Ask a leadership team to describe their ideal customer and you will usually hear an aspiration: bigger logos, a new industry, the account everyone wants on the website. That is a wish list, not an ideal customer profile. A useful ICP describes who you actually win, and it comes from your own data.
Start with the last 40 decisions
Pull your last 20 closed-won and 20 closed-lost opportunities. For each one, record the industry, company size, the title of the economic buyer, what triggered the first conversation, and how the deal was sourced. You now have a small but honest dataset.
Compare win rates, not volume
Group the deals by each attribute and calculate a win rate for every group. You are looking for combinations where you win more than half the time. Most companies find two or three of these pockets, plus a long tail where they win rarely and slowly. Volume can fool you. A segment that produces lots of opportunities but closes at 12% is costing you selling time.
Tier the result
- Tier 1: the segments where you win most often and fastest. These get your proactive prospecting and your best people.
- Tier 2: segments you win reasonably well. Accept inbound and referrals here, but don't chase.
- Outside the ICP: qualify hard, and be willing to walk away early.
Make it operational
An ICP only matters if it changes behavior. Add an ICP-fit field to every new opportunity, review it in pipeline meetings, and track win rate inside versus outside the ICP every quarter. If the gap is wide, your targeting is working. If it closes, your market is shifting and it is time to re-run the analysis.
We start by understanding where revenue is stuck, then scope the work to fit.