GTM Strategy

Your ICP Isn't a Spreadsheet. It's a Living Hypothesis.

Most teams define their ideal customer profile once and never revisit it. That's why their pipeline is full of accounts that will never close.

O
Own Outbound
January 25, 2026
5
min read

Somewhere in your company's Google Drive, there's a document titled something like "ICP Definition" or "Target Account Criteria." It was written during a strategy offsite. It lists firmographic attributes — industry, employee count, revenue range, tech stack — and maybe a few persona descriptions. Everyone agreed on it. Nobody's updated it since.

That document is quietly costing you pipeline every single week.

The Snapshot Problem

Here's what happens. A founding team gets together early on and defines their ICP based on a mix of intuition, early wins, and investor feedback. It looks something like: B2B SaaS, 50–200 employees, Series A to B, US-based, uses Salesforce. They hand it to the sales team and say "go."

And the sales team does go. They build lists against those filters, run sequences against those lists, and report pipeline against those accounts. Months pass. Some deals close. Many don't. And when leadership asks why win rates are low, the answer is always about execution — the emails weren't good enough, the reps need more training, the follow-up cadence needs work.

Almost nobody asks the more fundamental question: are we going after the right accounts in the first place?

The ICP isn't wrong because the team was careless. It's wrong because it was built on a snapshot of the business that no longer exists. Markets move. Products evolve. What you're actually good at selling becomes clearer over time — but only if you're paying attention.

What Your Closed-Won Data Is Trying to Tell You

Every company has a goldmine of signal sitting in their own CRM that they're not using. Your closed-won deals — the accounts that actually bought, stayed, and expanded — contain patterns that are far more reliable than any theoretical ICP exercise.

But most teams never do the analysis. They'll spend weeks building outbound sequences and hours on email copy, but they won't spend an afternoon looking at what their best customers actually have in common.

Questions your closed-won data can answer

What triggered the buying process? Not the first touch — the business event that made them start looking. Funding round? New hire? Competitive loss? Product launch?

How long was the sales cycle? If deals under 200 employees close in 30 days but deals over 500 take 6 months, your ICP should reflect that — especially if you need near-term revenue.

Who was the internal champion? Not the signer — the person who fought for you. Is it always a VP of Sales? A RevOps lead? An SDR manager? That tells you exactly who to target.

Which deals expanded? Your best customers aren't just the ones who bought — they're the ones who grew. What do those accounts share that your churned customers don't?

When I've run this exercise with founders, the results almost always surprise them. The ICP they've been targeting doesn't match the profile of their best customers. They think they sell to mid-market SaaS companies, but their fastest-closing, highest-retention deals are actually with professional services firms going through digital transformation.

Or they think their buyer is the CRO, but their champion is consistently a director-level ops person who found them through a peer recommendation.

The data doesn't lie. But you have to actually look at it.

From Static to Living

The fix isn't complicated, but it does require a mindset shift. Your ICP isn't a document to be written and filed. It's a hypothesis to be tested and updated — continuously, based on real outcomes.

Static ICP

Defined once at an offsite

Based on firmographics

Updated annually (maybe)

Treats all matching accounts equally

Living ICP

Refined after every closed deal

Based on behavioral triggers

Reviewed monthly with real data

Prioritizes by timing and signal strength

In practice, this means a monthly ICP review. Thirty minutes. Pull up every deal that closed in the last 30 days and every deal that was lost. Look for patterns. What changed in the winning accounts before they entered the pipeline? What was different about the ones that stalled?

Then update your targeting. Not dramatically — incrementally. Tighten one criteria. Loosen another. Add a new signal. Remove one that isn't predictive. Run the updated target list for the next month and measure again.

The Signal Stack

What I've found most effective is building what I call a signal stack — a ranked list of observable indicators that predict buying intent for your specific product. It's different for every company, but the structure is the same.

At the top of the stack are high-confidence signals — events that historically precede a purchase. A new VP of Sales joining a company that matches your profile. A funding round at a company already using a complementary tool. A job posting that describes the exact problem you solve. These are the accounts you prioritize immediately.

In the middle are moderate-confidence signals — things like website visits, content downloads, or industry trends that suggest growing interest. These are worth watching but not worth an immediate all-out push.

At the bottom are baseline fit criteria — the firmographic filters that define your broad addressable market. These are necessary but not sufficient. An account can match every firmographic criteria and still be a terrible prospect if the timing isn't right.

The best outbound teams I've seen don't just know who could buy. They know who's likely buying soon. That distinction is the entire game.

What This Looks Like in Practice

I work with a founder who used to send 2,000 emails a month against a static ICP. Win rate was around 3%. After rebuilding the ICP as a living hypothesis — reviewing closed-won data monthly, building a signal stack, and targeting accounts with active buying triggers — volume dropped to 400 emails a month.

Win rate went to 11%. Pipeline value actually increased because the accounts they were targeting were better fits with larger deal sizes. And the sales team stopped burning out on rejection because a higher percentage of their conversations were with people who actually wanted to talk.

That's the compounding effect of a living ICP. Better accounts lead to better conversations, which lead to better data about what works, which leads to even better account selection. The loop tightens over time.

Your spreadsheet ICP can't do that. A living hypothesis can.

Start with what you know — your last 10 closed-won deals. Find the patterns. Build the signal stack. And commit to updating it every month. The returns will surprise you.

O
Own Outbound

Helping founders and GTM teams move from activity to accuracy. Exploring the intersection of AI, outbound strategy, and human judgment.

Keep Reading