The state of change 2026: change capability is systemic.
IMA Worldwide is the firm behind the Accelerating Implementation Methodology (AIM), developed by founder Don Harrison. Peacock Hill Consulting acquired IMA Worldwide in 2024. · By Ann Marvin · Updated August 2026
What 38,757 assessment records across 409 organizations reveal about their history of successful change. The finding is not that organizations are bad at one thing. It is that the same pressures recur nearly everywhere, and the conditions move together.
The Implementation History Assessment (IHA) is IMA Worldwide's diagnostic, measuring the climate an organization brings to a change before it is deployed. Unlike surveys that ask how people feel about a change, it scores what happened to previous ones. Across 13 industries and three decades the same three conditions come out weakest: no involvement, no room for the change, no reward for the new way.
Everything below, in one minute
What the numbers say.
Three decades of assessments, and the shape of the answer barely moves. Organizations are reliably good at the same things and reliably weak at the same things, and it has almost nothing to do with what business they are in.
Across 38,757 assessment records and 409 organizations, from 1995 to 2026, five things hold up.
- 90% of organizations rate one of two conditions highest: Structure 57%, Personal Belief 33%.
- 90% of organizations rate one of three conditions lowest: Involvement 38%, Organizational Stress 35%, Reinforcement 17%.
- Industry does not change this.
- When sponsorship is higher, all 10 other dimensions are higher too.
- Only 5.5% of assessment records reach the High implementation-readiness band.
The results
The Implementation History Assessment
Harrison rates it above everything else IMA Worldwide runs. If I'm limited to one assessment in an organization, it's going to be this one.
His reason is that it is the only one that reads the organization's own track record back to it.
Every organization here came to IMA Worldwide for help with a change. Before that work begins, their people take the Implementation History Assessment. It does not ask about the change they are facing. It asks how change has gone here before.
People are asked to name up to three major changes from the past three years, then answer 50 questions about how those went.
The reasoning is blunt: people repeat the strategies they have used before, so past problems are likely to be repeated. The overall score is a forecast, the probability that the next implementation succeeds.
IMA Worldwide judges an implementation on five things: on time, on budget, and the technical, business and human objectives it was meant to achieve. Organizations already push hard on the first three. Harrison's argument is that the last two decide whether it worked, and the human objectives are the ones he pushes hardest of all.
The assessment asks about four of those five, scored on how the last changes went: whether they landed on time, stayed within budget, and met their technical and human objectives.
Each question is a statement, and people say whether they agree or disagree with it. Agreeing means people say that condition has been in place here. Whether that helps or hinders is the second half of this page.
Where organizations land
Everything adds up to a single score out of 100. The guide calls it the Implementation History Factor. The higher it is, the better the odds that the next change takes hold. Scores fall into four groups, from Very Low to High. Here is where all 38,757 records land.
Across 38,757 assessment records collected between 1995 and 2026, the average is 61.0, one point above the line where the odds stop being poor and start being middling. Just under half of those records fall below that line, and only 1 in 18 reaches the top group.
Every condition, scored
This is the guide's own scoring page. Every section gets its own bar running from 20 to 100, the dashed lines mark where one group ends and the next begins, and the marker is where the average lands.
Not one of the eleven reaches the top group. The best of them, Structure at 67.4, sits mid-table. Three fall below the halfway line into the second-lowest group: Involvement at 56.3, Organizational Stress at 57.5 and Reinforcement at 58.7.
That is the average organization's starting position for its next change.
Harrison's reading of a score like that is not a grade. It is a schedule and a budget. When confidence is low, resistance is higher in the organization,
he tells teams walking through their own results, which means change is going to be slower and it's going to take more resources.
If a world class electronic medical records rollout takes eighteen months, an organization in the low bands should not assume eighteen months.
Your strength is always your resistance.
Don Harrison, on reading an implementation history assessment
Where people disagree most
The score above is the average. This is the spread behind it: for each condition, the share of people who disagreed that it was in place, and the share who agreed.
The rest answered neither, so the two never add to 100. Nothing clears 54%. Even the strongest condition has a quarter of people disagreeing, and on the bottom three more people disagree than agree.
Which condition comes out top, and which comes out bottom
Rated highest
Which condition comes out top, across 156 organizations
- Structure 57%
- Personal Belief 33%
- anything else 10%
Rated lowest
Which condition comes out bottom, across 156 organizations
- Involvement 38%
- Organizational Stress 35%
- Reinforcement 17%
- anything else 10%
Of all 11 conditions, only three have more people disagreeing than agreeing. They are the same three that come out weakest.
Split the data any way you like, by industry, by decade, by organization, and the same conditions keep landing in the same places.
Does your sector change it
Industry does not change this.
We can identify 13 industries. For each one we took the two conditions people rate lowest. Almost every industry lands on the same short list.
