Methodology
Six instruments
Most commercial advice fails in the same place: the analysis is right and nobody acts on it. Each instrument answers one commercial question and ends in an artifact your team owns and can run without me.
No engagement uses all six. If a problem needs none of them, I say so in the first conversation.
1. The three-layer read
The read finds which of three layers a commercial fault sits in. Most get diagnosed one layer too high: the pipeline looks soft, so the fix is a new deck or a new hire.
- Market fit. Who the product actually wins with, on real win and loss data, not assumed segmentation.
- Motion design. How a deal moves from first contact to signature: the channel, the pricing logic, who owns each handoff.
- Operating infrastructure. The metrics that show whether the motion is healthy before the quarter closes.
What you keep: an architecture map and a prioritised roadmap a new commercial hire can execute on day one.
2. The working-assumptions table
It sits on page three, ahead of the argument, with an owner and a severity per row.
Every row states what must be true for a recommendation to hold:
- The assumption, in one sentence.
- Its status: confirmed, working assumption, or unknown.
- Its severity if it turns out to be wrong.
A high-severity unknown is never softened into a finding, which is how it surfaces in week one, not month four.
What you keep: a living assumptions log. Every row has a named owner. Statuses change as facts arrive.
3. The build, buy, or partner routing scorecard
It names the cheapest credible route to the next proof point. One page, three routes side by side, plus the fourth that usually goes unwritten: stop.
All four on the same criteria:
- Strategic fit.
- Cost and time to the next milestone.
- The capability gap that has to be closed either way.
Scoring is coarse on purpose: a defensible ranking beats a precise number.
What you keep: a scored one-pager, and a written stop condition for the route chosen.
4. Evidence-gated discovery
This is for creating something new, not fixing what exists. Four stages, with a go or no-go meeting between each. The sequence decides which ideas deserve real money and when to stop paying for the rest.
- Discovery. A long list, clustered.
- Conviction. The stage most programmes skip. A ranked idea earns a written, evidenced case before anyone spends money testing it.
- Test. That case under real pressure on desirability, feasibility, and viability. Those three are Alexander Osterwalder's framing, from Strategyzer's work on testing business ideas, used as published rather than rebadged.
- Launch. First evidence from the market.
Machines do the sweep, on tooling I build and run myself. A person makes every gate call.
What you keep:
- A scored long list.
- A validated short list, with evidence attached to each claim.
- A written kill decision for everything dropped.
5. The one-page value brief
Is the revenue real, and what is the honest range on the outcome? A market-size slide cannot answer that. The buy-side and portfolio format can.
Three gates, run in order:
- Deal killers. Fail here and the work stops; gate two never opens.
- Economics and structure.
- Conviction builders.
I score what survives dimension by dimension. Then I model it under bear, base, and upside, each with a stated probability. The value-creation half names the work and who owns it.
What you keep: one decision-ready page, with the gated assessment behind it. This is the format used for the buy-side diligence on a leading Asia-focused PE fund's acquisition, in Alex's prior advisory role. The deal closed.
6. Pilot-first 90-day sequencing
The 90-day target is a pilot running, not a contract signed, stated on page one.
Tracks run in parallel, ranked by one rule:
- Conversion probability.
- Times deal size.
- Divided by the effort.
That usually promotes the routes around procurement and demotes the one everyone expected to lead.
Enterprise cycles run longer than a quarter. The plan separates what can run in 90 days from what is a next-year event.
What you keep:
- A five-track plan, with a named owner and a dated window per track.
- A stated conversion moment further out.
Working with these
The first step is a 30-minute conversation. No deck, nothing to prepare.