You can be completely right and still lose. The proposal that dies in the room rarely dies on its merits, it dies because the person who could have killed it was never spoken to, the team it affected first heard about it from the announcement, and the one ally who mattered wasn't sure enough to spend their credibility on it. Building a coalition is the unglamorous work that prevents all three.

The quick version

  • A coalition is the group of people, across functions and levels, who will actively back a change, not just tolerate it. Buy-in is what you've earned when they'll spend their own credibility to support it.
  • One sponsor is not enough. John Kotter's research on change names "build a guiding coalition" as a separate step precisely because solo champions stall.
  • Start by mapping who holds power and who holds interest, then engage each group differently, you manage the powerful-and-interested closely and keep the powerful-but-disengaged satisfied.
  • Resistance is usually rational. Kotter & Schlesinger identified four honest reasons people push back, and the fix is almost always to involve them before you've decided, not to sell harder afterwards.

The idea in depth

The most quoted argument for coalitions comes from John Kotter, whose 1995 Harvard Business Review article and 1996 book Leading Change set out an eight-step model for organisational transformation. The second step, right after creating urgency, and before any vision or plan, is "build a guiding coalition." Kotter's point is structural, not motivational: major change overwhelms what any single executive, however senior, can drive alone. You need a group with enough position power (so opponents can't easily block it), expertise (so decisions are sound), credibility (so people take it seriously), and leadership (so it actually moves). A title alone supplies only the first of the four.

So the move is to stop asking "how do I convince my boss?" and start asking "who are the five to ten people whose visible support would make this hard to stop?" Then recruit them as a group, early, before the proposal is finished, because people defend what they helped build and resent what's handed to them complete. The coalition isn't an audience for your idea; it's a co-author of it.

An honest limitation. Kotter's model is one of the most cited frameworks in management, and also one of the most criticised. It was built from his own observation of roughly 100 companies, not from a controlled study, and the "70% of change efforts fail" figure often attached to it has been traced by academic reviewers, notably Mark Hughes of Brighton Business School, to no reliable empirical source at all. Treat the eight steps as a well-worn checklist of what tends to matter, not as proven law. The coalition step survives that scrutiny better than most, because it restates something older and sturdier: you cannot impose change on people who can quietly refuse to deliver it.

Map the room before you work it

"Get the right people on side" is useless advice until you know who the right people are. The most practical tool here is the power–interest grid, usually credited to Aubrey Mendelow's work on stakeholder analysis from the early 1990s (the exact citation date is muddled in the literature, some sources say 1981, some 1991, so credit the idea, not a precise page). It plots every stakeholder on two axes: how much power they have to help or block you, and how much interest they have in the outcome.

quadrantChart
  title Power vs interest, who to engage, and how
  x-axis "Low interest" --> "High interest"
  y-axis "Low power" --> "High power"
  quadrant-1 "Manage closely"
  quadrant-2 "Keep satisfied"
  quadrant-3 "Monitor"
  quadrant-4 "Keep informed"
  "Sponsoring exec": [0.78, 0.85]
  "Finance lead": [0.30, 0.80]
  "Affected team": [0.82, 0.30]
  "Adjacent dept": [0.25, 0.28]
The power–interest grid (after Mendelow): four groups, four different jobs. Leaders Loop

The grid earns its keep because it tells you what to do with each group, not just who they are. High power and high interest, your sponsoring executive, the budget holder who actually cares about the outcome, you manage closely: these are your coalition core, and they deserve real time. High power but low interest, the finance lead who can veto on cost but isn't paying attention, you keep satisfied: no surprises, brief them early, never make them learn about your plan in a forum where they feel ambushed. High interest but low power, the team whose daily work changes, you keep informed and listen to, because they hold the practical knowledge that makes or breaks delivery. The rest you simply monitor. The classic failure is treating everyone the same: drowning the disengaged-but-powerful in detail while neglecting the affected team whose quiet non-cooperation will sink you.

People defend what they helped build and resent what's handed to them finished.

An honest limitation. A grid is a snapshot, and politics moves. Power shifts with reorganisations, interest spikes the moment a change touches someone's budget, and a "keep satisfied" stakeholder can become your fiercest opponent overnight if they feel cut out. Redraw it as the situation changes, and never mistake the neat quadrants for the real, shifting relationships underneath.

Resistance is information, not an obstacle

The instinct when someone pushes back is to argue harder. The evidence says argue less and ask more. In their 1979 Harvard Business Review article "Choosing Strategies for Change," John Kotter and Leonard Schlesinger set out four honest reasons people resist: a fear of losing something they value, a misunderstanding of the change and its implications, a genuine belief that it's a bad idea, and a low tolerance for change itself. Notice that only one of those is "they don't get it", the rest are either self-interest you haven't addressed or a judgement that might be correct.

