Why does bad news get softer as it travels up?

The mechanism has a name

A group project manager responsible for a collaboration product, covering 100–150 engineers across ten teams, described his organisation's reporting in three words:

"Human summation at every level."

That is the whole problem in one phrase. Consider the path a piece of bad news has to travel.

An engineer notices that a dependency isn't behaving as documented. She isn't sure yet whether it costs two days or two weeks. At standup she says she's looking into it.

The team lead hears this alongside eight other updates. Writing the weekly summary, he has 200 words for a two-week sprint. "Investigating an issue with the vendor API" makes the list; the fact that his engineer sounded less certain than usual does not, because it isn't a fact and he'd struggle to justify including it.

The dev manager receives four such summaries and produces one. The vendor API line becomes part of "some integration risk, being tracked."

The head of engineering receives that alongside eleven other projects, and reports upward that the programme is on track with known risks under management.

Four people. No dishonesty at any point. Every step is a reasonable act of professional judgement. And the thing that would have been most useful to the person at the top — that an experienced engineer is uneasy and can't yet say why — was eliminated at the first step, because it wasn't the kind of thing a status format can carry.

Why summarising removes doubt specifically

Compression doesn't discard information at random. It discards according to a rule, and the rule is roughly: keep what you can defend.

Facts are defensible. "The vendor API returns malformed responses under load" survives every level, because anyone can check it. Uncertainty isn't defensible. "I have a bad feeling about the integration" invites the question why?, and if you can't answer, you look unreliable. So it gets dropped, or converted into something firmer and smaller.

This matters because uncertainty is the early signal. By the time a concern has hardened into a defensible fact, it is usually also a schedule impact. The window in which you could have done something cheap about it has closed.

Then there is the deliberate part

Most softening is structural. Some isn't. A project manager we interviewed was candid about the proportion:

"Ninety per cent of the time it's a direct update. Ten per cent embellished if the team is almost shipping the thing — leading to sugar coating."

Ten per cent sounds tolerable until you notice where it lands. It isn't spread evenly across projects — it concentrates entirely on the ones that are struggling, which are the only ones you needed the report for in the first place. The reports you can trust are the ones about projects you didn't need to read about.

A developer put the incentive plainly:

"Everyone wants to be green. Executive sponsor wants to look good."

And an engineering leader described what happens when the culture makes that incentive sharper:

"A culture where everything was green. You'd get raked over the coals for saying something was red."

Why the format makes it worse

The traffic-light status that most organisations use turns a continuous quantity into three buckets. A team lead who was 90% confident last month and is 70% confident today has no way to express that. Green is now an overstatement; amber, in most organisations, means something quite specific and expensive.

"Switching a ticket from green to yellow is such a big step change that it sets off all the alerts. There is a cost to calling out problems." — project manager

Faced with a choice between overstating confidence and triggering an escalation, most people wait another week to see whether it resolves itself. Often it does. When it doesn't, they've lost a week — and the same logic applies again next week, which is how a project stays green until a fortnight before the deadline.

What actually shortens the path

You cannot fix summation by asking people to summarise better. Every level has a legitimate reason to compress, and the person doing it is doing their job correctly.

What works is removing the need for the signal to pass through those levels at all.

If everyone on the team answers the same question directly — how confident are you that we'll hit the goal? — and the answers are aggregated mechanically rather than editorially, no summarisation happens. What the person at the top sees is a distribution and a direction, derived from the same inputs the team lead would have compressed.

Two properties make this work where a status field doesn't:

• It captures degree, not category. Moving from 4 to 3 expresses growing unease without asserting a fact you'd have to defend, and without triggering an alarm.
• Aggregation is arithmetic, not judgement. An average has no professional reputation to protect and no reason to round anything off.

It also surfaces something summarisation destroys completely: disagreement. When one person on a team of six is markedly less confident than the rest, a summary will smooth that to the mean. Kept intact, it's often the most valuable thing on the page — one person who has looked closely at the thing about to break.

How Genchi does this

Genchi collects a one-to-five confidence rating from every team member at a cadence you choose, through Slack, email or the app, in about two seconds. Responses are aggregated on our servers before anyone sees them, so nobody's individual answer is attributable.

What a leader sees is the distribution across the team and the trend across recent check-ins, for every initiative, without anyone having written a report. The path from the person who knows to the person who needs to know is one step.