The weakest two, by industry
13 industries, grouped by which two conditions come out lowest
- Involvement + Organizational Stress 8
- Involvement + Reinforcement 3
- Reinforcement + Organizational Stress 1
- Implementation History + Involvement 1
Every industry draws its weakest two from the same four conditions. Eight of the thirteen land on exactly the same pair.
| Industry | The two rated lowest |
|---|---|
| Healthcare | Involvement, Organizational Stress |
| Pharmaceutical | Involvement, Organizational Stress |
| Insurance | Involvement, Organizational Stress |
| Technology | Involvement, Organizational Stress |
| Oil and Gas | Involvement, Organizational Stress |
| Telecommunications | Involvement, Organizational Stress |
| Retail | Organizational Stress, Involvement |
| Financial Services | Organizational Stress, Involvement |
| Government | Involvement, Reinforcement |
| Aerospace | Involvement, Reinforcement |
| Energy | Involvement, Reinforcement |
| Utilities | Involvement, Implementation History |
| Manufacturing | Organizational Stress, Reinforcement |
Line up the 13 industries from highest to lowest and the top and the bottom are only about 5.1 points apart, 57.9 against 63.1.
Now line up the 156 organizations with at least 30 records instead, and the top and the bottom are 28.8 points apart, 43.9 against 72.7. More than five times the spread.
So your industry tells you almost nothing about your score. The organization you work for tells you a great deal.
Who you work for matters. What sector you are in barely does.
What everything hangs on
When sponsorship is higher, the other 10 are higher too.
We compared the 4,297 assessment records where respondents disagreed that leaders were behind the change with the 7,407 records where they agreed. Then we scored everything else for each group, the way the guide scores it.
Every condition crosses a band. With leaders not behind the change, all ten sit in Very Low or Low. With leaders behind it, all ten sit in Moderate or High. Not one of them overlaps.
The eleven are not eleven separate dials. Take all 55 possible pairs of them. All 55 are positively correlated. Even the weakest relationship, between structure and agent capacity, is 0.50 at the reported precision. Structure sits at the edge of that web and has the weakest relationships with several of the other conditions.
Is this just gloomy people marking everything low? Partly, and we cannot rule it out.
But a bad mood would lift every condition by the same amount, and it does not. The smallest gain is 26.0 points, on reinforcement. The largest is 38.6, on personal belief.
Look at reinforcement against personal belief. Where leaders were not behind the change, reinforcement scores higher, 44.7 against 44.4. Where leaders were behind it, belief reaches 82.9 and reinforcement only 70.7, the lowest of the ten. A mood cannot flip an order.
This is the through-line Don Harrison has pointed to for three decades. The condition that moves everything else is whether leaders are visibly behind the change: of the eleven conditions it tracks most closely with whether people believe the change will land, more closely than any of the other ten, and where it is present all ten other conditions rise with it, without exception.
AIM treats leader-driven implementation as the primary lever for exactly this reason. The data does not prove leadership causes the rest, and it records the conditions people reported, not whether a change succeeded. What it shows, across every era, is that nothing else in the instrument travels as widely.
What it is for, and what it is not
The assessment has one job.
The Implementation History Assessment is a baseline diagnostic. It is run once, at the start of a piece of work, with the people a specific change is about to affect. Its job is to show where that change is most at risk, so effort goes there instead of everywhere.
Everything on this page is those baselines, pooled. That is a large and unusual body of evidence, and it is also a particular kind of evidence. It answers some questions well and is not built to answer others.
| Not built to answer | Because |
|---|---|
| Whether a given change went on to succeed | The assessment runs before the work, so the work can be aimed at the weakest conditions. What happened next was shaped by that work, which is the point of running it, and also why these scores cannot grade their own forecast. No number on this page is a success or failure rate. |
| Whether organizations are improving over time | Each assessment is commissioned for one organization, with the population a particular change affects. Two runs years apart are two different groups answering about two different situations. That is the right design for targeting work, and the wrong one for a trend line. |
| What causes what | The eleven conditions move together. Co-movement is not cause, and this design cannot separate the two. |
| How one person changes over time | Responses are anonymous by design, so that people answer honestly. The trade is that nobody can be followed from one assessment to the next. |
| How leaders compare with everyone else | The assessment is normally run across a whole affected population rather than split by level. Separate leadership cuts are commissioned occasionally, for the client's own use, and are not part of this dataset. |
| How this compares with organizations that never assess | Everyone here works somewhere that brought in outside help and chose to measure first. That is not a random sample of companies, and the bias could run either way: more deliberate than most, or in more difficulty than most. |
Where these numbers come from
How the research was done.
Every record answers all 50 questions. Nothing was estimated, filled in or left blank.
| Assessment records | 38,757, each one completed by a participant in an AIM class assessing their own organization. |
| Used for the figures on this page | all 38,757 for record-level figures. The 35,796 we could tie to a named organization are used only where a figure breaks down by organization or industry. |
| Organizations | 409 |
| Years covered | 1995 to 2026 |
| How each question is answered | 1 to 5, from strongly disagree to strongly agree. 3 is neither agree nor disagree. On this page, agreed means someone answered 4 or 5, and disagreed means they answered 1 or 2. |
| What it measures | the conditions people reported around a change, not whether the change succeeded |
What counts as an organization
An organization here is one client company, with its divisions, business units and spelling variants merged back into the parent. That gives 409 of them across the three decades.
