market-sizing-analysis

Overview of market-sizing-analysis skill

What the market-sizing-analysis skill does

The market-sizing-analysis skill helps an AI agent produce structured TAM, SAM, and SOM estimates for startup and product opportunities. It is built for market-opportunity work where you need more than a loose “how big is this market?” answer: you need a defendable sizing approach, explicit assumptions, and a methodology that can be reviewed by founders, operators, or investors.

Who should use market-sizing-analysis

This market-sizing-analysis skill is best for:

• founders validating a new market
• startup operators preparing fundraising material
• consultants doing quick but structured opportunity analysis
• analysts who want a repeatable market-sizing workflow for Data Analysis

If you need a first-pass market model with clear logic, it is a strong fit. If you need audited research or highly regulated industry forecasts, it should be a starting framework, not the final source of truth.

The real job-to-be-done

Most users do not just want TAM/SAM/SOM definitions. They want to turn a rough idea like “AI software for mid-market retailers” into:

• a scoped target market
• segment-level assumptions
• one or more sizing methods
• realistic 3-5 year obtainable share logic
• a narrative suitable for planning or pitching

That is where market-sizing-analysis is more useful than a generic prompt.

Key differentiators vs ordinary prompting

The main value of market-sizing-analysis is that it pushes the agent toward three complementary approaches:

• top-down sizing from industry reports
• bottom-up sizing from customer counts and pricing
• value-theory sizing from willingness to pay

That matters because market sizing fails when a model relies on only one lens. This skill gives users a more decision-ready structure and encourages cross-checking instead of presenting a single impressive but fragile number.

What matters most before you install

The biggest adoption question is not “can it calculate TAM?” but “will it reduce guesswork?” For this skill, the answer is usually yes if you can provide:

• a defined product or service
• target customer characteristics
• geography
• rough pricing or contract value
• timeline and go-to-market constraints

Without those inputs, the output becomes generic quickly.

How to Use market-sizing-analysis skill

market-sizing-analysis install context

The repository excerpt does not surface a built-in install command inside SKILL.md, so users typically add the parent skills repository and then invoke the skill by name in their agent environment. If your setup supports Skills-style installs, the common pattern is:

npx skills add https://github.com/wshobson/agents --skill market-sizing-analysis

After install, verify that your agent can see the skill under the startup-business-analyst plugin path.

Read these files first

For practical market-sizing-analysis usage, start with:

• plugins/startup-business-analyst/skills/market-sizing-analysis/SKILL.md
• plugins/startup-business-analyst/skills/market-sizing-analysis/examples/saas-market-sizing.md
• plugins/startup-business-analyst/skills/market-sizing-analysis/references/data-sources.md

This reading order works well:

1. SKILL.md for the workflow and method choices
2. examples/saas-market-sizing.md for the shape of good output
3. references/data-sources.md for where assumptions should come from

What input the skill needs to work well

For strong market-sizing-analysis usage, give the agent a compact operating brief:

• product description
• buyer type
• industry or use case
• geography
• pricing model
• time horizon
• known competitors
• constraints on what the product can actually serve today

A weak input is: “Size the market for AI legal software.”

A stronger input is: “Size the 3-5 year market for AI contract review software for U.S. mid-market legal teams at companies with 200-5000 employees. Assume annual pricing of $18k-$60k depending on seat count and a direct sales motion.”

How to turn a rough idea into a complete prompt

A good invocation prompt for market-sizing-analysis for Data Analysis should ask for method, assumptions, and output shape in one request. For example:

“Use the market-sizing-analysis skill to estimate TAM, SAM, and SOM for an AI-powered email marketing platform for North American e-commerce companies with $1M+ revenue. Use bottom-up as the primary method, top-down as a cross-check, and state all assumptions. Include segment counts, ACV ranges, 3-5 year obtainable share logic, and a short risk section on uncertainty in the source data.”

This works better than “estimate the market size” because it reduces ambiguity on:

• target segment
• preferred methodology
• output format
• confidence and caveats

Choose the right methodology first

Do not default to top-down because it feels faster. This skill is most credible when you match the method to the market:

• Use bottom-up when you know customer segments, pricing, or seat counts.
• Use top-down when the market already has published category estimates.
• Use value theory when pricing depends on economic value created rather than standard category pricing.

For startup work, bottom-up is often the best primary method because it is easier to defend in a board deck or pitch.

Suggested workflow in practice

A good workflow with market-sizing-analysis looks like this:

1. Define the exact offering and buyer.
2. Narrow geography and segment constraints.
3. Pick a primary sizing method.
4. Ask the agent for assumptions before final numbers if uncertainty is high.
5. Run a cross-check with a second method.
6. Adjust SAM and SOM based on product scope, GTM capacity, and competition.
7. Export the result into a memo, pitch slide, or planning doc.

This sequence prevents the common failure mode where TAM is large but unrelated to the business you can actually build.

Use the example file as a quality benchmark

examples/saas-market-sizing.md is especially useful because it shows what “complete enough” looks like:

• clear segment criteria
• count-based bottom-up logic
• explicit ACV assumptions
• formulas
• realistic obtainability framing

If your output does not include those ingredients, ask the agent to revise rather than accepting a narrative-only answer.

