When to Update Your ICP: Signs It Has Gone Stale

An ideal customer profile (ICP) is a description of the companies most likely to buy from you and keep buying. It is a snapshot. Your product, your pricing and your competitors keep moving, and the snapshot ages.

It is easy to miss. Outreach keeps going out to the same list, and the numbers drift. This post covers the signals that say your ICP needs a second look, a review routine that does not eat a week, and how to re-run the analysis so that changes are approved before outreach changes. If you have not written an ICP yet, start with How to Define an Ideal Customer Profile. For the term itself, see the ICP glossary entry.

How often to review

There is no honest universal number. How fast your market moves depends on your product, your sales cycle and how many deals you close. A fixed interval copied from someone else's blog is a guess.

Use two triggers instead:

A date you set and keep. Attach the review to a planning meeting or pipeline review your team already holds, so it does not need its own invite. Put it on the calendar and keep the review short.
A signal. If any item in the next section shows up, review now, whatever the calendar says.

The date catches slow drift. The signals catch sudden change.

What changed: signs your ICP has gone stale

Each sign below means something about your buyers or your market has changed. None of them proves the ICP is wrong. Each one is a reason to look at recent deals.

Reply rates are falling

Pick one definition of reply rate and keep it. Belkins' 2026 cold email study, which covered 7.5M B2B emails sent in 2025, defines reply rate as replies divided by emails sent [[1]](#sources). The same study found that 0.45% of emails get a reply across industries [[2]](#sources). Treat that as an outside baseline, not a target.

Your own trend matters more than that baseline. Compare your reply rate this period with your reply rate in earlier periods, and compare it by segment. One way to read the pattern is to check two things. If the drop is across everything, look at message and deliverability as well as the ICP. If one segment drops while another holds steady, check whether the ICP is the cause.

Smartlead sends reply, bounce, unsubscribe and category-change events back to Outlevo by webhook [[3]](#sources). Those events give you the counts to compare by segment.

New loss reasons appear

Read the closed-lost reasons on your recent deals. A reason you did not see a year ago is a signal. It might be a missing feature, a new compliance need, or a buyer type who now blocks the deal.

Outlevo reads each deal's pipeline, stage, amount and closed-lost reason from HubSpot [[4]](#sources). From Salesforce it reads accounts, opportunities and their stages [[5]](#sources). If you keep deals elsewhere, a CSV import can carry stage, amount, outcome and loss reason [[6]](#sources).

A competitor keeps showing up

If a competitor appears in more of your late-stage deals, one hypothesis to test is that your ICP includes accounts that competitor serves well. Outlevo's competitor tracking covers quoted facts from competitors' public pages, changes, win/loss by competitor and competitor signals on accounts [[7]](#sources). Win/loss by competitor shows whether you lose to one vendor in one kind of account.

Deal sizes are shifting

Look at the amount on won deals. If recent wins are consistently larger or smaller than the ones your ICP was built on, the profile may describe the wrong size of company. There is no standard threshold for "meaningful" here. Decide with your sales lead what size of shift would change who you target, and write it down before you look at the data.

Your wins do not look like your ICP

Each qualified account in Outlevo gets a Readiness Card with fit, intent and data quality scores [[8]](#sources). If accounts that score well on fit keep losing, and accounts that score poorly keep winning, the fit definition is out of date.

Website visitors point the same way. Outlevo shows which companies visit, which pages they read and whether they fit your ICP [[9]](#sources). A steady flow of non-fitting companies reading your pricing page is worth a look.

How to re-run the analysis

You do not need a long project. Re-run the same analysis on recent closed deals.

1. Pull recent won and lost deals. Outlevo learns your ideal customer from the deals you won and lost in HubSpot [[4]](#sources). You can also import past deals from a CSV or your CRM, which produces a win/loss analysis that proposes ICP hypotheses [[10]](#sources). A CSV import takes up to 200,000 rows [[6]](#sources). See CSV and Excel import and export for the fields you can map. Connecting and syncing a CRM uses no credits [[4]](#sources), and importing a CSV uses none either [[6]](#sources).

2. Mind the sample size. Win/loss analysis counts companies, not deals [[11]](#sources). A company with several closed deals is one company in that count, not several. Small samples are labeled, and under 10 won companies is very low confidence [[11]](#sources). If your recent window has few wins, treat what comes back as a lead to follow, not a conclusion. Widen the window or wait for more closed deals.

3. Compare against the current ICP. Look for segments that win more than the ICP expects, segments that lose more, and loss reasons that are new. Write the differences down in plain words.

4. Cross-check with your public site. The free ICP assessment reads up to 5 public pages of a website and returns up to 5 experimental segment hypotheses [[12]](#sources). It gives you a view of how you present yourself, to set beside what your deals show.

Get changes approved before outreach changes

A new analysis proposes. A person decides.

AI-proposed ICP segments are labeled experimental until a person approves them [[11]](#sources). A changed ICP does not take effect without that sign-off. Market scope also stays with you. Outlevo asks which countries you sell into and never proposes markets [[11]](#sources).

Approval also governs what goes out. Nothing is sent outside a policy that a person approved [[13]](#sources). Policies set who can be contacted, on which channels, how often and with what budget [[13]](#sources). Once you approve a revised ICP, review the policy for each campaign that used the old one. Every external action is checked against the approved policy before it runs, and recorded in a ledger [[13]](#sources).

If you use HubSpot, Outlevo writes the ICP version, fit score and reason to company fields it owns, and nothing is written until an owner or admin approves the field mapping. Your standard HubSpot fields are never edited [[4]](#sources).

If something looks wrong after a change, kill switches pause a campaign, a channel or the whole workspace at once [[13]](#sources).

What to do

Set a review date. Choose one you will keep. Make the review short.
Name your triggers. Agree with sales on what counts as a real drop in reply rate and a real shift in deal size. Write the thresholds down before you look at the data.
Track reply rate by segment. Use one definition, replies divided by emails sent [[1]](#sources), and compare against your own history first.
Read the closed-lost reasons. Note any that are new since the last review.
Check competitor appearances. Look at win/loss by competitor [[7]](#sources).
Re-run the win/loss analysis on recent deals. Count companies, and treat fewer than 10 won companies as very low confidence [[11]](#sources).
Have a person approve the changes. Experimental segments stay experimental until then [[11]](#sources).
Update policies, then outreach. Adjust who can be contacted, on which channels, how often and with what budget [[13]](#sources). Change the campaigns last.

For how an ICP differs from a buyer persona and a total addressable market, read ICP vs buyer persona vs TAM. To see how approvals, policies and the ledger fit together, see How Outlevo works.