Day 3: Build an Ideal Customer Profile That Filters Out the Wrong Deals
The Concept
Almost every sales organisation has an Ideal Customer Profile document. Almost none of them actually use it.
The reasons vary. Sometimes the ICP was built in a strategy workshop and reflects the customers the leadership team aspires to serve, not the ones who actually buy and stay. Sometimes it is so broad — "mid-market companies in North America" — that it filters out almost nothing. Sometimes it sits in a shared drive that was last updated eighteen months ago and does not reflect how the market or the product have changed. Whatever the reason, the result is the same: reps are making qualification decisions based on intuition and gut feel rather than a clear, applied framework, and the pipeline is full of deals that look possible but are actually low-probability time sinks.
A useful ICP is not a document. It is a decision-making tool you consult every time a new prospect enters your world.
Why ICPs fail and what makes them work
The most common failure mode is that ICPs are built from aspiration rather than evidence. A leadership team decides they want to move upmarket, so the ICP gets updated to reflect enterprise accounts — even though 80% of the last year's revenue came from mid-market companies who closed in three weeks. The field team then receives an ICP that describes customers they almost never see, and quietly continues qualifying on feel.
The second failure mode is that even an accurate ICP never gets operationalised. It describes who the ideal customer is in general terms but does not tell a rep how to recognise one in the first five minutes of a discovery call, or how to score a prospect from a LinkedIn profile. There is no checklist, no scoring rubric, no set of questions to ask.
What makes an ICP actually work is two things: it is built from your real data, and it is specific enough to use in a conversation. That means green flags you can actually look for in a prospect's public footprint, and red flags you can recognise in the way they describe their situation — not abstract categories like "budget-conscious buyer" but concrete signals like "they are replacing a tool they built in-house" or "they mentioned the project has been on hold for six months."
How AI turns anecdotes into patterns
The raw material of a good ICP is already in your memory and your CRM — you just have not organised it. Every rep has a mental model of what a good client looks like. They know which deals felt easy and which were exhausting. They know which clients renewed without negotiation and which churned after six months of complaints. The problem is that this knowledge stays tacit, never gets written down, and is almost impossible to transfer to a new rep or apply consistently.
When you give AI structured descriptions of your best and worst clients, it does the extraction work for you. It looks for commonalities across the good ones — similar stages, similar buying triggers, similar decision-making patterns — and contrasts those against what made the bad one difficult. The output is a first draft of a pattern you already knew but had not articulated. Your job is to review it critically: where does the AI's pattern match your experience, and where has it missed something that is actually important?
The green flag and red flag model
The most practical element of today's output is the green flag and red flag lists. Green flags are specific, observable signals that a prospect is likely to close fast, at full price, and stay. Red flags are signals that should prompt you to either dig deeper before investing time, or qualify out early.
The critical word in both cases is specific. "Engaged leadership team" is not a green flag — it is a platitude. "The economic buyer joined the first discovery call without being asked" is a green flag. "They asked for a discount before they had seen a proposal" is a red flag. The more specific the signals, the more useful the list becomes in a real qualification conversation.
Use the checklist from today's output on your next new prospect. The habit of running a structured qualification check before you invest significant time in a deal is one of the clearest differentiators between average reps and top performers.
Prompt of the day
Copy this into your AI tool and replace any bracketed placeholders.
Prompt
You are a B2B sales strategist helping me build a working Ideal Customer Profile (ICP) — not a slide deck, but a practical tool I can use to qualify prospects and prioritise my pipeline. Here is the raw material I am giving you: My three best clients (describe each briefly — what they do, their size, their situation when they bought, why the deal worked, and what made them easy to retain or expand): - Client 1: [DESCRIBE] - Client 2: [DESCRIBE] - Client 3: [DESCRIBE] My worst client in the past 12 months (describe — what made them difficult, where the deal went wrong, what signs were there early that I missed): - [DESCRIBE] What I sell: [DESCRIBE YOUR PRODUCT OR SERVICE — what it does, the core outcome, typical contract size or deal structure] Based on this, please produce: 1. A firmographic profile: company size range, industry verticals, growth stage, and geography that best fits the pattern of my good clients 2. A behavioural profile: how good-fit prospects typically behave before they buy — what triggers their search, how they make decisions, what their buying process looks like 3. A green flag list: 6–8 specific signals that indicate a prospect is likely to be a fast, smooth deal 4. A red flag list: 6–8 specific signals that indicate a prospect will drain pipeline, discount heavily, or churn early 5. A qualification checklist I can run through in 10 minutes for any new prospect to give them a fit score of Low, Medium, or High Be direct. Use the language of a senior sales leader, not a marketing strategist.
Your 15-minute task
Write honest, specific descriptions of your three best clients and your worst client in the past 12 months. Do not generalise — use real details. Run the prompt. Read the green flag and red flag lists carefully and mark any that do not match your actual experience. Edit them. Then print or save the qualification checklist and use it on the next new prospect that enters your pipeline.
Expected win
A complete working ICP with firmographic profile, behavioural signals, a green flag and red flag list, and a 10-minute qualification checklist you can run on any new prospect from tomorrow.
Power user tip
After reviewing the output, send this follow-up: 'Now apply this ICP to my current pipeline. Here are 5 prospects I am currently working: [LIST THEM WITH A SENTENCE OF CONTEXT EACH]. Score each one as Low, Medium, or High fit using the ICP criteria you just built, explain the score in two sentences, and flag any red flags I should investigate before investing more time.' You will have a pipeline audit in under two minutes.