Kotter and Schlesinger then lay out six approaches, from education and participation through to negotiation and, at the blunt end, coercion, and they're clear that the right choice depends on the situation. But the one they single out as most powerful for building durable buy-in is participation and involvement: bring potential resisters into designing the change, and they tend to commit to it rather than fight it. So the move is to treat the first objection as data. Ask "what would have to be true for you to back this?", and mean it. Sometimes the answer reveals a flaw in your plan you can fix; sometimes it reveals a fear you can address; either way you've converted an opponent into a contributor, which is worth more than a silenced one.

flowchart TD
  A(["Someone resists
your proposal"]) --> B{"Why? (Kotter & Schlesinger)"}
  B -->|"Losing something"| C(["Address the loss,
or negotiate"])
  B -->|"Misunderstanding"| D(["Educate &
communicate"])
  B -->|"Thinks it's wrong"| E(["Involve them,
they may be right"])
  B -->|"Low tolerance for change"| F(["Support, pace,
reassure"])
  C --> G(["A contributor,
not an opponent"])
  D --> G
  E --> G
  F --> G
Diagnose the resistance before you respond to it, different causes need different moves. Leaders Loop

A worked example

Take a head of operations, call her Priya, who wants to move her company's support team from email to a proper ticketing system. (Illustrative scenario; the people and figures are made up to show the method.) The business case is obvious to her: faster resolution, less duplicated work, real data on what customers actually struggle with. Six months ago she'd have written it up, presented it at the leadership meeting, and watched it get "parked for further discussion", which is where good ideas go to die quietly.

This time she maps the room first. The CFO has high power, low interest: he'll wave it through if the cost is modest and there are no surprises, so she sends him a one-page cost summary a week early and asks nothing of him in the meeting. The support team lead has high interest, low formal power but enormous practical power, if his people don't adopt the tool, it fails, so she doesn't present to him, she co-designs with him, letting him shape which workflows move first. The one predictable resister is a senior agent who's been there fifteen years and built his own email-folder system; rather than overruling him, Priya asks what he'd lose and what would make it worth it. His answer, that his hard-won shortcuts would vanish, surfaces a real gap, so she has him define the system's saved-reply templates. He goes from blocker to the person training everyone else.

By the time the proposal reaches the leadership meeting, it isn't Priya's idea anymore. The CFO has already nodded, the team lead presents the rollout plan himself, and the senior agent is quoted as a supporter. The decision takes four minutes. That speed is the visible tip of weeks of invisible coalition work, and it's the opposite of how it looks from the outside, where it appears she simply had an idea everyone happened to agree with.

Frequently asked questions

Isn't this just office politics dressed up?

It's politics in the neutral sense, understanding where power and interest sit and working with them honestly, not manipulation. The line is intent and transparency: a coalition built by genuinely involving people, addressing real concerns, and improving the plan is the opposite of a backroom stitch-up. If your "buy-in" depends on people not knowing the full picture, you've built something fragile that will collapse the moment they find out.

What if I don't have the authority to assemble a coalition?

Coalition-building is most valuable precisely because you lack the authority to simply mandate things, it's how influence works without a hierarchy behind it. You don't need permission to map stakeholders, ask a respected colleague for early feedback, or bring an affected team into shaping a proposal. The whole point is that borrowed credibility and shared ownership substitute for formal power you don't have.

How big should a coalition be?

Big enough to be hard to stop, small enough to move. Kotter suggests the group needs real position power, expertise, credibility and leadership between them, which for a meaningful change usually means a handful of people, not a committee of thirty. Aim for the smallest group that covers those four bases plus the affected parties whose cooperation you need. A coalition that's too large stops being a coalition and becomes another meeting.

What do I do about someone who resists no matter what?

First, diagnose honestly using the four reasons, a persistent resister often values something specific you haven't addressed, or genuinely thinks you're wrong (and may be). If real involvement and negotiation don't move them, you don't always need them: a strong coalition can carry a change past a lone holdout, provided that person can't block it outright. The mistake is letting one immovable objector hold the whole thing hostage when the rest of the room is with you.

How is this different from just being persuasive?

Persuasion is one-to-one and happens in the moment; coalition-building is one-to-many and happens over time, before the moment. A persuasive case can win a single conversation, but it evaporates if the person you convinced has no stake in the outcome. Coalitions create durable backing by giving people ownership, which is why they survive the meeting, the reorganisation, and the inevitable first setback.

Related in the Toolkit

Coalition-building is influence applied to groups, so it sits next to the skill of moving people when you can't simply tell them, influence without authority, and rests on the underlying mechanics of why people say yes, which Cialdini's persuasion principles map out.

Where to go next