Data sources that materially improve results

The references/data-sources.md file is one of the strongest parts of this skill. It points users toward:

• premium analyst firms like Gartner, Forrester, and IDC
• accessible sources like Statista
• startup and private-market tools like CB Insights and PitchBook
• broader strategic sources like McKinsey insights

For practical use, combine one published market source with one count-based validation source. Example:

• published category estimate from Statista
• buyer count check from Census, platform ecosystem counts, or LinkedIn filters

That is usually more reliable than citing a single industry report.

What good output should include

High-quality market-sizing-analysis guide output should include:

• TAM, SAM, and SOM definitions applied to your case
• formulas or calculation logic
• segment assumptions
• time horizon
• pricing assumptions
• key uncertainties
• rationale for obtainable share

If the output gives neat market numbers without showing how they were built, ask for a recalculation with assumptions exposed.

Common constraints and tradeoffs

This skill is useful, but it does not remove core market-sizing limitations:

• public source data may use category definitions that do not match your product
• customer counts may be outdated or inconsistent across sources
• value-based sizing can become speculative fast
• SOM estimates are often more about GTM realism than market math

Use it to structure judgment, not to fabricate precision.

market-sizing-analysis skill FAQ

Is market-sizing-analysis good for beginners?

Yes, especially if you understand your product and customer better than you understand formal market-sizing methods. The skill gives a framework that is easier to follow than starting from a blank prompt. Beginners still need to review assumptions carefully because bad scope definition leads to bad numbers.

When is market-sizing-analysis not a good fit?

market-sizing-analysis is a poor fit when:

• you need audited market research
• the market is too undefined to describe a buyer
• pricing is unknown and impossible to estimate
• the real problem is demand validation, not market sizing

It is also a weak fit for highly technical categories where public segment data is extremely thin and domain experts are required.

How is this different from a normal AI prompt?

A normal prompt may produce plausible TAM/SAM/SOM language but skip the hard parts: segmentation, methodology choice, and defendable assumptions. The market-sizing-analysis skill is better when you want a repeatable workflow rather than a one-off answer.

Can I use market-sizing-analysis for investor decks?

Yes, but do not drop the first output into a pitch deck unchanged. Use the skill to create a traceable model, then tighten the sources, simplify the narrative, and make sure the SAM and SOM reflect your actual launch scope and GTM capacity.

Does it work only for SaaS?

No. The included example is SaaS-oriented, but the framework can also work for services, marketplaces, fintech, healthtech, and other startup categories. It works best where you can estimate customer counts, spending levels, or economic value created.

How to Improve market-sizing-analysis skill

Give narrower market definitions

The fastest way to improve market-sizing-analysis output is to narrow the market definition. Specify:

• exact buyer
• company size or user profile
• geography
• deployment model
• current product scope

“Healthcare AI” is too broad. “AI prior-authorization automation for U.S. regional health insurers” is much more usable.

Provide pricing and packaging assumptions

Bottom-up sizing gets much stronger when you supply one of:

• annual contract value
• monthly subscription range
• seat-based pricing
• transaction take rate
• average deal size

Without pricing, the model often has to invent weak proxies.

Ask for cross-checks, not just one number

A strong prompt asks the agent to produce:

• primary method
• secondary validation method
• explanation for any gap between them

That improves trust. Large differences between top-down and bottom-up estimates are often the most useful insight, because they reveal category-definition problems or unrealistic pricing assumptions.

Force the agent to separate TAM, SAM, and SOM logic

A common failure mode is that the model simply applies percentage cuts without explaining why. Improve results by asking for distinct logic:

• TAM based on total potential spend
• SAM based on current product and geography constraints
• SOM based on realistic acquisition capacity and competition

This makes the market-sizing-analysis guide more operationally useful.

Ask for source quality and uncertainty notes

Tell the agent to label assumptions as:

• sourced
• inferred
• placeholder

Also ask for a confidence note on each major input. This is especially helpful if you are using the skill in early-stage strategy work where some numbers will inevitably be directional.

Iterate after the first draft

Do not treat the first run as final. A good revision loop is:

1. correct buyer and geography mistakes
2. replace weak assumptions with real inputs
3. tighten pricing
4. challenge SOM realism
5. re-run with one more source cross-check

That usually improves output quality more than adding extra prose.

Use the example structure for your own domain

If your first result is messy, tell the agent to mirror the structure in examples/saas-market-sizing.md:

• segment table
• formula section
• calculation walkthrough
• takeaway summary

That file is a useful formatting model even when your market is not SaaS.

Watch for these common failure modes

The main quality problems in market-sizing-analysis are:

• category inflation in TAM
• vague segment counts
• ungrounded pricing assumptions
• SOM based on hope rather than GTM capacity
• mixing user counts, company counts, and revenue without clear conversion logic

If you see any of these, ask for the exact line of reasoning to be rebuilt.

Improve outputs for Data Analysis workflows

For market-sizing-analysis for Data Analysis, ask the agent to return assumptions in structured form:

• segment
• count
• pricing metric
• annual revenue assumption
• source
• confidence

That makes it much easier to move the result into spreadsheets, notebooks, BI tools, or downstream forecasting models.