Two figures on this page are per-organization rather than per-record: which condition an organization rates highest, and which it rates lowest. Those use the 156 organizations where at least 30 people answered, because below that a single team can decide the result.
Three things worth knowing
The questionnaire has not changed in three decades. The same eleven conditions reconcile in 990 of the 1,001 assessment workbooks, which is what makes it fair to compare the 1990s with today. The 11 exceptions are partial files holding 7 questions rather than 50.
The main finding has been checked against separate samples. A separate group of 16,897 people, assessed in 2002, came out with the same three weakest and same two strongest conditions, in almost exactly the same order. So did 905 responses collected between 2024 and 2026 on an entirely different assessment platform.
Client organizations are confidential. No individual organization's results appear anywhere on this page.
Citing this research
These figures are free to quote and chart with attribution. There is no paywall and no form. If you quote a single number, please carry its sample size and date range with it.
IMA Worldwide. (2026). The State of Change 2026: implementation readiness across 38,757 assessment records, 1995 to 2026. Implementation Management Associates, Inc.
Journalists and researchers can request a cut of the data by industry, condition or period through the IMA Worldwide contact page. Corrections are applied here and noted with the update date.
What these numbers cannot tell you
They describe conditions, not outcomes. The assessment records what people reported about the situation around a point in time in an organization. It does not record whether that change succeeded. Nothing here is a success rate or a failure rate, and none of it should be combined with published statistics about how often transformations fail, which measure something different.
They are not a trend. Records are dated by the year an organization first appears, not by the date of a survey wave, and there are no records at all for 1996, 1997 and 1999. The span from 1995 to 2026 is a description of coverage. We make no claim that readiness has risen or fallen over it.
They are client-weighted, not market-representative. These are organizations that chose to measure their implementation history and asked IMA Worldwide to do it. The largest single contributor is 7.3% of the identified records, and the ten largest are 41.5%. Read the figures as a large body of practice, not as a survey of the economy.
The conclusion
Change capability is systemic.
The eleven conditions are not eleven separate problems. They move together, and the same three come out weakest almost everywhere.
So a low score on any one of them is rarely an isolated fault. It is telling you about the environment around it.
Which raises the obvious question: if these conditions are this consistent, why has thirty years not fixed them? Harrison's answer is that they are not faults at all. Every pattern in the data is one the organization has been paying for. Every time you see a pattern of behavior replicated over and over again, somewhere there is or was a reward for that behavior.
That is also why the obvious response does not work. These are reinforced patterns, and many of them will not change simply by talking about them.
Change what gets rewarded, or expect the pattern back.
The next edition
Get next year's report
This is reissued annually at this same address, so next year's edition measures change against the baseline above rather than starting over. Leave your email and we will send it when it publishes, along with the occasional read for practitioners. Confirm the subscription in the email that follows and you are on the list. No other use, and one click unsubscribes.
Common questions
Change readiness, answered.
What conditions are associated with weaker implementation environments?
Excessive change load above all. Across all 38,757 responses, 54% of people report too many changes underway at once, and those people report weaker conditions on every one of the other ten measured. Limited involvement and weak reinforcement recur alongside it. These are associations, not proof that one condition caused another.
What is change readiness, and how do you measure it?
Change readiness describes the conditions surrounding a change before it is deployed. IMA Worldwide measures it with the Implementation History Assessment, a 50 question survey scoring eleven conditions from sponsorship and involvement to communication and organizational stress. Every question is answered 1 to 5, and the guide converts those answers to a score from 20 to 100, sorted into Very Low, Low, Moderate and High probability bands.
What is a good change readiness score?
Scores run from 20 to 100 and fall into four groups, each one describing the odds that a change will take hold: Very Low below 40, Low 40 to 59.9, Moderate 60 to 79.9, High 80 and above. Across 38,757 records the average is 61.0, one point above the line where poor odds become middling ones. Of the 38,757 records, only 5.5% reach the top group, and not one of the eleven conditions averages High.
Does my industry affect change readiness?
Very little. Industry explains only about 1% of the variation between individual scores. Involvement is among the two lowest rated conditions in 12 of the 13 industries measured. Comparing yourself to an industry benchmark tells you far less than most organizations assume.
What distinguishes stronger implementation environments?
Stronger environments are associated with visible sponsorship, credible prioritization, consistent reinforcement, accountability, involvement and sufficient capacity. Where sponsorship is strong, all ten other measured conditions are stronger with no exceptions. These relationships do not prove which condition caused the result.
When should we assess readiness?
Before you deploy. Each assessment is taken by the people involved in a particular change, so it tells you about the environment that change is landing in: where it is weakest, and which of the eleven conditions is the biggest risk. Act on the highest risks first. Assessing again for a later change tells you about that change, not about whether the organization has improved, because it is a different group of people answering.
This page tells you where everyone else lands. Interested in learning where you land?
That is what the assessment does. Tell us what you are changing and we will show you where the risk sits.
Enterprise transformation or organizational change? Talk with Peacock Hill Consulting, which applies AIM diagnostics, training and consulting to enterprise transformations. IMA Worldwide is a Peacock Hill Consulting company.
What is AIM? · Why transformations fail · Reinforcement in change management