// diagnose
Market Sizing
Use when the user wants to size a market - TAM, SAM and SOM - for a pitch deck, a board plan, a new segment or a go-to-market decision. Also use when the user mentions market size, total addressable market, serviceable market, "how big is the opportunity", bottom-up sizing, or has a top-down analyst number they want to use. Builds the size bottom-up from counted accounts and sourced assumptions, derives SOM from the team's real capacity rather than a share of the market, and flags every link in the chain nobody sourced.
npx skills add sarojkjha/aaj-marketing-skills --skill market-sizingClaude Code
/plugin marketplace add sarojkjha/aaj-marketing-skills/plugin install aaj-marketing-skills@aajOpenAI Codex
codex plugin marketplace add sarojkjha/aaj-marketing-skillscodex plugin add aaj-marketing-skills@aajClaude or ChatGPT — no install
Add this as a custom connector:
https://aaj-engines-mcp.sarojkjha10.workers.dev/mcp// when to use it
Use when the user wants to size a market - TAM, SAM and SOM - for a pitch deck, a board plan, a new segment or a go-to-market decision. Also use when the user mentions market size, total addressable market, serviceable market, "how big is the opportunity", bottom-up sizing, or has a top-down analyst number they want to use. Builds the size bottom-up from counted accounts and sourced assumptions, derives SOM from the team's real capacity rather than a share of the market, and flags every link in the chain nobody sourced.
// what you give
The segments you could sell to, a count of accounts in each with its source, the share that fit your ICP and that your channels can reach, what one account pays per year, and your sales or signup capacity
// what you get
TAM, SAM and SOM with the arithmetic shown per segment, a verdict on whether the number is defensible, every unsourced assumption, the one to verify first, and a check of any claimed market share against capacity
Engine spec
// decision it supports
How big is the market we can actually sell to, counted bottom-up, and how much of it can our team win in a year?
// inputs
The segments you could sell to, a count of accounts in each with its source, the share that fit your ICP and that your channels can reach, what one account pays per year, and your sales or signup capacity
// output
TAM, SAM and SOM with the arithmetic shown per segment, a verdict on whether the number is defensible, every unsourced assumption, the one to verify first, and a check of any claimed market share against capacity
// acts or analyzes
Analyzes onlyMakes no network calls, writes no files and changes nothing in your tools.
// where your information goes
| How you run it | Where your inputs go |
|---|---|
| Skill installed in your agent, or the local MCP server | The engine runs on your machine and AAJ receives nothing. Your AI agent still sends what you give it to its model provider (Anthropic, OpenAI and so on), under your own account's terms. |
| Hosted MCP server | Your agent's model provider, as above, plus AAJ's server on Cloudflare, which receives the engine inputs over HTTPS, runs them in memory and keeps no copy. Cloudflare keeps error logs (no inputs) for up to 7 days. |
No engine needs customer names, emails or other personal data; leave them out.
// limitations
- A missing fit share or reachable share is taken as 100%. The engine flags it, but SAM equals TAM for that segment until you enter a real share.
- It checks that each count and share has a source written down, not that the source is right, current, or counts the same market your segment name describes.
- SOM is one year of wins from the capacity you enter, at the average ACV in SAM, capped at SAM. An optimistic rep, pipeline or conversion input gives an optimistic SOM.
- It cannot tell if segments mix currencies or monthly and annual values. ACV must be annual, in one currency, at what customers pay now, for every segment.
// review before acting
// setup
npx skills add sarojkjha/aaj-marketing-skills --skill market-sizinghttps://aaj-engines-mcp.sarojkjha10.workers.dev/mcpAgents listed: Claude Code, Cursor, OpenAI Codex, Windsurf, Cline. Tested and recorded: Claude Code.
Market Sizing
Most market sizes are built backwards. Someone finds an analyst figure with a B on the end, takes one per cent of it, and calls that the plan. The number is large, it is quoted with confidence, and nobody can say which companies it is made of.
A reader who has seen a few of these - an investor, a board member, a sceptical co-founder - discounts it immediately. Not because it is too big, but because it cannot be traced.
When to use
The user needs a market size for a deck, a plan, or a decision about which segment to go after, or is holding a top-down figure and wants to know whether to use it.
Do not use this to produce a bigger number. If the honest bottom-up size is small, that is the finding, and it is usually more useful than the flattering one: it tells you whether the business needs a second segment before it needs more marketing.
