Stop Building a 20,000-Company TAM

Most early-stage SaaS teams do the same thing after landing their first handful of customers. They look at who bought, find the commonalities, and build a TAM. "We sell to mid-market tech companies in Europe with 50-200 employees." That's 20,000 companies. A nice-looking spreadsheet. And a completely useless targeting strategy.

The problem isn't the research. The problem is the abstraction. When you average out your first customers into a single ICP description, you lose the thing that actually made them buy: the specific situation they were in, the specific problem that became urgent, and the specific reason your product was the right fix.

I call the alternative Atomic Segmentation. Instead of one broad market, you build a portfolio of tiny, hyper-specific segments, each defined by three things: who has the problem, why your product specifically fits, and what just happened that made it urgent. Each of those is an atom. And atoms are where pipeline actually comes from.

What an Atom Actually Is

An atom is the smallest targetable segment where the "who," the "why us," and the "why now" are all baked into the definition. You can't split it further without losing the logic of why these companies would buy.

Compare this to a traditional ICP:

Traditional segment: "Logistics companies, 100-300 employees, in Europe."

That's a filter. It tells you who to put in a spreadsheet. It tells you nothing about what to say when you reach them, or why they'd care right now.

Atom: "Logistics companies with large fleets of field workers still submitting expenses manually, that just closed a Series A, where the Finance Lead and their first hire are now under board pressure to professionalize financial operations."

That's a story. You know who they are. You know the people inside the company who feel the pain (a two-person finance team drowning in manual receipts from 200 field workers). You know why your expense management tool beats their current process (built for volume and simplicity, not the complexity that SAP Concur solves for). You know what just happened that made this urgent (new board, new expectations, need to clean up ops fast). And you know that Finance Lead is probably already Googling "SAP Concur alternatives for small teams."

The difference matters because every part of the atom does work:

• Who narrows the universe. Not all logistics companies. The ones with field-heavy teams.
• Why us confirms fit. Your product specifically wins against their current alternative. If they were using SAP Concur and had a dedicated finance team, you're not the right fit. The atom excludes them.
• Why now identifies the trigger. Series A funding + board pressure = urgency. Without the trigger, they might have the problem but not the motivation to solve it.

How to Build Your First Atoms

You don't start from market research. You start from your existing customers.

Take your first 5-10 paying customers. Don't look at what they have in common on paper. Look at what happened right before they bought. What was the situation? What was the trigger? Why did they pick you over the alternatives?

You'll probably find 2-4 distinct patterns. Each pattern is a candidate atom.

One might be: companies in a specific vertical that share a technical culture and want to solve the problem the way you solved it. Another might be: companies that just hired a specific role, which signals they're building out a function your product supports. A third might be: companies currently using a tool that's too expensive or too complex for their size, which makes your simpler product a natural fit.

The signals that define each atom are what make them targetable at scale. These aren't abstract personas. They're observable, researchable attributes:

• Tech stack signals. They're using Software X, but they only have a small team in that department. Probably overpaying. Your product is a better fit at their scale.
• Hiring signals. They just posted a job for Head of [Function]. That means they're building out or scaling exactly the area where your product fits.
• Funding signals. They just raised a Seed or Series A. New money, new pressure, new willingness to invest in infrastructure.
• Firmographic clusters. Not "all companies in Industry X," but "companies in a specific sub-vertical that share a buying culture and problem profile."
• Implicit fit signals. They're ISO certified (indicates process maturity and compliance needs). They have a partner program that looks custom-built (suggests they've outgrown manual workflows). They have a department above a certain headcount threshold. These are the things a good salesperson would notice during manual research. The trick is making them researchable at scale.

You research these through Clay, Apollo, or whatever enrichment tools you use. The key is that the atom definition tells you exactly what signals to look for. You're not running a broad filter and hoping for the best. You're confirming a specific hypothesis about why this cluster of companies would buy.

Start with 3-5 atoms. That's enough. Each one might contain 200-500 companies. That's not a 20,000-company TAM. That's a portfolio of 600-2,500 companies where you have a genuine reason to believe they have the problem, right now, and that your product is the right solution.

The Inbound Radar

Atoms are your proactive bet. You've studied your customers, defined hypotheses about who else looks like them, and built targeting around those hypotheses. But here's the thing about hypotheses: they're limited by what you already know.

What about the companies you haven't thought of yet? The verticals you haven't considered? The use cases you didn't anticipate?

That's where the inbound radar comes in. While your atoms handle outbound targeting, you run a parallel motion designed to catch demand you didn't predict.

