Two people from the same organization give a customer slightly different accounts of the same policy. Neither is lying and neither is careless.
The customer does not conclude that the organization has a communication problem. They conclude that they have been told two things, that one of them is the real one, and that somebody is managing them. That inference is nearly always wrong and it is entirely reasonable.
Why the assumption is concealment
People do not have access to your internal disorder. They have two statements and must explain the gap between them.
Two explanations are available. Either the organization is not coordinated, which is unflattering but harmless, or somebody is being told what they want to hear, which is not. Most people reach for the second, because it accounts for the difference in a way that mere sloppiness does not: sloppy accounts vary randomly, and the accounts people receive tend to vary in the direction of what each listener wanted.
That is not usually deliberate. It is what happens when each person adjusts slightly toward the listener in front of them, which is ordinary human behaviour and produces a pattern indistinguishable from manipulation.
Consistency in the reasoning, not the wording
The instinct is scripts, and scripts fail predictably.
They cover the anticipated questions and nothing else, so the first unusual case produces improvisation with no guidance behind it. They sound like scripts, which listeners detect and dislike. And they atrophy, because nobody updates them and the organization moves.
What travels intact is the reasoning. A person who understands why the policy exists will produce a consistent answer to a question nobody anticipated, in their own words, and it will be right. A person holding only the wording will be stuck the moment the question is phrased differently.
This is why the aligning function covered earlier in this chapter is the one that scales. Consistency is downstream of shared reasoning, and no amount of message discipline substitutes for it.
Drift over time
The version across people is visible. The version across time is not, and it is the more common failure.
A leader explains a decision in March with one emphasis, in June with another, and in October with a third, each shaped by whatever was pressing that week. Nobody hears all three except the people who have been there throughout, and they are the ones who conclude the organization does not know what it thinks.
The protection is having written the reasoning down at the time. A leader who can reread what they actually said in March will say something compatible in October. One relying on memory will reconstruct it against present circumstances without noticing.
Consistency is not rigidity
Positions should change when circumstances do. What must not happen is a changed position presented as though it were the original one.
Say that it changed and why. In March we expected the grant to renew. It did not, and that changes the answer. This is entirely defensible and preserves the credibility of everything else you have said.
What destroys credibility is the quiet substitution: the new position stated as though it had always been the position, delivered to people who remember otherwise. They now know that your account of the past adjusts to suit the present, and they will apply that knowledge to everything you tell them going forward.
The audiences compare notes
Leaders adjust the account by audience and consistently underestimate how connected those audiences are.
The version given to the board, the version given to staff, the version given to a funder, the version given to a reporter. Each is tailored, each is true, and each carries a different emphasis. In a small state these people know each other, sit on things together, and talk.
The workable test is whether the four versions could be laid side by side without embarrassment. Differing in depth and framing is fine. Differing in what they imply about the situation is not, and the reckoning is a matter of timing rather than of whether.
What to actually do
Three things, none elaborate.
Write the reasoning down once, in a few sentences, and give it to everybody who will have to explain the decision. Not a script. The two or three considerations that actually drove it.
Tell people what to do when they do not know. I do not know, let me find out is a better answer than an improvised one, and staff will only give it if they have been told it is acceptable.
Check occasionally. Ask two people in different parts of the organization to explain the same policy and listen to whether the reasoning matches. Where it does not, the fix is upstream of both of them.
Edited by Patrick J. Wolf, PhD