Part II · Strategy before tools · Chapter 4
Sizing the market from the bottom up
Count real accounts instead of quoting analyst reports, and use the count to choose where to start.
10 min read
Market sizing in pitch decks is often a top-down guess built from an analyst report. GTM engineers can do something more useful: count the actual companies that fit, using the same data you will later build lists from. The count tells you which segment to start with and whether your outbound capacity matches the market.
What you'll take away
- 01Why a bottom-up account count beats a top-down market figure for outbound planning.
- 02Total, serviceable and obtainable markets, defined in plain words.
- 03How to count accounts per segment with the data you already have.
- 04How to pick a starting segment, and check it against team capacity.
Why count accounts
"A $40bn market" doesn't tell a sales team anything they can act on. "4,300 companies in our best segment, of which 900 hired a security lead last year" does. A bottom-up count is concrete, it can be checked, and it doubles as the first draft of your target list.
Three numbers, defined plainly
| Number | The question | How to estimate it |
|---|---|---|
| Total addressable (TAM) | How many companies could use this at all? | Count every company matching the broad profile, any region |
| Serviceable (SAM) | How many can we sell to today? | Apply real constraints: regions served, languages, integrations, compliance |
| Obtainable (SOM) | How many could we realistically win this year? | Your team's reach and historical conversion rates applied to the serviceable count |
If you need a money figure, multiply the counts by an average deal size from your own data. Keep the unit of counting as accounts, though. Accounts are what lists, signals and reps work with.
Counting from the bottom up
- 1
Translate the ICP into filters
Industry codes, headcount ranges, countries, technology used. Use the filterable version from chapter 3. - 2
Count in two or three sources
Databases disagree. Counting the same filter in more than one gives you a range and shows where coverage is weak. - 3
Remove what you can't serve
Unsupported regions, existing customers, competitors, the anti-ICP. - 4
Split into segments
By industry, size band or region, whichever splits change how you'd sell. - 5
Add situational counts
For each segment, how many showed a relevant signal in the last year? That is your warmest slice.
Market sizing for outbound is a count of accounts you could name, not a percentage of a big number.
Choosing where to start
Score each segment on a few practical questions rather than on size alone.
| Question | Why it matters |
|---|---|
| How well have we done here before? | Past win rates and deal sizes are the best predictor you have |
| How reachable is it? | Good contact coverage and working channels |
| Are there visible signals? | Segments with detectable 'why now' events are easier to time |
| Is it big enough to learn from? | A few hundred accounts gives enough sends to learn; thirty doesn't |
| Do we have proof? | Case studies or logos from that segment make messages credible |
Does the market fit your capacity?
Check the plan against what the team can actually do. If two reps can each run about 400 new accounts a month through a multi-step sequence, the team covers about 9,600 accounts a year. A serviceable market of 6,500 means you will exhaust it in under a year. Then the work is depth (more contacts per account, better timing, more channels) rather than breadth. A market of 60,000 means the opposite: tighter prioritisation, because you will never reach most of it.
- Small market, big team: go deeper per account and lean on signals for timing.
- Big market, small team: prioritise ruthlessly by fit and signals; don't try to cover everything.
- Recheck counts every quarter. Markets and databases both change.
Try it
Size a segment
You sell payroll software for German companies with 20 to 200 employees. Database A returns 41,000 companies for those filters, database B returns 33,000. You only support companies already using one of two accounting tools, which about 30% of them use. Your two reps can each start 300 new accounts a month.
Estimate the serviceable market, how long the team would take to reach it, and what you would do about it.
Write your answer down first, then reveal the worked answerWorked answer
Serviceable: take a range: 30% of 33,000 to 41,000 is roughly 10,000 to 12,300 companies. The 8,000-company gap between the databases is itself worth investigating, because one source may include dormant entities.
Time to reach: 2 reps × 300 × 12 = 7,200 accounts a year, so about a year and a half to touch everything once.
What to do: prioritise. Start with companies showing signals such as hiring an HR or finance role, growing headcount, or recently switching accounting tools, and with industries where you already have customers. A strong answer treats the database disagreement as a data-quality finding, not just noise.