The mechanics are straightforward:

• Search marketing (SEO + GEO): Make sure you show up when someone is actively searching for the problem you solve. Not your product name. The problem. This catches individuals already experiencing the pain.
• Thought leadership (organic + paid social): Run content on LinkedIn, both organic posts and paid Thought Leadership Ads, that speaks to the pain point broadly. Not pitched at a specific atom. Pitched at anyone who recognizes the problem.

Both of these motions drive individuals toward your website or content. People who read your blog post. People who engage with your LinkedIn post. People who visit your pricing page from a Google search.

These are signals. And they're valuable because they come from people you didn't target. They self-selected.

From here, you do three things:

1. Identify the company. Website deanonymization tools or just looking at who's engaging on LinkedIn. You're turning an individual signal into a company-level insight.
2. Run an automated ICP check. Does this company have at least X people in the relevant department? Are they ISO certified? Do they have a custom-built solution on their website that suggests they've outgrown manual processes? Are they in a size range where your product fits? This is the manual prospecting a salesperson would normally do before reaching out, but automated through Clay or similar tools.
3. Promote or discard. If they pass the ICP check, they enter your active targeting. If not, they stay in the broad awareness layer and might come back later.

The inbound radar doesn't replace your atoms. It complements them. Atoms tell you where to proactively invest your outbound effort. The inbound radar tells you where demand already exists that you hadn't accounted for.

How the Two Sides Work Together

Early-stage SaaS teams have limited resources. You can't run 20 campaigns and chase every lead. The power of this system is that it self-prioritizes.

Here's how:

Atom-only signal: A company matches one of your atoms but hasn't engaged with any of your content. They're a good proactive target. You include them in your outbound and paid campaigns, but you don't invest heavy sales time yet.

Inbound-only signal: Someone from a company you've never heard of reads three blog posts and visits your pricing page. They pass the ICP check. They're interesting, but you don't have a hypothesis about why they'd buy. Worth reaching out, but with curiosity, not a pitch.

Both signals: A company that matches an atom AND someone from that company is engaging with your content. Maybe that Finance Lead from the logistics company is reading your "SAP Concur alternatives" comparison page. This is where you go deep. The atom tells you why they should care. The engagement tells you they already do. This is your highest-priority action.

The system creates a natural hierarchy. Instead of treating all 20,000 companies in your TAM equally (or worse, randomly), you have a clear framework for where to spend time, budget, and attention.

Atoms Are a Portfolio

This isn't a one-time exercise. The biggest mistake would be defining your atoms once and treating them as permanent.

Your product evolves. Your first atoms might be in one vertical, but as you build new features and solve more complex use cases, new verticals open up. What starts as a single-industry play can grow into three or four distinct atoms across different markets.

A healthy GTM strategy looks like this:

• 2-3 proven atoms where you have strong signal definitions, validated messaging, and consistent pipeline. These get the majority of your outbound budget.
• 1-2 emerging atoms that you're testing. Maybe you noticed a pattern in recent inbound leads. Maybe a new feature unlocked a new use case. You're sculpting these, running smaller campaigns, seeing if they convert.
• The inbound radar running continuously, detecting signals you haven't formalized into atoms yet.

Over time, emerging atoms either graduate to proven atoms or get retired. New candidates come in from inbound signals or product evolution. The portfolio grows and diversifies.

This is how you build a grounded, diversified GTM strategy. Not by expanding a spreadsheet from 20,000 to 40,000 companies. By adding new atoms, each with its own thesis about who, why you, and why now.

What I'd Actually Recommend

If you're an early-stage SaaS team currently staring at a big TAM spreadsheet and wondering why your outbound isn't converting:

Stop averaging your customers into one ICP. Go back to your first 5-10 deals and ask: what was the situation, what was the trigger, and why did they pick us? Define 3-5 atoms from those answers.

For each atom, identify the signals that make them findable. Tech stack, hiring patterns, funding events, firmographic indicators. Build your outbound targeting around those signals, not around broad firmographic filters.

In parallel, make sure your inbound channels are open. SEO for the problem you solve. Thought leadership on LinkedIn that resonates with anyone experiencing the pain, not just your atoms. Let people self-select into your world, then check if they're a fit.

When a company shows up in both your atom targeting and your inbound signals, that's your best lead. Prioritize accordingly.

And keep evolving. Add new atoms as your product grows. Test emerging segments. Retire the ones that don't convert. Treat your GTM like a portfolio, not a list.

The companies that struggle with pipeline at early stage aren't usually missing data or tools. They're missing specificity. Atomic Segmentation is how you get it back.