Every client eventually asks the same question, and they are right to ask it. What am I going to get for this, and how will I know I got it?

For twenty-five years our answer has started the same way. You should distrust anyone in our industry who hands you a clean number.

Here is what a clean number sounds like. “We improved that client’s efficiency by 30 percent.” It sounds like proof. It is not. Ask the person who said it how they isolated that 30 percent from the competitor who stumbled the same quarter, from the product line that finally shipped, and from the two strongest people in the building landing in the right jobs that year. They cannot. Nobody can. We can’t either, and we have been doing this work since 2002.

Organizational development is confounded by design. We work on strategy, structure, people, and leadership. Every one of those touches everything else the organization is doing, and all of it moves at the same time as the market, the budget, the personnel, and the competition. There is no control group. There is no second version of your company that skipped the work, running in parallel, so you can compare the two.

It is like taking a handful of vitamins every morning, eating right, exercising, and getting a full night of sleep. A month later you feel better. Now point to the one that did it. You can’t, and it would be a strange thing to conclude from that failure that none of them worked.

* * *

This leaves the whole field with two problems, and both of them are expensive.

The first is that the clean numbers get produced anyway, because clean numbers close deals. A firm arrives, runs a two-day offsite, leaves a laminated card of values behind, and sends an invoice. Six weeks later nothing has changed except the wall art. Any buyer who has lived through that once is right to be skeptical the next time.

The second problem is worse, because it costs you money quietly. Leaders conclude that if organizational development cannot be measured, it cannot be justified. So it becomes optional. Optional work is the first thing cut when the quarter gets tight, and the bill arrives later in a form nobody traces back to the decision. The silos. The decisions stacked three layers above the people who understand them. The experienced person who left because nobody ever explained how her work connected to anything.

* * *

Here is what twenty-five years of being asked this question has taught us. Most leaders asking us for metrics are not running an analysis. They are running due diligence. The question underneath the question is not “what is the expected return on this line item.” It is “how do I know you are not going to waste my money and my people’s time.”

That is a question about trust. A clean number is a bad answer to a trust question, because anyone can generate one, and the people most willing to generate one are exactly the people you should worry about.

There are better answers. Three of them, in the order of weight we give each.

1. What clients do with their own money.

Stated satisfaction is cheap. Behavior is not.

We hand every client the entire methodology. The processes, the templates, the checklists, the assessment tools, all of it. We teach them to do what we do so they can do it without us. People warned us for years that we were going to put ourselves out of business.

Twenty-five years later, our longest running client has been with us sixteen years and renewed twenty-two times. Another has been with us fifteen years across fourteen renewals. A third has renewed ten times. These are organizations that account for every penny they spend and have to prove a return on investment to somebody else in order to keep spending it. Close to 90 percent of our revenue comes from returning clients, and better than 95 percent of new work arrives by referral. Until two years ago we had no business development function at all, because we were too busy with clients to build one.

That is the number to hold any firm to. Ninety percent of our clients hire us again after being handed everything they would need to do the work themselves. They own the free substitute. They keep paying anyway.

You can dismiss a survey result. It is much harder to dismiss what an organization does repeatedly with its own budget. If you are evaluating any provider, ask for this. Renewal counts, relationship lengths, referral share. If they will not tell you, that is your answer.

2. Process metrics instead of outcome metrics.

Outcome metrics are confounded. Process metrics generally are not, because they measure the specific thing that got done rather than the organization’s overall performance.

We hear some version of this same sentence on nearly every engagement. “We have been trying to pin that process down for eighteen months, and you walked in and had it documented and agreed to in two days.”

Eighteen months against two days. That is not a rounding error, and nothing else happening in the business explains it. Price out eighteen months of a team’s partial attention, then price out everything that did not happen while the process stayed broken. Nobody has to take our word for either number, because it is the client’s own calendar.

Others in the same category. Roles, responsibilities, and relationships documented for a full team in a day. Decisions that used to travel three layers up now getting made on the shop floor. The number of approvals standing between a finished deliverable and the customer. Time to get a new hire productive, once a document exists that tells her who owns what and which relationships she has to work through to get anything out the door.

None of that requires us to take credit for your quarterly results.

3. The gap you measured on the way in.

This is the one most organizations skip, and it costs nothing.

Every engagement we run starts with an assessment, and the assessment is itself a measuring instrument. Most organizations we walk into have no formally documented roles and responsibilities. Almost none have defined the relationships between those roles, which is the half that actually prevents silos. The ones that captured roles at some point did nothing to enforce the behavior, so they finished half the job and got none of the benefit.

Count it on the way in. How many key roles have no documented responsibilities. How many relationships between functions have never been established. How many decisions have no named owner. How much work in flight cannot be traced to a stated objective. Then count it again when the work is done.

That is a before and after you own, on the specific thing you set out to change, and no competitor stumble or lucky product launch touches it. If you run your own internal development effort and take nothing else from this article, take that one. Measure the gap before you start. Most organizations never do, and that is the actual reason they cannot demonstrate anything at the end.

* * *

Then report your survey data, honestly, for exactly what it is.

Ours says that within roughly six months clients see a 35 percent improvement in traceability, meaning work that genuinely aligns to organizational goals, along with a 25 percent increase in overall efficiency and a 20 percent improvement in leadership and morale. McKinsey studied 1,700 teams across 75 organizations in 2023 and landed in close to the same place, reporting up to 27 percent gains in efficiency, customer satisfaction, engagement, and operational performance.

Those are not lies. They are also not proof. They are directional, self-reported, and aggregated, and we say so before we say the numbers. State the sample. State how it was collected. State what it cannot establish. Survey data presented honestly is more persuasive than survey data presented as physics, because the audience can tell the difference.

* * *

There is an entire category of work that gets judged this way, and organizational development belongs to it.

Nobody has ever been handed an invoice for the engine that did not fail. Maintenance intervals are not judged by what happened on the day the work was done. They are judged across years and across fleets, by what stopped happening.

Ed and I both came up flying Marine helicopters. The value of a preflight is invisible on every flight where somebody did the preflight. It is only ever visible on the one where somebody skipped it.

You judge that kind of work by what it prevents and by how long the performance holds, not by a single before and after. Organizational development is preventive work. Measure it like preventive work.

So when a leader asks us how to measure organizational development, the honest answer is that you measure it the way you measure anything that works through people and prevents problems instead of producing widgets. Track what people do, not what they say. Measure the process, not the weather. Count the gap before you start so you can count it again at the end. Report your survey data for exactly what it is and nothing more.

And stay suspicious of anybody, ourselves included, who offers you something cleaner than that.

Mike Westman and Ed Gillcrist are co-owners and co-CEOs of Shackleton Group, Inc., and the authors of the forthcoming Suits, Guitars & 3D Glasses: A Practical Guide to Corporate Maneuver Warfare®.