Building an ICP List from Secondary Market Research, Not Guesswork
Most ICP definitions are written from memory: a founder's gut feel about who buys, dressed up as a slide with firmographic ranges. Secondary research replaces the guessing with public data you can pull today — registries, filings, industry reports, association membership lists — so your ICP reflects the market as it actually is, not as your last five closed deals happened to look.
- Secondary research means data someone else already collected — registries, filings, industry reports, directories — not a new survey you run yourself.
- A usable ICP needs three layers: firmographic (size, industry, geography), technographic (what they run), and trigger-based (what recently changed).
- Cross-reference at least two independent sources before treating a segment as validated; single-source ICPs tend to overfit to whichever list was easiest to find.
- Sizing the list early prevents wasted campaign setup — an ICP with fewer than a few hundred matching companies rarely supports sustained outbound.
- Secondary research narrows the target; it does not replace the contact-level enrichment and verification you still need before the first send.
Why secondary research beats an ICP written from memory
Ask most B2B teams to describe their ideal customer and you get an answer shaped by the last few deals that closed easily, not by the market. That is survivorship bias, and it quietly narrows your addressable list to whoever resembles your existing logos — which is fine until growth requires reaching companies you have never sold to.
Secondary research is any data someone else already gathered: government business registries, industry association reports, market sizing studies, regulatory filings, trade publication rankings. None of it requires you to run a survey or buy access to a proprietary panel. It exists specifically so you do not have to guess at market structure before you start prospecting.
The payoff for address-based B2B outreach is direct. A campaign built against a properly sized, evidence-based ICP gets a cleaner reply rate because the offer actually matches the company's situation. A campaign built against a hunch-based ICP burns list volume on companies that were never going to buy, and the bad signal pollutes your sender reputation along the way.
Where usable secondary data actually lives
Start with sources that are free, public, and reasonably current. Company registries — the equivalent of a companies house or state business registry — give you legal entity data: registration date, industry code, sometimes headcount and revenue bands. This is the most reliable layer because it is regulatory data, not marketing data.
Industry and trade association reports are the next layer. Most verticals have at least one association that publishes annual market sizing, member directories, or salary and headcount benchmarks. These reports tell you how big a segment actually is and how it is trending, which matters when you are deciding whether an ICP is worth building a campaign around at all.
Regulatory filings — procurement disclosures, licensing databases, import-export records depending on the industry — surface companies that are otherwise hard to find through generic search. A construction firm holding an active government contract, or a logistics company with recent customs filings, is both identifiable and, often, in an active buying cycle.
Finally, trade publications and analyst rankings (top 100 lists, sector reports, conference sponsor and speaker rosters) are a fast way to sanity-check a segment. If a niche has an active trade press and recurring events, it has enough commercial activity to be worth targeting.
- Business registries: legal status, industry code, incorporation date, headcount band
- Association reports: market size, growth trend, member directories
- Regulatory and procurement filings: licenses, contracts, customs, permits
- Trade press and analyst rankings: top-N lists, sector deep dives
- Conference sponsor and speaker lists: active budget, category leadership
Turning research into ICP criteria
Raw data becomes an ICP only once you translate it into three layers of criteria. The firmographic layer is the baseline: industry code, employee range, revenue band, geography, legal structure. This is what registries give you directly, and it should be specific enough to exclude clearly, not just include vaguely — 'B2B software companies, 20 to 200 employees, US and UK' filters far better than 'tech companies.'
The technographic layer answers what the company runs — which platforms, which category of tool they already use or conspicuously lack. Association reports and trade press sometimes name specific tools in case studies; job postings (a form of secondary data most teams overlook) reveal stack requirements directly. A company hiring for a role that implies a gap in their tooling is a stronger fit signal than headcount alone.
The trigger layer captures timing: recent funding, a new executive hire, a office expansion, a regulatory change affecting their industry, a public complaint about a problem you solve. Triggers turn a static ICP into an active target list, because they tell you which companies in your fit segment are likely to be receptive right now rather than at some undefined future point.
A vendor selling compliance software might define ICP as: mid-market financial services firms (50 to 500 employees, registry data), currently running spreadsheet-based reporting (inferred from job postings requesting manual reconciliation skills), that recently expanded into a new regulatory jurisdiction (filings data) — three layers, three sources, one defensible segment.
