Mini-MBA Part 2: Fair process beats being right

Being right is a claim about the decision. Fairness is a claim about how the decision was made. Only one of them predicts whether anything happens.

Some years ago, in a group-level role at an industrial company, I was responsible for replacing a set of fragmented regional CRM systems with one global platform. The business case was sound. The process design had been through workshops and pilot sessions with every region. I was, on the substance, largely right.

One regional leader refused. I asked for specifics and got some, and then got the sentence that actually mattered, which I did not hear properly at the time: it always seems like we're the contingency, not the priority.

He was not wrong. The sequencing of that programme had been weighted towards the largest markets, because that is how such programmes get sequenced. His region had been consulted, and had also been told, implicitly and repeatedly, where it sat. Both were true at once. I spent the following weeks defending the correctness of the design, which was the wrong argument, because the design was not what he was contesting.

What actually transferred

1. Fair process is a design constraint, not a communication style.

The most durable idea I met on the leadership side is also the most misused. W. Chan Kim and Renée Mauborgne set out three conditions for fair process – engagement, explanation and expectation clarity – and the finding that matters is counter-intuitive: people will accept an outcome they dislike, and execute it, if they judge the process that produced it to be fair. The reverse also holds. A favourable outcome delivered through a process people consider illegitimate produces compliance at best and quiet sabotage at worst.

The version that stayed with me is the operational one developed at INSEAD by Ludo Van der Heyden, which turns the three conditions into five sequential steps and names the failure mode at each. Kim and Mauborgne tell you what fairness consists of. Van der Heyden tells you where in your calendar it goes.

The framework, explained: Fair Process Leadership

Five steps, in order. Each has a characteristic failure.

1. Seeing, engaging and framing. Set the agenda jointly. The diagnostic question is a role reversal: "what would you set as priorities if you were in my place?" – asked before you have an answer, not after. Failure: skipping straight to solutions, so people who disagree with the agenda spend the rest of the programme litigating it sideways.

2. Exploring and eliminating options. Generate multiple genuine alternatives. Your job here is impartial judge, not advocate; if you argue for your preferred option at this stage, you have ended the step. Failure: the single-option process, which readers correctly interpret as a decision already taken and a theatre performance in progress.

3. Deciding, explaining and setting expectations. State the decision, the why, the trade-off it makes and who bears its cost. Then say who does what and who gets what benefit, before implementation, not after. Failure: explaining the business case instead of the trade-off. People can already read the business case.

4. Acting and executing. Fair process demands resolute action, which surprises people who mistake it for consultation. Once expectations are set, reopening the decision privately is the betrayal – the correct line to a late lobbyist is "you should have raised this in step two; raise it at the review." Failure: the quiet reversal, which teaches everyone that the visible process is not the real one.

5. Evaluating, learning and adapting. Publicly review outcomes and admit errors. Failure: the uncorrected mistake. Most failures are experienced as unfair not because they happened but because they happened and were left standing.

Five conditions run across all five steps – the qualities that make the sequence credible rather than procedural: communication, clarity, consistency, changeability (the process can itself be challenged and revised), and culture (the same standard applies to the leader).

Scoring it. Rate each of the five steps 0–5 for your live programme, honestly, and have two other people rate it independently. Van der Heyden's field research across German engineering firms found the relationship monotonic: the fairer the process, the better the results – on revenue from new products, time to market and first-to-market share. Convergent low scores tell you where the programme will stall. Divergent scores tell you that step three never happened.

The trap for a competent executive is that consultation feels like engagement. It is not the same thing. Engagement means the affected party influenced a decision – that there is a specific, nameable thing that is different because they spoke. Consultation means they were asked. My programme had a great deal of consultation and, for the smaller regions, very little engagement, and that regional leader could tell the difference precisely because he could not name a single decision that had changed after he spoke.

The operational test is quick, and it is a ratio rather than a feeling.

The diagnostic: the engagement ratio

Engagement ratio = (leaders who can name one decision that changed because of them) ÷ (leaders affected by the programme)

Ask them, not yourself, and ask for the decision by name. Below about 0.5 you do not have a communication problem to be solved with better slides. You have a process problem, and no amount of presence training will touch it.

The ratio has one property that makes it worth tracking over a spontaneous impression: it is falsifiable, it is cheap to re-run at each programme gate, and the names in the denominator without an entry in the numerator are your resistance map – before the resistance arrives.

2. Negotiate internally before you negotiate externally.

The line that reorganised my thinking on negotiation came out of the negotiations track and applies far beyond vendors: before negotiating with the other party, the team must first negotiate internally. Most weak external outcomes are unresolved internal disagreements arriving at the table in disguise, where the counterparty reads them instantly and prices them.

The framework, explained: BATNA, reservation value and the ZOPA

BATNA – best alternative to a negotiated agreement – is what you do if this deal dies. It is not a mood; it is a specific, costed, verified alternative. You cannot improve your position at the table. You can only improve it away from the table, before you arrive.

Your reservation value is the worst deal still better than your BATNA. It is a number, decided in advance, in writing, by the team rather than by whoever is in the room at 6pm.

The ZOPA – zone of possible agreement – is the range between the two sides' reservation values. If the buyer will pay at most 100 and the seller will accept at minimum 85, the ZOPA is 85–100 and the negotiation is about who captures the fifteen. If the buyer stops at 80, there is no ZOPA and no amount of relationship will manufacture one. Recognising that early is worth more than any tactic.