Before you start
- Read the brand/product context (
.agents/product-marketing.md) for the ICP and pricing. - Name the segments as lists you could actually pull. "Mid-market SaaS" is a category. "US B2B software companies with 50 to 500 staff" is a list with a count.
- Find a source for every count. A register, a directory, a database export, an association member list. Write down where the number came from and when.
- Check the ICP fit share by hand. Pull a sample of real accounts from the list and count how many actually fit. That share is the most commonly guessed number in any market size.
- Get the capacity. How many deals the team can close in a year, or how many opportunities and at what win rate, or how many signups and at what paid conversion.
Method
Three numbers, each a subset of the one before, each built from things that can be counted.
| Number | What it is | Built from |
|---|---|---|
| TAM | Everyone in the named segments, at what one account pays | accounts x ACV |
| SAM | The share that fit your ICP and that your channels reach today | accounts x fit share x reachable share x ACV |
| SOM | What your capacity can win in a year | wins per year x average ACV in SAM, capped at SAM |
SOM comes from capacity, never from a percentage. "We will take 5 per cent of SAM" is a wish unless the team can close that many deals. The engine computes SOM from capacity and, if someone wants to present a share, reports how many wins the share needs against how many the team delivers.
Every link needs a source. A market size is a chain of multiplications, and it is only as credible as the weakest link. The engine reports each unsourced count or share instead of refusing to run, because an unsourced number is a finding to fix, not an error.
Price at what people pay now. Sizing on an aspirational ACV inflates every number downstream. If the pricing is changing, show both.
Triangulate, do not average. A top-down figure is useful as a cross-check. If it is more than three times larger or smaller than the bottom-up TAM, the two are almost certainly not counting the same market - compare definitions before quoting either. The three-times threshold is a rule of this engine, not an industry benchmark.
Workflow
Step 1 - List the segments. One row per list you could pull, with the count and its source.
Step 2 - Apply fit and reach. Hand-check a sample for fit. Say which channel reaches the rest and how you counted.
Step 3 - Add capacity and run the engine. It returns TAM, SAM and SOM with the arithmetic per segment, and every gap.
Step 4 - Verify the weakest link first. The engine names the unsourced input that sits under the largest share of SAM. Because every factor multiplies, that is the assumption whose error moves the answer most.
Step 5 - Decide what the number is for. If SOM is small against SAM, the constraint is capacity, and the plan is hiring or a channel that scales. If SAM is small against TAM, the constraint is fit or reach, and the plan is a second segment or a new channel.
Run the tool
node resources/market-size.js --demo
size(config) returns TAM, SAM and SOM, the per-segment chain, the verdict (defensible, needs
sources, or incomplete), every finding, and the assumption to verify first.
Use the engine rather than a spreadsheet you build on the spot. The same checks every time is what stops a flattering assumption going unnoticed.
Present the result
- The verdict first - defensible, needs sources, or incomplete - and why.
- TAM, SAM and SOM with account counts, not only currency. A reader trusts "1,500 accounts" more than "$15M", because they can picture it.
- The chain for each segment, so every multiplication is visible.
- What to verify first, and every unsourced assumption.
- Any claimed share against capacity, with the number of wins it would need.
- The top-down cross-check, if there is one, with its source and scope.
Guardrails & common mistakes
- Never present a top-down figure as the market size. Use it only as a cross-check, and only with its source and what it counts.
- Never derive SOM as a percentage of SAM. Derive it from capacity, and state the capacity.
- Do not leave the fit share at 100%. If it is not given, the engine treats every account as a fit and says so - that is almost never true.
- Do not round up. 1,501 accounts is more believable than "about 2,000".
- Do not mix currencies or periods. ACV is per year, in one currency, for every segment.
- A small honest number beats a large untraceable one. It is also the one that survives due diligence.
Related AAJ resources
- Scoring the accounts inside a segment: https://aajconsult.com/tools/icp-fit-scorer
- What one account is worth: https://aajconsult.com/tools/pricing-arpu-modeler
- Turning the size into a plan: https://aajconsult.com/playbooks/90-Day-Growth-Playbook
Related skills
target-account-list— the named accounts behind a segmentpricing-and-packaging— the ACV the size depends onunit-economics— whether winning that SOM pays back
// faq
// install
npx skills add sarojkjha/aaj-marketing-skills --skill market-sizing// worked example
Get the worked example for Market Sizing
One email. The filled-in config, the output, and how to read it.
// human in the loop
Need a human in the loop?
AAJ builds the strategies, tools, and playbooks these skills are made of — with engagements for growth-stage and enterprise teams.