Where to Play and How to Win

The idea in one minute#

A strategy is two choices and nothing else. Where to play is the segment, the geography and the part of the problem you are taking. How to win is the reason a customer in that segment picks you over the alternative they would otherwise use, which is usually not a competitor. Both have to be specific enough to be wrong. The common failure is answering the first question with “everyone” and the second with a list of features, which produces a company that is the third best option for several segments rather than the obvious choice for one.

A picture#

flowchart LR
  ALL[":i-globe: <b>Everyone</b><br/><small>the non answer</small>"] -.-> THIRD[":i-triangle-alert: <b>Third best for many</b>"]
  SEG[":i-target: <b>Where to play</b><br/><small>segment, geography, slice of the problem</small>"] --> WIN{":i-crown: <b>How to win</b><br/><small>versus their real alternative</small>"}
  WIN --> PROOF[":i-badge-check: <b>Proof</b><br/><small>why it is believable</small>"]
  PROOF --> REF[(":i-ban: <b>Refusals</b><br/><small>what this choice costs you</small>")]
  class ALL neutral
  class SEG,PROOF compute
  class WIN queue
  class REF memory
  class THIRD warn

The last box is the one that makes it a strategy. Choices that cost nothing were not choices.

How it really works#

Where to play, specifically#

A usable answer names a segment you could count, and the three axes are usually who, where and which part of the problem. “Finance teams at companies between 200 and 2000 people in the UK and Ireland, for month end close only” is a place to play. It can be sized, it can be wrong, and it tells you what to build.

The test is whether somebody could draw a list of target accounts from the sentence without asking a follow up question.

How to win, against the real alternative#

The alternative is rarely a competitor. It is usually a spreadsheet, a junior hire, an agency, or doing nothing. Your basis of winning has to be stated against whatever they are actually doing, and there are a limited number of honest ones:

Being better is not on the list, because it is not a mechanism. Each row above says why you are better and why that persists, which is the part that makes it strategic rather than a claim.

Proof, and why it is the hard part#

A basis of winning needs a reason it is believable now, not aspirationally. If you win on implementation speed, what makes you fast that a competitor cannot copy in a quarter? If the answer is “we work hard at it”, you do not yet have a strategy, you have an intention. Durable answers usually come from a structural asset: an integration nobody else built, data you have and they do not, a cost structure, or a position in a workflow.

The refusals are the test#

Choosing mid market means declining enterprise deals that would distort the roadmap. Choosing one geography means saying no to revenue elsewhere. If your strategy has never cost you a deal, it is not operating. The useful artifact is a short list of what you have actually turned down, kept visibly, because it is evidence the strategy is real.

Scoring segments rather than arguing about them#

Segment choice is usually settled by whoever is most senior in the room, and it does not have to be. Four factors make it tractable: how many there are, how much each is worth, how well you can serve them with what you have, and how hard they are to reach. The program below makes the comparison explicit, and its main value is that it exposes which factor somebody is weighting without saying so.

Run the numbers#

Four candidate segments, scored on size, value, fit and reachability, against the honest alternative each one currently uses.

// segments.go: scoring where to play.
package main

import (
	"fmt"
	"sort"
)

type segment struct {
	name        string
	firms       float64 // how many exist
	annualValue float64 // what one is worth to you a year
	fit         float64 // how well today's product serves them, 0 to 1
	reach       float64 // how cheaply you can get to them, 0 to 1
	alternative string
}

// winnable is the share of the segment you could plausibly take in three years,
// which is dragged down by both poor fit and poor reach.
func (s segment) winnable() float64 { return 0.25 * s.fit * s.reach }

func (s segment) prize() float64 { return s.firms * s.winnable() * s.annualValue }

func main() {
	ss := []segment{
		{"enterprise, global", 4_000, 140_000, 0.35, 0.25, "an incumbent suite"},
		{"mid market, UK and Ireland", 11_000, 18_000, 0.80, 0.70, "a spreadsheet"},
		{"mid market, all Europe", 68_000, 18_000, 0.55, 0.30, "a spreadsheet"},
		{"small business, global", 900_000, 1_900, 0.45, 0.15, "doing nothing"},
	}
	sort.Slice(ss, func(i, j int) bool { return ss[i].prize() > ss[j].prize() })
	fmt.Printf("%-28s %8s %7s %7s %9s %13s  %s\n",
		"segment", "firms", "fit", "reach", "winnable", "three yr prize", "their alternative")
	for _, s := range ss {
		fmt.Printf("%-28s %8.0f %7.2f %7.2f %8.1f%% %13.0f  %s\n",
			s.name, s.firms, s.fit, s.reach, 100*s.winnable(), s.prize(), s.alternative)
	}
	fmt.Println("\nThe largest segment is rarely the largest prize, because fit and reach multiply.")
	fmt.Println("Anybody arguing for the big number is implicitly assuming fit and reach they do not have.")
}

Read the winnable column against the firm count. Small business has eighty times the firms of the UK and Ireland segment and produces a prize only slightly larger, because a fit of 0.45 and a reach of 0.15 multiply into a winnable share of under two percent. Enterprise has eight times the annual value per customer and finishes last for the same reason. The UK and Ireland segment has the highest winnable share on the table by a factor of three, on the fewest firms.

So the arithmetic does not say “pick the small segment”. It says the two soft numbers, fit and reach, swing the answer more than the two hard ones, and they are exactly the numbers people assert rather than measure. When somebody argues for the largest market, they are implicitly claiming a fit and a reach they have not demonstrated, and this table is how you ask them for it.

Three honest limits. Fit and reach are estimates, so run it pessimistic and optimistic and see whether the ordering survives; if it does, stop arguing. There is no cost side here, and a segment needing a year of product investment to reach that fit is not comparable to one you can serve today. And it cannot capture whether a segment is a stepping stone to a larger one, which is sometimes the whole reason to choose it.

Sector lens#

Remember this#

• Two choices: where to play and how to win. Everything else is a plan.
• A place to play is specific enough to draw a target list from the sentence.
• The alternative is usually a spreadsheet, a junior hire or nothing, and you win against that.
• Being better is not a basis of winning. Name the mechanism and why it persists.
• Fit and reach multiply, so the biggest segment is usually not the biggest prize.
• If the strategy has never cost you a deal, it is not operating.

Try it#

1. Run segments.go with your real candidates. Then set fit and reach to pessimistic values and check whether the ordering survives.
2. Write your basis of winning in one sentence, against the alternative your customers actually use. Name the structural reason it persists.
3. List the deals or segments you have turned down this year because of the strategy. An empty list is the finding.

Check yourself#

1. Why is “we are better” not a basis of winning, and what would turn it into one?
2. Fit and reach multiply into winnable share. Explain why that makes the largest segment usually the wrong choice.
3. A strategy has produced no refused deals in a year. What are the two possible explanations, and how would you tell them apart?