Say what happened, not what they are like

Here is the conversation almost everybody has the first time. You open with how well things are going. You mention, lightly, that there is one small thing. You say the small thing in about nine words, wrapped in four qualifiers. Then you close with how pleased you are to have them on the team.

They walk out thinking that went well, and they are right, because nothing happened. The feedback sandwich is not a way of delivering criticism. It is a way of surviving the delivery of criticism.

Feedback is not automatically good for people

Kluger and DeNisi’s 1996 meta-analysis is still the most sobering thing written about this. They gathered 607 effect sizes from 23,663 observations. On average, feedback improved performance, d = .41. In more than a third of cases it made performance worse.

Their explanation is about where attention goes. Feedback moves the receiver’s attention somewhere. If it moves onto the task, people work on the task. If it moves onto the self, onto what this says about me and whether I am any good, people spend their energy defending themselves and the work does not improve.

That is the whole rule, and it is a rule about your sentences, not your intentions. “You are not detail-oriented” points at the person. “The deck went to the client with last quarter’s numbers in it” points at the work.

The shape that keeps it on the task

The Center for Creative Leadership’s SBI is the best known version: situation, behaviour, impact. COIN, DESC and NVC are the same move with different labels on it. Steady uses four parts, because the fourth is the one managers most often leave out:

1. What happened. A thing you could point at. A date, a meeting, a number.
2. Why it matters. The effect on the work, the client, the team, you.
3. What you need instead. Something specific enough that they could do it tomorrow.
4. What happens next. Who does what, by when, and when you will look at it again.

Take the person who is ten minutes late to everything. Six weeks of it. Everyone has noticed.

What most people say: “I wanted to have a quick chat about timekeeping. It is not a big deal at all, and obviously you have a lot on, but a few people have mentioned it, so if you could maybe try to be a bit more on top of it, that would be great.”

What it should be: “I want to talk about start times. Stand-up is at 9:30 and you have arrived after it started on nine of the last ten days. When you come in late, Priya re-runs the update for you and the team starts twice. From Monday I need you in the room at 9:30. What is getting in the way?”

The second one is thirty words longer, contains no criticism of anybody’s character, and is impossible to misunderstand. It also ends in a question, which is where the actual conversation starts.

Ask before you conclude

Everything above assumes you already know what happened. You know half of it. HBR’s guide to talking to someone who is not meeting their goals makes getting their view first the opening move, and it is right for a self-interested reason as well as a kind one: if the deadlines are moving because they are covering for a colleague on sick leave, you would quite like to know that before you deliver your carefully prepared paragraph about ownership.

Lara Hogan’s feedback equation ends on a question for the same reason. Observation, impact, question. The question is not politeness. It is the part that gets you the information you are missing.

The three questions any feedback has to answer

John Hattie and Helen Timperley’s review of feedback in education (so: classrooms, not offices) reduced effective feedback to three questions it has to answer for the person receiving it. Where am I going? How am I going? Where to next?

Managers are usually fine on the second one and vague on the other two. “Where to next” is the single most commonly missing piece of workplace feedback. It is also the cheapest to add: one sentence, at the end, saying what you want to see and when.

Watch the words that take it back

Two habits undo good feedback in real time.

Hedges. Just, a bit, sort of, kind of, maybe, I mean, no big deal. Each one is a small apology for having spoken. Four of them in a paragraph and the message has effectively been withdrawn while you were saying it.

The word “but”. Everything before “but” is heard as throat-clearing. “Your analysis was excellent, but the deadline” is received as one piece of information, and it is not the compliment. If both halves are true, say them as two separate sentences, on two separate occasions, meaning both.

Brené Brown’s line is the one worth keeping: clear is kind, unclear is unkind. Softening the message until it disappears feels like consideration. What it actually does is transfer the discomfort from you today to them in six months, with interest.

What about the good stuff

Marcus Buckingham and Ashley Goodall argue that the whole corrective apparatus is overrated, and that people grow faster when you point at what worked. We do not buy all of it, because the deck still went out with the wrong numbers and somebody has to say so. But they are right that most managers are strikingly specific about failures and strikingly generic about success.

So be as precise with the good as with the bad. Not “great job on the client thing”. “When she pushed on the timeline you did not rush to fill it, you asked what was driving the date. That is the reason we got two more weeks.” People can repeat that. They cannot repeat “great job”.

The test

After the conversation, ask yourself one question: if I asked them tomorrow what I wanted them to do differently, could they tell me, in a sentence, without guessing?

If the answer is no, it does not matter how comfortable the conversation was. You have not had it yet.