Give every market one bet a quarter. Watch the network compound.

The last post made the case for playbooks over meet-ups. But a library of plays only matters if the network keeps generating new ones — deliberately, not by accident. And this is where most distributor networks leave the biggest prize on the table.

They have ten, twenty, fifty markets, each perfectly capable of trying something. And they run, in effect, one experiment at a time: central, cautious, and slow.

The problem with one central roadmap

Headquarters builds a plan and rolls it out everywhere at once. If it works, wonderful. If it doesn’t, the whole network just paid for the mistake in unison. Either way you learned one thing, slowly, and you treated a diverse set of markets as if they were a single one. They aren’t — which was rather the point of the responsibility boundary I started this series with.

One market, one bet, one quarter

Here’s the alternative I’ve run with management teams. Each market commits to one focused bet for the quarter. Not five. One clearly defined thing it will try — a new segment, a new offer, a new channel — chosen because it fits that market. Ninety days. Then it reports back in a shared format.

Do the arithmetic. Ten markets, one bet each, and the network is suddenly running ten experiments at once: in parallel, in real conditions, at no extra cost. That’s a portfolio, not a roadmap.

A network of ten markets can run ten experiments at once. Most run one, and call it a strategy.

The single bet is the feature, not the limitation

Restricting each market to one bet feels stingy. It’s the most important rule in the whole system. One bet forces a real choice about what matters most this quarter, and it keeps the result legible — you can actually tell whether the thing worked.

Harvard Business Review’s work on disciplined experimentation is blunt about this: a test is only worth running if you’ve committed in advance to act on the result, and if you’re willing to walk away when it fails. Five simultaneous bets per market give you neither the focus to act nor the clarity to walk away.

The report-back is the real infrastructure

The magic isn’t the experiment. It’s the report-back — same short format, every quarter, win or lose. What we tried, what happened, what we’d tell the next market that tries it. That single document is the unit that turns a local result into network knowledge.

McKinsey’s channel research describes the same discipline from the manufacturer’s side: run pilots to cut through analysis paralysis, then hold regular reviews that feed the lessons back into the model. The format is boring, and it is the entire game.

Promoting a winner

When a bet works, it doesn’t get a round of applause and an email. It gets promoted: written up as a play, and handed to two or three more markets to run the following quarter — to check whether it travels. If it works in three, it becomes a network standard. Within a couple of quarters, a good idea from one country is running everywhere it fits, verified rather than mandated.

What kills it

Three things, reliably. Too many bets, so focus dissolves and nothing is legible. No report-back, so experiments become private hobbies and the learning never leaves the market. And punishing the misses — the single fastest way to stop honest reporting is to treat a failed bet as a failed manager. A bet that fails and is reported clearly is a win for the network. Protect that, fiercely.

Six months of this and the network runs differently — not because headquarters imposed a new plan, but because the markets built one together, in the open.

Which raises the obvious question: who holds all of this together, and stops it collapsing back into a pile of unread reports? That’s the last post in the series.

Give every market one bet. Let the network place ten.

Sources & further reading

External
Harvard Business Review — The Discipline of Business Experimentation
McKinsey — A new dawn for industrial channels

Related posts
Where headquarters ends and the distributor begins
Learning with one another, not from one another