Then sort every issue into three buckets:

– Distributive – a gain for one side is a loss for the other. Price, usually. This is the only bucket where the metaphor of a fixed pie holds.

– Integrative – the two sides rank the issues differently, so trades create value. You care about payment terms; they care about volume commitment. Trade them and both sides gain. Most value left on mid-market tables is here, uncollected, because nobody asked about ranking.

– Compatible – both sides want the same thing and neither has said so. Astonishingly common: a longer contract, a named account manager, a quarterly review. Settle these early; they build the momentum that makes the distributive trade survivable.

The internal step, mechanically. Before the external meeting: list every issue. Have each person on your side rank them, on paper, independently, before any discussion. Then compare. Where your own team's rankings diverge is the seam the other side will open – and they will find it within twenty minutes, because they are looking for it and you are not. Agree the reservation value, assign roles to strengths rather than to seniority, and decide who is permitted to make a concession.

The same sequence is what makes an internal change programme survivable. A management team that has not privately ranked its own priorities will present a rollout as unanimous and then leak dissent through twenty regional conversations.

3. Meet a threat by diagnosing it, not by answering it.

The most immediately useful thing in our negotiation course pack was a short piece by Katie Liljenquist and Adam Galinsky on defusing threats at the bargaining table, and it maps directly onto internal resistance.

The framework, explained: the four-move threat protocol

Move 1 – Classify before responding. Every absolute objection is one of three things: a genuine constraint (real, external, verifiable), a leverage play (real person, negotiable position, dressed as immovable), or a bluff (no substance, high volume). The three require opposite responses, and you cannot tell them apart from the tone of voice.

Move 2 – Express understanding first. Not agreement. Understanding. This costs nothing and removes the escalation that makes classification impossible.

Move 3 – Ask, do not rebut. "What specific constraint is driving that conclusion?" A genuine constraint answers immediately and in detail. A leverage play answers vaguely. A bluff changes the subject. One question does the classification work of a week of speculation.

Move 4 – Name the move, calmly, if it is one. "It sounds like this is being framed as impossible so we stop asking." Said without heat, this strips most of the force from a positional threat while leaving the other party a face-saving exit.

The rule underneath: rebutting a stated constraint, even correctly, converts a technical disagreement into a contest over whose word counts. You may win it. You will not get the execution.

In my case the objection was that a local legal framework made the platform impossible. My instinct was to rebut it, because I believed it to be untrue and could name counter-examples. The better move was to treat it as a diagnosis problem: ask what specifically was blocked, get the answer in writing, and put it in front of the vendor and the local lead together. Where the constraint is real you have learned something important. Where it is a leverage play, the request to specify it dissolves it without anyone losing face.

What did not transfer

Simulation confidence. Organisational change simulations are excellent, and they mislead in one respect: they can be rerun. You cannot rerun a quarter, and the sequencing instinct you build against a resettable model does not price the fact that a leader you lose in month two is often lost for two years.

Presence without process. Communication and leadership-presence training is real, and it is the most dangerous thing to bring home first. Polished delivery of a decision people consider illegitimate does not reduce resistance; it raises the temperature, because now the process is unfair and the messenger is smooth. Fair process first, presence second. In that order it compounds. Reversed, it corrodes.

The workshop illusion. Alignment is not produced by an offsite. It is produced by a cadence – a standing forum, a written decision record, and visible consistency between what was decided and what happens next. Goodwill generated at an offsite does not survive the first inconsistent decision after it. This is step four of the fair process model, and it is where most well-designed programmes actually die.

And one thing about myself. In my first genuinely non-executive setting I answered three questions that were not mine to answer before I noticed I was doing it. The instinct to solve is an asset running a company and a liability in a room whose job is oversight, where it crowds out the capacity to listen properly and to frame a matter as a question rather than as a recommendation with a decision already attached. The most useful correction I have found is exactly that discipline – present the situation, the challenge, and the question, then let the room answer. Unlearning a reflex is slower than learning a model, and it is the part that decides whether a competent executive becomes a useful director.

What to do Monday morning

Three actions for the operator whose correct plan is stalling.

Compute the engagement ratio for your live change programme – list the affected leaders, ask each of them to name one decision they influenced, and count. Before your next external negotiation of any consequence, run the internal one first: BATNA established away from the table, reservation value written down, every issue sorted into distributive, integrative and compatible, and every team member's ranking on paper before the discussion. And the next time someone hands you an absolute constraint, ask one question instead of giving one answer, then get the answer in writing.

The five numbers for the change-readiness dashboard

One page, per programme, reviewed monthly.

1. Engagement ratio: affected leaders who can name a decision they changed, over affected leaders.
2. Fair Process Leadership score, five steps rated 0–5, scored independently by three people – and the spread between them.
3. Number of internal alignment sessions held before the last significant external negotiation, with the reservation value written down before the meeting.
4. Adoption ninety days after go-live, as a percentage of licences or capacity paid for.
5. Ratio of questions to statements in your last leadership meeting – have someone count.

The fifth is the one people resist measuring, which is why it is worth measuring. A management team whose leader speaks in statements will bring him statements, and he will be the last person in the building to learn that the plan is not moving.

That programme did eventually land, later and differently than designed, and the difference was almost entirely procedural rather than technical. Learning it took an expensive detour through a classroom. What the classroom cost, and what it actually bought, is the arithmetic almost nobody publishes – and it is where part three ends this series.

Photo: Unsplash