Problems hide in averages

Hi ,

Last week I wrote about TransDigm treating its operators like owners: tiny corporate office, real equity, real autonomy. Which raises the obvious question. With fifty autonomous business units and almost nobody at headquarters, how did anything stay on track?

This week is the other half of the machine.

Nick Howley’s partner Doug Peacock had a saying: “If you want to confuse, you conglomerate. If you want to illuminate, you disaggregate.” TransDigm turned that into a ritual. Roughly 150 product lines, reviewed every quarter, 15 to 20 minutes each. The product line manager stands up with a standard format: performance against plan, bookings by segment, pricing by market, cost projects, new business pipeline. Presidents in the room. Trust, then verify.

A few details worth stealing:

Nothing hides in a slice. A company P&L is an average, and averages are where problems go to hide. The winners subsidize the losers and the blended number looks fine. Cut the business into 150 pieces and every piece has to explain itself. When a similar product elsewhere got a 4.5% price increase and yours got 3, the question asked itself.

Bad news was fine. Sloppy thinking wasn’t. Howley was explicit about the standard: things go wrong, everyone knows that. But there was no excuse for numbers you couldn’t explain. His word for facing your own issue once a quarter was “clarifying.” Most problems compound not because nobody could see them, but because the operating rhythm never forced anyone to look.

The pipeline rule. For cost savings and new business alike, the standing rule was twice the coverage: if the plan needs $100 out, show up with $200 of identified projects, because half won’t land. Simple, and it kills the most common form of plan fiction.

The hidden benefit was people. Howley called the reviews the best integrative mechanism in the company. Watch someone explain their own numbers four times a year and you learn quickly who’s upwardly mobile, who hires well, and who’s coasting on an average.

The small-company version doesn’t need 150 product lines. It needs two moves. First, split your P&L along one honest dimension, product, customer, or channel, and look at the slices. In my experience the first pass is always uncomfortable: there’s usually a segment everyone loves that the numbers don’t. Second, build the light ritual: once a quarter, twenty minutes per slice, standard format, the person responsible walks it. Not a performance review. A facing of the numbers.

Autonomy tells your people the business is theirs. Disaggregation makes sure the truth is, too.

Talk soon,

Matt

P.S. When I’m not writing this, I’m buying and operating founder-led businesses for the long term at Eidolon Capital. If you’re a founder thinking about your next chapter, or you advise one who is, just hit reply. I read every note.