Undercover Economist

Let’s be blunt: criticism works

‘If Amazon encourages its staff to be straight with each other about what should be fixed, so much the better’

Last month’s Amazon exposé in The New York Times evidently touched a white-collar nerve. Jodi Kantor and David Streitfeld described what might euphemistically be called an “intense” culture at Amazon’s headquarters in a feature article that promptly became the most commented-on story in the newspaper’s website’s history. As Kantor and Streitfeld told it, Amazon reduces grown men to tears and comes down hard on staff whose performance is compromised by distractions such as stillborn children, dying parents or simply having a family. Not for the first time, The Onion was 15 years ahead of the story with a December 2000 headline that bleakly satirised a certain management style: “There’s No ‘My Kid Has Cancer’ In Team.”

Mixed in with the grim anecdotes was a tale of a bracingly honest culture of criticism and self-criticism. (Rival firms, we are told, have been hiring Amazon workers after they’ve quit in exasperation, but are worried that these new hires may have become such aggressive “Amholes” that they won’t fit in anywhere else.)

At Amazon, performance reviews seem alarmingly blunt. One worker’s boss reeled off a litany of unachieved goals and inadequate skills. As the stunned recipient steeled himself to be fired, he was astonished when his superior announced, “Congratulations, you’re being promoted,” and gave him a hug.

It is important to distinguish between a lack of compassion and a lack of tact. It’s astonishing how often we pass up the chance to give or receive useful advice. If Amazon encourages its staff to be straight with each other about what should be fixed, so much the better.

We call workplace comments “feedback”. This is an ironic word to borrow from engineering, because while feedback in a physical system is automatic, with a clear link between cause and effect, feedback in a corporate environment is fraught with emotion and there is rarely a clear link between what was done and what is said about it.

The story of the Amazon worker who thought he was about to be fired is instructive. A list of goals not yet accomplished and skills that need improving is actually useful. Yet we’re so accustomed to receiving uninformative compliments — well done, good job — that a specific list sounds like grounds for dismissal.

Consider the contrast between a corporate manager and a sports coach. The manager usually wants to placate workers and avoid awkward confrontations. As a result, comments will be pleasant but too woolly to be of much use. The sports coach is likely to be far more specific: maintain lane discipline; straighten your wrist; do fewer repetitions with heavier weights. Being positive or negative is beside the point. What matters is concrete advice about how to do better.

A similar problem besets meetings. On the surface these group discussions aim at reaching a good decision but people may care more about getting along. People who like each other may find it harder to have sensible conversations about hard topics.

In the mid-1990s, Brooke Harrington, a sociologist, made a study of Californian investment clubs, where people joined together to research possible stock-market investments, debate their merits and invest as a collective enterprise. (The results were published in a book, Pop Finance.) Harrington found a striking distinction between clubs that brought together friends and those with no such social ties.

The clubs made up of strangers made much better investment decisions and, as a fly on the wall, Harrington could see why. These clubs had open disagreements about which investments to make; tough decisions were put to a vote; people who did shoddy research were called on it. All rather Amazonian. The friendlier clubs had a very different dynamic, because here people were more concerned with staying friends than with making good investments. Making good decisions often requires social awkwardness. People who are confused must be corrected. People who are free-riding must be criticised. Disagreements must be hashed out. The friendly groups often simply postponed hard decisions or passed over good opportunities because they would require someone to say out loud that someone else was wrong.

None of this should be a blanket defence of Amazon’s workplace culture — which if the New York Times exposé is to be believed, sounds dreadful. Nor does it excuse being rude. But the problem is that honest criticism is so rare that it is often misinterpreted as rudeness.

In some contexts, letting politeness trump criticism can be fatal. From the operating theatre to the aeroplane cockpit, skilled professionals are being taught techniques such as “graded assertiveness” — or how to gently but firmly make your boss realise he is about to kill someone by mistake.

Scientists have wrestled with a similar challenge. As the great statistician Ronald Fisher once drily commented, “A scientific career is peculiar . . . its raison d’être is the increase of natural knowledge. Occasionally, therefore, an increase of natural knowledge occurs. But this is tactless, and feelings are hurt . . . it is inevitable that views previously expounded are shown to be either obsolete or false . . . some undoubtedly take it hard.”

Nobody likes to be told that they are wrong. But if there’s one thing worse than someone telling you that you are wrong, it’s no one telling you that you are wrong.

Written for and first published at ft.com.