Sizing the list before you build the campaign
Once criteria are set, count how many real companies match before writing a single email. This step gets skipped constantly, and it is the single biggest cause of outbound campaigns that fizzle after two weeks. If your ICP criteria return a few dozen companies, you have a target account list for a hand-crafted, high-touch motion — not a sustained cold email program, which needs enough volume to run cadences, test messaging, and absorb the normal drop-off from bounces and disqualifications.
As a rough planning number, a sustainable outbound program benefits from a segment in the low thousands of matching companies at minimum, so you can run multiple campaigns, refresh lists as they age, and still have headroom after removing suppressed and unreachable contacts. If secondary research turns up fewer, either widen one criterion (geography or headcount range) or accept that this segment is a targeted-account motion, not a volume one.
Cross-check the size estimate against a second source before committing. A registry count and an association's published market size rarely match exactly, but they should be in the same order of magnitude — if a trade report says a niche has 400 companies globally and your registry pull returns 15,000, one of the two definitions is wrong and needs tightening or loosening.
Common mistakes that quietly wreck the segment
The most common failure is stopping at a single source. A list pulled only from one registry or one directory inherits that source's blind spots — geographic coverage gaps, industry code misclassification, stale entries for companies that closed or pivoted years ago. Two independent sources that roughly agree is the minimum bar for calling a segment validated.
The second failure is treating firmographic fit as sufficient. Two companies with identical headcount and industry code can be in completely different buying situations — one just signed a three-year contract with a competitor, the other has an open budget line for exactly your category. Skipping the trigger layer means every company in the list gets the same generic pitch regardless of timing.
The third failure is letting the research go stale. Registries update on lag, association reports come out annually, job postings disappear within weeks. Treat secondary research as a snapshot with a shelf life — refresh the trigger layer at least quarterly and the firmographic layer at least twice a year, or you end up prospecting companies that no longer exist in the form your data describes.
From ICP to a working company list
Once the three-layer ICP is validated and sized, the next step is converting it into named companies with the right contacts, not just a criteria document. This is where the operational tooling matters: a company database that lets you filter by industry, headcount, and custom fields for the technographic and trigger signals you pulled from research, then build a working list without re-doing the research manually for every campaign.
In LDM, this maps onto the company list and custom fields mechanics — ICP criteria become saved filters, trigger data goes into custom fields with a source note so the team can see where each signal came from, and the resulting list feeds directly into campaign targeting without a manual export step. Whatever platform you use, the discipline is the same: keep the source of every criterion documented, because the moment someone asks why a company is on the list, 'gut feel' is not an answer that survives a pipeline review.
- Define firmographic, technographic, and trigger criteria separately, each tied to a named source
- Validate segment size against at least two independent sources
- Set a minimum viable segment size before greenlighting a sustained campaign, not just a target-account list
- Store trigger data with a timestamp and refresh it on a set cadence
- Convert validated criteria into saved filters or custom fields so the list rebuilds without re-research
FAQ
What counts as secondary research versus primary research?
Secondary research is data someone else already collected and published — registries, industry reports, filings, directories. Primary research is data you collect yourself, such as customer interviews or a survey you run. ICP building usually starts with secondary research because it is faster and covers the whole market, then narrows further with primary input from sales conversations.
How many sources do I need before trusting an ICP segment?
Two independent sources that roughly agree on size and characteristics is a reasonable minimum. A single source inherits that source's coverage gaps and classification errors, which can silently distort the whole segment.
How often should I refresh secondary research for an active ICP?
Firmographic data (industry, size, geography) is stable enough to refresh twice a year. Trigger data (funding events, hiring signals, regulatory changes) decays fast and is worth refreshing quarterly, or continuously if you have the tooling to monitor it.
What if secondary research shows my ICP segment is too small for cold email?
Treat it as a target-account motion instead — a short, hand-researched list with highly personalized outreach and multi-channel follow-up, rather than a sustained cold email cadence that needs volume to function. Alternatively, widen one criterion, such as geography or headcount range, and re-check the size.
Does secondary research replace the need to verify individual contacts?
No. Secondary research validates the company-level segment; you still need contact-level enrichment and email verification before sending, since registries and reports rarely include current, deliverable contact details for the actual decision-maker.
Want to apply this to your outreach?
We will map it to your segment and product — before any work starts.
Talk to us