Key takeaways
- An ICP needs two layers: company fit (therapeutic area, modality, stage, size) and buyer fit (which function and seniority actually owns the decision).
- Your existing best customers are usually the fastest, most reliable starting point for a first-draft ICP.
- Buyer titles in life science are inconsistent across companies, so a buyer-fit model needs to reason about the role, not just match exact title strings.
- An ICP should be revisited against real conversion data as your product, customers, and best-fit segment change.
A lot of ICP advice is written for generic B2B software and translates poorly to a CRO, CDMO, or life science tools and services company. "Company size" and "industry" are not enough: two biotech companies of identical size and headcount can be in completely different buying postures depending on development stage, modality, and how their program mix maps to what you actually sell.
Why generic ICP frameworks fall short in life science
The attributes that predict whether a life science company will buy from a services or tools vendor are domain-specific: therapeutic area and modality, where a company sits in the development pipeline, and what kind of operational complexity they are dealing with right now. A generic firmographic filter (industry, headcount, revenue) misses almost all of that nuance, which is why "good-fit on paper" accounts so often turn out to be a poor match once a rep actually talks to them.
Why so many CROs and CDMOs end up reactive instead of proactive
Without a working ICP, the default motion becomes reactive: wait for inbound interest, work whatever account happens to be in the pipeline, and rely on existing relationships and referrals for new business. That is a reasonable way to fill capacity when demand is strong, but it means missing the emerging and newly funded biotechs that have not discovered you yet, exactly the accounts that are cheapest to win early and hardest to win once a competitor has already locked in the relationship. A clear ICP is what turns "find clients" into a repeatable, proactive search instead of waiting for the phone to ring.
Layer 1: company fit
A working company profile for a life science seller usually needs to define:
- Therapeutic area(s). Which indications overlap with your team's expertise or your existing customer base, and which are explicitly out of scope.
- Modality. Small molecule, biologics, cell and gene therapy, and other modalities require different capabilities from a services partner. This often matters more than therapeutic area for a CDMO or CRO specifically.
- Development stage. Preclinical, Phase 1 through 3, or commercial-stage companies have very different needs and buying cycles. A CDMO selling process development work has a different sweet spot than one selling commercial-scale manufacturing.
- Company size and funding stage. A rough proxy for budget and how formalized the company's vendor evaluation process is likely to be.
- Overlap with your best existing customers. The single most reliable starting point: look at the accounts you already win and keep, and reverse-engineer what they have in common before writing down anything else.
Layer 2: buyer fit
Buyer fit is who inside the account actually owns the decision, which company fit alone does not tell you. Getting the buyer wrong is one of the most common reasons a "good-fit" account list underperforms.
- Function. The department that actually owns the pain you solve, whether that is manufacturing and CMC, clinical operations, regulatory affairs, quality, or business development, depending on what you sell.
- Seniority, adjusted for company size. A Director at a 20-person biotech can carry the budget authority of a VP at a larger organization. Scoring buyer fit purely on title level, without adjusting for company size, systematically undervalues small biotechs.
- Title variability. Life science titles are inconsistent across companies. Chief Scientific Officer, Head of Scientific Affairs, and VP of R&D can all describe the same science leadership function; VP of External Partnerships, Head of Alliance Management, and Director of Business Development can all describe the same partnerships function. A buyer-fit model needs to reason about the role, not just match exact title strings.
A practical way to build the first draft
Start from your best current customers, not a blank page. Pull the last 10-20 accounts you won and are happy to have as customers, and look for what they share on both layers: similar modality or therapeutic focus, a similar development-stage range, and a similar buyer function and seniority pattern among the people who actually drove the deal. Write that pattern down as your first-draft ICP, narrow enough that it functions as a real filter. An ICP that matches half your addressable market is not doing its job.
Keeping the ICP alive
An ICP is a living definition, not a document you write once at company founding and forget. As your product or service offering expands, as you win customers in a new modality or stage, or as your best-fit segment shifts, the ICP should be revisited, ideally against real data on which accounts converted and stuck, not just intuition.