Practitioner Framework
An ICP that's "B2B SaaS, 50-500 employees, US/UK" isn't an ICP. It's a guess. Here's the 8-layer framework ORRJO uses on every client to build an Ideal Customer Profile that actually moves pipeline. Eight steps, two hours, one document. Free template at the end.
An Ideal Customer Profile is the set of attributes that define your best-fit customer at the company level. It's not a buyer persona. That's the individual. The ICP is the company.
A useful ICP filters which accounts to target. A weak ICP makes outbound generic, lengthens sales cycles, and inflates churn. Most B2B founders write their ICP in 20 minutes and don't revisit it. The result is target lists that include companies who'll never buy and miss the ones who would.
Work through each layer in order. Each takes 10-20 minutes. Skip none.
List the ten customers with the highest LTV, lowest churn, and shortest sales cycle. Not your favourites. Not the ones who pay the most. The ones with the cleanest commercial profile. These ten are the truth.
Industry, sub-industry, employee count, revenue range, geography, growth stage, funding status, ownership structure. Find the patterns across the top 10. Not "B2B SaaS, 50-500 FTE". Specific. "Series B SaaS in financial services compliance, 80-300 FTE, UK or Ireland HQ".
What stack do they run? What tools imply they're a fit? Salesforce or HubSpot? Amazon or Azure? Outreach or Apollo? Use a technographic data source like BuiltWith or Wappalyzer to score accounts. Tech-stack patterns often predict fit better than firmographic patterns.
Identify economic buyer, champion, blocker, and influencer roles. List the job titles and seniority that show up across closed-won deals. If your champion is always Head of Revenue Ops but your economic buyer is always CFO, your sequencing matters.
What was happening at the company when they bought? Funding round, new hire, leadership change, M&A, expansion, technology change, regulatory deadline. Triggers turn ICP fit into pipeline. Without them, ICP is static and you target everyone equally.
Articulate the specific commercial problem they have that you solve. In their words, not yours. If you can't articulate it in one sentence that a non-customer would recognise, the ICP isn't tight enough yet. Go back to step one.
List five lost deals or churned customers. Identify what was different. Maybe their buying committee was bigger. Maybe their tech stack was wrong. Maybe they were too early-stage. Subtract those attributes from the ICP. Bad-fit data sharpens the ICP faster than good-fit data.
Use firmographic + technographic + trigger filters to build a 200-1,000 account list. Under 200 means the ICP is too narrow or the market is too small. Over 1,000 means it's too loose. This list is the input for outbound and demand gen.
Before scaling outbound, run a 50-account pilot. Track reply rate, meeting rate, conversion to pipeline. If reply rate is under 2% or attendance under 70%, the ICP is broken. Iterate. Don't scale a broken ICP.
The deck says "we serve mid-market B2B SaaS". Reality says you've closed 8 of your last 10 deals in compliance-heavy fintech with 80-300 FTE. The deck is aspirational. The actual ICP lives in the deal data.
If you sell more than one product, you have more than one ICP. The product that closes in 30 days has a different ICP than the one that closes in 9 months. Treat them separately.
Most teams only look at customers who closed. The cleanest signals come from comparing closed-won and closed-lost. Bad-fit data is what tightens the ICP fastest.
Persona is a person. ICP is a company. Both matter. They're different documents. If your "ICP" is "VP of Marketing at a B2B SaaS company", you've described half a persona, not an ICP.
Markets shift quarterly. Your ICP should review quarterly. Most teams write one and never touch it again. Annual rebuild minimum.
An ICP without trigger events is just a static target list. Triggers are what tell you who's buying right now. Without them, every account in the list looks equal. They're not.
Once the ICP is locked, the rest of the engine plugs in. Demand gen warms the target list. Outbound runs through it. Reporting scores the activity by ICP fit, not just volume.
LinkedIn campaigns, paid distribution, and content syndication run against the ICP target list. Every impression is qualified by ICP fit before it serves.
SDRs work the ICP-tier-A accounts first. Tier B is secondary. Tier C is parked. Most agencies skip the tiering and run the same cadence on a 5,000-account list. ORRJO doesn't.
Every meeting booked is scored against ICP fit. Out-of-ICP meetings flag for review. If too many slip through, the ICP filter is loose and we tighten it.
An Ideal Customer Profile is the set of attributes that define your best-fit customer at the company level. Not a buyer persona (that's the individual). The ICP describes the company. It includes firmographics, technographics, buying committee, and the specific problem you solve.
Tight enough that 200-1,000 accounts match it. Tighter than 200 means the addressable market is too small to support a sales motion. Looser than 1,000 means outbound becomes too generic.
ICP describes the company. Persona describes the people inside it. You need both. ICP filters which accounts to target. Persona shapes how you message individuals once you've targeted the account.
Quarterly review, annual rebuild. Trigger events for a rebuild: product pivot, new market entry, three or more lost deals from the same gap, or a meaningful change in win rate by segment.
Run a 50-account pilot. Track reply rate, meeting rate, pipeline conversion. Reply rate under 2% or attendance under 70% means the ICP is wrong. Don't scale a broken ICP.
ORRJO runs a two-week intelligence sprint that locks the ICP and produces the target account list. Book a 30-minute call.
Book a Strategy Call →