

I'm not sure it always is.

If one manager has become unusually good at something the organisation considers essential, while everybody else is allowed to do a wildly different version of it, their excellence may actually be exposing the absence of a system. If coaching depends on having a great coach, you don't have a coaching system. You have a great coach.

We romanticise coaching

There is something about coaching that encourages businesses to make it more mystical than it needs to be. Good coaches are described as naturally curious, emotionally intelligent, excellent listeners and instinctively good at asking the right question. All of those things can be true, but they also allow us to treat coaching as an activity where enormous variation is somehow inevitable because some people simply have “the knack”.

We don't generally tolerate that logic in other areas where consistency matters. We wouldn't be comfortable discovering that one finance manager is unusually good at paying people the correct amount while everybody else has developed their own interpretation of payroll. Nor would we celebrate a warehouse where one particularly gifted supervisor has developed a brilliant informal approach to fire safety. Where the consequences matter, organisations usually build enough structure around the work that acceptable performance doesn't depend entirely on getting an unusually capable individual.

Coaching is obviously different because human conversations require judgement. Trying to standardise them completely would solve one problem by creating a considerably more boring one. But there is an important distinction between standardising the conversation and standardising the conditions that make a useful conversation more likely. I think businesses often avoid the latter because they are understandably worried about the former.

The research gives us good reason to take coaching seriously without pretending that every intervention labelled “coaching” is equivalent. Grover and Furnham's systematic review found broadly positive effects across outcomes including performance, skills, well-being and goal attainment, while also highlighting considerable variation in how workplace coaching was defined and delivered.1 Cannon-Bowers and colleagues reached a similar conclusion in their later meta-analysis: workplace coaching has a positive overall effect, but outcomes vary according to characteristics of the intervention and its delivery.2

That nuance matters. Telling a Sales Director with 18 managers that “coaching works” isn't particularly useful. The interesting question is what those 18 managers are actually doing when the calendar says they're coaching.

The calendar can give you a false sense of security

I worked with one business where this became obvious when we started looking at what managers actually meant by coaching. The leadership team was reasonably confident that coaching was happening. It was in diaries, managers had received development around it and, if you asked whether they coached their people, almost everybody said yes. On the surface, there wasn't an obvious coaching problem.

Spend some time with the teams themselves, though, and you discovered several completely different activities operating under the same name. One manager would take a recent client conversation, ask the person to diagnose what had happened, challenge their interpretation and agree something specific to try next time. Another was essentially conducting a pipeline review, with most of the conversation spent moving through individual opportunities and deciding what needed to happen next. A third was doing something closer to problem solving: people brought them difficult situations and they gave them good answers.

All three managers were capable, conscientious people trying to help their teams perform. They weren't doing good coaching, bad coaching and no coaching. They simply had very different ideas about what the activity was for. Only one of those approaches, however, was consistently requiring the employee to develop their own ability to diagnose and solve the next problem.

That changed the conversation with the leadership team. The interesting issue wasn't which manager was “best”, or even whether managers needed more coaching training. The business had invested in coaching without ever getting particularly clear about what it expected coaching to produce. Once you see that, another workshop starts to look like a fairly optimistic response. Without enough shared architecture around the activity, you can train 18 managers in coaching and still end up with 18 different versions of it.

This is one reason I think frequency can be a surprisingly misleading measure. It is certainly better to know that conversations are happening than to discover they aren't, but the existence of an activity tells us relatively little about its quality. A business can have excellent compliance with monthly coaching sessions while employees experience almost no useful coaching at all.

There is a parallel here with Atul Gawande's argument in The Checklist Manifesto. Gawande isn't suggesting that skilled professionals can be replaced by checklists. His argument is almost the opposite: as work becomes more complex, expertise remains essential, but simple structures can protect important actions from being lost to memory, pressure and complexity.3 Aviation didn't become safer by making pilots less skilled, and medicine doesn't become safer by pretending professional judgement is unnecessary. The structure exists partly to protect that judgement when conditions are imperfect.

I think coaching needs some of the same thinking. Not a script telling a manager what to say next, but enough of an architecture that a reasonably competent manager, on a busy Tuesday, has a decent chance of having a useful conversation.

That is a much more demanding design problem than sending managers on a coaching course.

Design for the average Tuesday

When organisations design coaching approaches, there is a tendency to imagine the ideal conversation. The manager has prepared. The employee is receptive. There is enough time. The problem is sufficiently important to be interesting but not so urgent that somebody simply needs to fix it. Everybody remembers the coaching model and nobody has another meeting starting in eleven minutes.

Unfortunately, most management happens under rather less convenient conditions.

If I were designing a coaching system, I'd therefore be less interested initially in what the best manager can achieve with an hour and more interested in what an average manager can achieve with 25 minutes, a difficult employee, six overdue actions and a forecast call immediately afterwards. Reliability is partly about what survives when the conditions aren't ideal.

That doesn't require every manager to use the same seven questions. In fact, forcing experienced managers through a rigid conversational script is an excellent way to make good coaching worse. I would start instead by creating consistency around purpose. Perhaps the employee should leave a coaching conversation with greater clarity about the performance problem, having done a meaningful amount of the thinking themselves, with one or two things they're going to try and a clear point at which the result will be reviewed. Managers can arrive there in different ways.

