01 /The workshop ICP problem
Most ICPs are written in a half-day off-site, drawn from gut feel and the loudest reference customers in the room. They look great on a slide, then quietly die the moment the team starts prospecting.
The issue is that workshop ICPs describe who you'd like to sell to, not who you actually win with. Those are very different lists.
02 /Your won-deal history is the only ICP that matters
Every closed-won opportunity in your CRM is a data point about who really buys from you — what industry, what size, what trigger event, which buyer persona, what sales cyclePipeline · GlossarySales Cycle LengthThe average number of days between deal creation (or first qualified meeting) and closed-won. A shorter sales cycle increases velocity, reduces cost-per-acquisition and shrinks the window in which deals can stall or be displaced by competitors.View full definition →, what deal shape.
An AI-derived ICP analyses that history at scale: it surfaces the traits your top deals share, ranks them by closed-won concentration, and flags the segments where you punch above your weight.
03 /What changes when reps prospect against a real ICP
Win rates climb because reps stop chasing accounts that look exciting but never close. Cycle times shorten because better-fit buyers move faster. Forecasts tighten because the pipeline is built on deals that resemble previous winners.
Marketing benefits too — campaigns aimed at proven-fit segments outperform broad ones by an order of magnitude.
04 /Re-running the analysis is the secret
ICPs aren't static. New products, new markets and new competitors shift the centre of gravity every couple of quarters. The teams that win re-run the analysis continuously, not annually.
When the ICPICP · GlossaryIdeal Customer ProfileA data-driven description of the customer most likely to buy, succeed and renew. A real ICP is built from your closed-won history — not aspiration — and answers: industry, employee band, geography, tech stack, deal size, and the buying triggers that preceded the win.View full definition → is alive in the system — visible to every rep, every manager, every pipeline review — the whole team starts pulling in the same direction.
D01 /The five questions a real ICP must answer
1. Industry concentration: which 3–5 industries account for 70%+ of your closed-won revenue? 2. Size band: at what employee count or ARR band do your win rates peak — and where do they collapse? 3. Buyer persona: who is the economic buyer in 80% of your wins, and what is their job title in your CRM? 4. Trigger event: what happened at the account in the 60 days before the deal opened? 5. Sales cyclePipeline · GlossarySales Cycle LengthThe average number of days between deal creation (or first qualified meeting) and closed-won. A shorter sales cycle increases velocity, reduces cost-per-acquisition and shrinks the window in which deals can stall or be displaced by competitors.View full definition → shape: how many meetings, how many stakeholders, how long from first meeting to close in your typical winning deal?
If you cannot answer all five with a number drawn from your CRM, you do not have an ICPICP · GlossaryIdeal Customer ProfileA data-driven description of the customer most likely to buy, succeed and renew. A real ICP is built from your closed-won history — not aspiration — and answers: industry, employee band, geography, tech stack, deal size, and the buying triggers that preceded the win.View full definition → — you have an ambition. The whole point of the AI-driven approach is that every answer is a query against your closed-won history, not a debate at an off-site.
D02 /The three patterns that almost always surface
Almost every B2B team that runs a real ICPICP · GlossaryIdeal Customer ProfileA data-driven description of the customer most likely to buy, succeed and renew. A real ICP is built from your closed-won history — not aspiration — and answers: industry, employee band, geography, tech stack, deal size, and the buying triggers that preceded the win.View full definition → analysis finds the same three surprises. First, the industry they thought they sold into is not the industry they actually win in (services teams often discover they are really winning in fintech rather than 'professional services'). Second, the company-size sweet spot is narrower than they thought — usually a 4× band rather than the 20× band they were prospecting against.
Third, and most important, the highest-fit segment usually has the lowest pipeline coveragePipeline · GlossaryPipeline CoverageThe ratio of qualified open pipeline to remaining quota. A common heuristic is 3× — meaning you need three pounds of pipeline for every pound of remaining quota — but the right multiple depends on your win rate and average sales cycle.View full definition → — because reps were never told to focus there. Reallocating prospecting effort toward this hidden best-fit segment is typically worth 15–30% on win ratePipeline · GlossaryWin RateThe percentage of qualified opportunities that result in a closed-won deal over a given period. Calculated as: closed-won deals ÷ (closed-won + closed-lost) × 100. Win rate is the single biggest lever in sales velocity.View full definition → inside two quarters, with no other change required.
D03 /How to operationalise an ICP across the team
An ICPICP · GlossaryIdeal Customer ProfileA data-driven description of the customer most likely to buy, succeed and renew. A real ICP is built from your closed-won history — not aspiration — and answers: industry, employee band, geography, tech stack, deal size, and the buying triggers that preceded the win.View full definition → that lives on a slide gets ignored. An ICP that lives in the CRM gets used. Predara writes the fit signal into a custom field on every account, so every rep sees it before they prospect, every manager sees it during pipeline review, and every marketer sees it before launching a campaign.
The single most powerful operational change is to add a fit-score gate at the qualification stage: deals below the threshold need a written justification before they progress. Within a quarter the average deal in the pipeline looks more like the average deal in closed-won — which is the precondition for everything else (faster cycles, higher win ratePipeline · GlossaryWin RateThe percentage of qualified opportunities that result in a closed-won deal over a given period. Calculated as: closed-won deals ÷ (closed-won + closed-lost) × 100. Win rate is the single biggest lever in sales velocity.View full definition →, tighter forecast) to compound.
D04 /Re-running the analysis: cadence and triggers
ICPs drift. New products, market shifts and competitor moves change the centre of gravity every two to four quarters. The discipline is to re-run the analysis on a fixed cadence (we recommend the start of every fiscal half) and on demand whenever you ship a meaningful new product or enter a new region.
Treat the output the way a finance team treats a forecast: review the changes from last cycle, debate the surprises, and update the prospecting and marketing plan accordingly. The teams that compound advantage over time are the ones that operationalise this loop — not the ones that run the analysis once and then forget about it.
Frequently asked questions
Your CRM already knows who your best customers are. The job is to listen to it.
Going deeper? The 2-week Predara Academy covers this live with peer feedback and instructor Q&A.
See cohorts →Share this article
