ADKAR, Kotter or Lewin: which change model for which job?

A kitchen-equipment distributor with 120 staff in Can Tho has three jobs this year. Fifteen salespeople still quote in Excel although new software is in place. The owner wants to move from selling to dealers to selling to construction projects, and the department heads are doubtful. Two warehouses must become one within six weeks because a lease is ending. The firm and the three jobs are my invention. Should it use ADKAR, Kotter or Lewin, and one model for all three jobs or one for each?

Which question does each of ADKAR, Kotter and Lewin answer?

Each answers a different question, so choose by the question, not by the name.

ADKAR (ADKAR model) asks which of five elements a person is stuck at: awareness, desire, knowledge, ability, reinforcement (Hiatt, 2006). Kotter (Kotter's eight-step change model) asks what the leaders of a large change have skipped. His 1995 article lists eight common errors, each tied to a phase of the process (Kotter, 1995). The Lewin picture (Lewin's three-step model) offers a shape: loosen the old habits, move to the new way, then make the new way stick (Cummings et al., 2016).

Fifteen people ignore the quoting software: what does ADKAR add?

It splits the fifteen into groups stuck in different places. Do not assume all fifteen are stuck at the same point.

According to the library's summary, the five elements run in order, and Hiatt argues that knowledge is not the place to start: effort spent building knowledge pays off only when people want the change and are looking for the way to do it (Hiatt, 2006). Running a training day before anyone understands why may be wasted effort. Ask each person one question and listen to the answer.

In my reading, the model suits a change that has already been decided, where the question is whether each person is keeping up. It does not suit a fault in the tool itself: if the software demands 14 fields for one quote, people are stuck at the design, not at any element. The library also notes that Prosci built and publishes the model, and that the book's evidence is mostly Prosci's own benchmarking (Hiatt, 2006).

Moving to project sales: how should Kotter's list be used?

As a checklist. This is a company-wide job, and Kotter's question fits: which of the eight errors has been made? He writes that skipping steps creates only the illusion of speed, and that a critical mistake in any phase can slow momentum (Kotter, 1995).

Two cautions. Kotter's examples include large companies (Ford) and a small one (Landmark Communications), but the article says nothing specific about a firm of 120; a post on a 20-person firm takes the small-firm question further. And the article has no systematic data or sampling method. It is one person's observation of more than 100 companies, and the 75 per cent threshold is his estimate, not a measured result (Kotter, 1995).

Two warehouses in six weeks: what can Lewin and 7-S do?

Lewin helps little here if treated as a procedure. By (2005) places Lewin in the planned approach and records four criticisms aimed at that approach: it stresses small, incremental change and so suits fast change poorly; it assumes stable conditions; it ignores situations that need a directive hand, such as a crisis with no room for wide consultation; and it assumes all parties are willing (By, 2005). A six-week deadline gives the first criticism some weight. A lease ending is not obviously a crisis, so I do not apply the third.

Lewin still works as a picture: the old way has to stop running, and the new way has to become habit. But do not treat it as something Lewin tested. Cummings and colleagues point out that the three-step passage is just two short paragraphs, 131 words, in the 1951 collection, and that in the 1947 article it sits in sub-section 24 of 25, with no empirical evidence. They argue he never built that model and that it took shape after his death (Cummings et al., 2016). I report that as one group's argument, not a consensus, and the post on whether the three steps are real goes through it in full. They also recount the study at Harwood Manufacturing, in which Lewin took part: the most successful approach had all parties work through the processes and alternatives together before agreeing the way forward (Cummings et al., 2016). For two warehouses that is usable advice, although six weeks probably allows only a few representatives from each warehouse to decide together how to rearrange the stock. The "probably" is my own judgement.

The remaining question is what else breaks after the merger. That is the job of McKinsey 7-S model. Odeh (2021) describes seven elements: strategy, structure, systems, skills, staff, style, with shared values at the centre. The elements affect one another, like dominoes, when one changes (Odeh, 2021). Applied to the warehouses: are the stock codes unified (systems), who reports to whom (structure), which warehouse's people keep which jobs (staff)? In my reading it is a checklist, and it does not say what to do first.

How do you choose a model?

Start from the question you actually have.

By (2005) observes that existing theories are often contradictory, mostly lack empirical evidence and rest on unchallenged assumptions (By, 2005). The two readings of Kotter below differ, which in my view suggests the boundaries between families of theory are not fixed. By puts Kotter's 1996 book alongside Kanter and Luecke in the emergent approach (By, 2005). Cummings and colleagues list Kotter's 1995 article among the n-step guides that the Lewin heritage may have inspired, perhaps subconsciously (Cummings et al., 2016). So treat the three models as three questions to put to the people involved, not as three proven procedures.

Going deeper: what does each source cover, and what is unchecked?

The Kotter sections rest on his 1995 article in Harvard Business Review, which presents eight errors. The library's eight-step framework also draws on his 1996 book and on Pollack and Pollack (2015); this post does not use either, so it does not give the order of the eight steps from them.

The five elements of ADKAR and the note about Prosci's data come from the library's summary of Hiatt's (2006) book. Everything about Lewin's 1947 article comes through Cummings and colleagues, whose manuscript was read from a university repository rather than the printed version, so there are no page numbers. Odeh (2021) is a case analysis of Best Western Italy based on secondary data; the post does not trace the origin of McKinsey's 7-S.

What is unchecked: none of the sources used here compares the results of the three models on the same kind of change. The table for choosing a model is my reading, not a research finding.