I'd also make coaching less dependent on the calendar and more connected to events in the work. A deal that has stalled, an opportunity somebody expected to win but lost, a change in conversion, a new starter attempting something difficult or a recurring performance problem all create natural coaching moments. There is something concrete to examine and, importantly, an opportunity for the conversation to influence what happens next.

This fits with a much broader body of behavioural research showing that context and cues play a substantial role in repeated behaviour. Wendy Wood's work on habits is particularly useful here: intentions matter, but behaviour becomes more reliable when it is attached to recurring situations and environmental cues.4 Applied to management, that suggests that “remember to coach your people more often” is relatively weak behavioural design. Defining the moments that should trigger coaching gives the behaviour somewhere to live.

The structure can then remain deliberately light. What are we seeing? What might be causing it? What does the employee think is happening? What could they try next? When will we look at the result? I wouldn't give those questions a proprietary acronym or put them on a laminated card. Their value is simply that they make it slightly less likely that a manager will see a problem, immediately provide the answer and accidentally call the interaction coaching.

Advice is efficient until you have to give it fifteen times

This is where coaching becomes economically interesting.

Managers usually have good reasons for giving people answers. If somebody brings me a problem I can solve in two minutes, it can feel faintly ridiculous to spend ten minutes asking them what they think. The manager gets the satisfaction of being useful, the employee gets an answer and everybody can return to work. Michael Bungay Stanier describes this tendency in The Coaching Habit: managers easily fall into advice-giving because it is immediate, familiar and often genuinely helpful.5

The problem is that solving something quickly and reducing the likelihood that you'll need to solve it again are two different measures of efficiency. If I answer your question, we've dealt with today's problem. If I help you develop the judgement required to answer that category of question yourself, we may have dealt with several future ones too. Coaching is therefore sometimes slower at the level of the individual interaction while being considerably more efficient at the level of the system.

I've seen the cost of getting this wrong become very visible in growing sales teams. A manager who was extremely good at the job themselves became the natural destination for anything difficult. If a salesperson wasn't sure how to handle an objection, they asked the manager. If a deal was becoming complicated, the manager joined it. If somebody wasn't sure whether an opportunity was worth pursuing, the manager made the call. The team liked them and regarded them as enormously supportive, which they were.

The problem appeared as the team grew. The manager became busier and busier while the same categories of question continued to travel upwards. Their capability was solving the immediate problem so effectively that it was concealing another one: very little of their judgement was being transferred. What looked like an extremely helpful management style at six people became a bottleneck at twelve.

This is particularly easy to miss because the manager's busyness can look like evidence of their value. They are in every important conversation, everybody wants their opinion and things genuinely improve when they get involved. Eventually, though, the organisation has to ask whether the manager's job is to be the person who knows what to do or to create more people who know what to do.

Scale makes that distinction unavoidable.

A founder can answer everybody's questions when there are ten people. A Sales Director can personally intervene in the important opportunities when there are six sellers. A manager can compensate for relatively low capability by remaining close to every decision while their team is small. As the business grows, the number of possible interventions increases while the amount of senior attention available to make them doesn't.

Poor coaching can therefore create a peculiar form of organisational debt. Managers become increasingly busy solving problems that their people have never developed the capability or confidence to solve themselves. The team, meanwhile, learns that difficult judgement belongs further up the hierarchy. Eventually somebody notices that managers are overloaded and suggests time-management training.

I would probably start somewhere else.

Don't build a coaching culture. Build better defaults.

“Coaching culture” is one of those phrases that sounds excellent in a strategy document and becomes considerably harder to locate at 4.15 on a Wednesday afternoon. I think businesses would often be better served by aiming for something less grand: make a useful coaching response more likely when a meaningful performance moment occurs.

That means giving managers enough structure that an average conversation is reasonably good, then enough practice that they can depart from the structure intelligently. It means making performance visible enough that coaching can be based on something real rather than vague impressions. It means giving managers prompts, data and examples at the point they need them, and using technology where it removes friction rather than creating another system managers are expected to remember to visit.

Most importantly, it means measuring something beyond whether coaching happened. If coaching is doing useful work, people should gradually become better at diagnosing their own problems. The same issues should require less managerial intervention over time. Judgement should improve. Managers should be able to spend less time rescuing work and more time developing capability, and performance conversations should happen close enough to the event that somebody can actually do something differently next time.

None of this removes the value of exceptional coaches. Great managers will still notice things others miss, ask better questions, judge the moment more accurately and handle difficult conversations with a degree of elegance that is hard to codify. Their excellence should improve the ceiling. The mistake is allowing them to become the floor as well.

So if somebody tells me an organisation has a strong coaching culture, I'm not particularly interested in meeting its best manager. Show me the manager who is reasonably competent, extremely busy and having a fairly bad Tuesday. Show me what the system makes easier for them, what happens when they encounter a performance problem and whether their people become progressively less dependent on them for the same categories of decision.

If useful coaching still happens there, you've probably built something. If it only happens when you've got an unusually gifted manager, you haven't built a coaching system yet.

You've got some good coaches.

There is a very important difference.

RB