ICP drift
your icp is not who it was
12 months ago.
|
Summary
There is a document somewhere in your company that describes your ideal customer.
It was written during a planning session — maybe a company offsite, maybe a positioning sprint — and everyone agreed it captured exactly who you were building for. Alice: a VP of Sales at a 50-person B2B SaaS company, technically literate, frustrated with manual prospecting, looking for a tool that saves his team three hours a day.
Alice was right. Six months ago.
Latest insights. Straight to your inbox
01 – The ICP drift
when did you last check in with Alice?
The problem is nobody has talked to Alice recently. Not the current Alice , anyway. And the current Alice has different questions, different expectations, and a different baseline of sophistication than the Alice you originally described. Your messaging is still written for the old one.
That is not a minor inconsistency. It is a slow leak in everything downstream.
Across the 300+ companies I have worked with or mentored on positioning, the pattern is consistent: ICP work happens once. It gets done at the beginning, or during a growth plateau when someone decides “we need to get clearer on who we serve,” and then it gets filed. It becomes a document. A reference artifact. Something that lives in Notion and gets cited in sales onboarding decks.
Then the market moves, and nobody updates the document.
This is not laziness. It is the nature of how companies think about ICP work — as a deliverable with a start and an end, rather than a continuous signal you stay close to. Lincoln Murphy, who has spent more time than most people thinking about customer success and ICP design, put it plainly:
“Your Ideal Customer Profile is a living, breathing ‘definition’ that you’ll come back to — and modify — often.”
The companies that actually treat it that way are genuinely rare.
There are three signals hiding in your business right now that will tell you exactly who Alice has become. None of them require a research sprint, a consultant, or a dedicated budget. They require listening to what is already happening — in your sales calls, in your onboarding flow, and in your email replies. We will get to those. First, it is worth understanding why Alice drifted in the first place.
ICP is not a strategy. It is a hypothesis. And hypotheses need updating when the evidence shifts.
Work with Andrej
Is your growth stalling?
A quick discovery call to find out if positioning is the problem.
"Your Ideal Customer Profile is a living, breathing 'definition' that you'll come back to — and modify — often."
02 – Root cause
THE MARKET changed. your buyer profile didn't.
Here is the part nobody talks about: your ICP can drift without any strategic decision on your part.
You do not have to pivot. You do not have to change your pricing or launch a new product line. The category matures around you, and the people walking through your door are materially different from the people who walked through it a year ago. Same product. Different buyer.
Geoffrey Moore mapped this in Crossing the Chasm. As a product moves from innovators and early adopters into the early majority, the buyer profile shifts fundamentally. Innovators tolerate rough edges in exchange for vision. Early majority buyers want proven ROI, workflow fit, and peer validation before they commit. The questions are completely different. The objections are different. The language is different.
If your messaging is still calibrated for the early adopters who first believed in you, you are invisible to the pragmatists who now make up the majority of your addressable market.
The customers who stayed are not your whole ICP. They are a sample of one cohort. The new ones arriving are often a different cohort entirely.
03 – the reason
AI changed who your buyer is
Partly, I blame AI. And I mean it.
According to Forrester’s 2024 Buyers’ Journey Survey, 89% of B2B buyers have adopted generative AI as a primary self-research tool across every phase of the buying journey. The 6sense Buyer Experience Report puts it even higher: 94% of buyers now use LLMs during their buying process. Gartner has been tracking the downstream effect for years: 80% of the B2B buying journey now happens without vendor contact, up from 57% in 2015.
Forrester framed the implication directly:
“B2B buying today is a process of confirmation, not selection. Decisive buyers already know who they want to work with before they start gathering requirements or talking to vendors.”
The person researching your category in 2023 had to read blog posts, watch demos, and talk to peers to understand the space. That took weeks. The same person in 2026 asks ChatGPT or Perplexity a few pointed questions and arrives at your door already knowing your category, your competitors, your pricing model — before they have ever spoken to your team.
This is exactly what I watched happen at Aimfox. Six months ago: users asking whether LinkedIn automation was safe. Whether it was against LinkedIn’s terms. Today: do we have residential IP addresses? What is the warm-up protocol for new accounts? The product did not change. The buyer did.
What happened next — how this shift quietly ate our growth metrics — I covered in the previous post on why SaaS growth stalls are often a positioning problem. The short version: by the time we noticed, the drift had already cost us pipeline. That is how it always goes.
Informed buyers do not need your 101 content. They need to know you understand their specific situation. If your ICP hasn’t caught up to who they now are, your messaging won’t.
"B2B buying today is a process of confirmation, not selection. Decisive buyers already know who they want to work with before they start gathering requirements or talking to vendors."
04 – The pattern
FIVE SIGNS YOUR ICP HAS ALREADY DRIFTED
Not collapsing — softening. MQL-to-SQL rates down a few points. Cost per qualified lead climbing without a corresponding change in spend. The funnel looks the same on paper. The output has quietly changed.
01
Conversion is softening without a clear cause.
02
Sales is rejecting leads that look right on paper.
03
Churn is rising in recent cohorts specifically.
04
Onboarding friction is increasing.
05
The team disagrees on who the product is really for.
05 – The fix
THREE SYSTEMS TO KEEP UP WITH alice
Work with Andrej
Is your growth stalling?
A quick discovery call to find out if positioning is the problem.
Talk to sales about what they are actually hearing.
Not CRM data — actual language. What kinds of questions are coming in? What do prospects assume the product can do that it can’t? What did they ask about that you hadn’t heard six months ago? That is a 20-minute conversation per week. The signal is in the language people use when they don’t yet know whether you are right for them. That is when the ICP is most visible.
Use your onboarding survey as an intelligence instrument
At Aimfox, our onboarding survey qualifies users — not for their sake, but for ours. High-ticket accounts go to a sales conversation. Low-ticket users are routed to the product itself. The side effect: we continuously collect data on who is arriving — their role, their use case, their current stack, their expectation of what the product will do for them. A rolling picture of ICP, not a snapshot from a planning session.
Train your users to reply to your email automations.
At Aimfox, we write certain automated emails to invite a reply — not with a survey link or a book-a-call button, just a genuine question a real person would naturally respond to in a sentence or two. And when they do, we read every single one. The language in those replies is unfiltered ICP intelligence. Not what users say when they think they’re filling in a form. What they actually mean. The voice of the customer is most useful when the customer is not performing for an audience.
06 – what everyone misses
Alice HAS MOVED ON. HAVE YOU KEPT UP?
At some point in the next 12 months, you are going to look at a lead source or a cohort report and notice that something is not performing the way it used to. Conversion has dropped. Sales is pushing back. Churn in recent cohorts is above your historical average.
The instinct will be to optimize. Change the landing page. Rework the sequence. Add a case study.
Before you do that, ask a simpler question: when did you last actually check in with Alice ? Not the Alice from your last positioning session — the Alice who is arriving at your product today. The one who has already read three comparisons of your category on ChatGPT, who has a specific technical question before they even speak to a human, and who will decide on fit within the first 90 seconds of using your product.
That Alice has different needs. Different language. A different set of things that will make him say “yes, this is for me.”
If you have not updated your picture of him, your funnel is telling a story to someone who has already left the building.
