ICP Decay: Why Your Recruitment Agency's ICP Is Wrong by Q3 (and How to Refresh It Quarterly)

TL;DR

Most recruitment agencies define their ideal client profile once, at a January planning session or when a new desk launches, and then keep prospecting against it for years. The market does not wait. UK permanent placements only stabilised in July 2026 after the longest downturn on record, while temporary billings rose for a fourth consecutive month [1]. A profile written for last year's hiring conditions is already pointing consultants at the wrong companies. The data underneath it decays too: commonly cited B2B benchmarks put contact data decay at roughly 22.5% a year and often far higher [2]. The fix is not a bigger ICP workshop. It is a short, repeatable quarterly ICP refresh: treat the ICP as a hypothesis, test it against your own placements, win-loss debriefs and reply data, change at most two or three criteria, and write the result somewhere that survives consultant churn. This guide gives you the process, a drift checklist, the KPIs to watch and a 90-day plan. boilr's Company Brain is the shared memory that holds the current ICP and the evidence behind it, so a new hire inherits this quarter's truth rather than last year's.

Why an ICP Decays (Faster Than You Think)

An ideal client profile is a prediction: "companies that look like this will brief us, pay our fee and come back." Predictions go stale when the conditions behind them change. In recruitment BD, several conditions change at once:

• Hiring demand rotates between sectors. The KPMG and REC Report on Jobs for July 2026 describes stabilising permanent placements, subdued permanent demand and strengthening temp growth [1]. A profile built around permanent mandates in one sector can quietly stop matching where briefs now come from.
• Fee pressure moves the economics. When candidate supply is rising and clients feel leverage, the companies that used to accept your standard fee start asking for discounts, PSL terms or contingent-only arrangements. A client that fitted at 20% may not fit at 14%.
• Client behaviour changes. In-house talent teams grow, a competitor wins the PSL, a founder who always called you has left. The firmographics are unchanged; the buying behaviour is not.
• Your own delivery changes. A desk that has just placed three platform engineers in fintech Series B companies is better at that than at anything else you wrote on the ICP slide.
• The data layer rots. Contact and company data decays steadily as people change jobs and companies reorganise; commonly cited benchmarks range from about 22.5% a year upwards [2]. An ICP list that was accurate in January is partly wrong by autumn even if the strategy is still right.

There is no single research-backed "ICP half-life", and most published guidance on cadence is practitioner advice, not a controlled study. The consistent recommendation is a quarterly review, with an extra one after major changes such as a pricing change or a new market [3] [4]. Treat "wrong by Q3" as a working assumption to test with your own numbers, not a law.

What decay looks like on a recruitment desk

Symptom	What is probably happening
Reply rate falls on a segment that used to convert	The segment's hiring has cooled, or your message no longer matches their problem
You win briefs but at lower fees every quarter	Your ICP selects for price-sensitive buyers
Consultants "ignore" the score and pick their own targets	The scoring rules reflect an old profile and nobody trusts them
Lots of first meetings, few briefs	You are reaching companies that fit on paper but lack the budget or urgency
One desk performs, the rest do not	The real ICP lives in one consultant's head, not in the agency

ICP as Hypothesis: The Mindset Shift

If you have not yet written a first version, start with our guide to defining an ICP for recruitment agencies. This article assumes you have one and asks the harder question: is it still true? Practitioners who recommend revisiting the ICP on a fixed rhythm make the same point, that the profile you state and the customers who actually buy drift apart unless you compare them on purpose [3]. Some argue for a lighter cadence of twice a year [7]; for an agency whose briefs shift with the hiring market, quarterly is the safer default.

• Write it as a claim. "Series A to C fintechs with 50-250 staff, hiring their first Head of Engineering, will brief us on retained or exclusive terms." A claim can be proven wrong. A mood board cannot.
• Attach evidence and a date. Every criterion carries the data that supports it and the quarter it was last checked.
• Limit changes per cycle. Two or three edits per quarter. More than that and you cannot tell which change moved the numbers.
• Keep a "watch list". Segments you do not target yet but whose early signals look promising get a small test budget, not a full pivot.

The Quarterly ICP Refresh Process

The whole review should fit in one half-day workshop plus a few hours of data prep. Run it in the first two weeks of each quarter, ideally with the agency owner, one lead from each desk and whoever owns your CRM data.

Step 1: Pull your own data, not the market's

Generic market reports tell you what is happening to hiring in general. Your own records tell you what is happening to your revenue. Export the last 90 days from your CRM or ATS (Bullhorn, RecruiterFlow or similar):

1. Placements and fees - by client, sector, size, role level and fee percentage.
2. Briefs received vs. briefs won - including terms (retained, exclusive, contingent, PSL).
3. Outreach outcomes - replies, meetings and briefs by segment and by signal type.
4. Lost and stalled deals - with the reason, taken from your win-loss debriefs.
5. Time to first brief and time to placement - by segment.
6. Client repeat rate - which new clients from the last 12 months briefed you a second time.

A common rule of thumb in sales advice is to revisit the profile at least quarterly or after a fixed number of closed deals [4]. For a small agency, "every 20 closed-won deals" may mean twice a year, so use whichever comes first.

Step 2: Compare stated ICP against actual winners

Build a simple two-column view. On the left, your written ICP. On the right, the profile of the clients who actually placed with you at good fees in the last two quarters. The gaps are your drift.

Dimension	Written ICP	Actual winners (last 2 quarters)	Gap?
Sector	Fintech	Fintech plus climate-tech	New segment emerging
Company size	50-250 staff	20-80 staff	Skewing smaller
Trigger	Funding round	Exec hire / new office	Trigger has changed
Buyer	Head of Talent	Founder or COO	Different decision maker
Terms	Retained or exclusive	Mostly contingent	Economics shifted

The table above is an illustrative template, not agency data. Fill it with your numbers.

Step 3: Run the drift checklist

For each ICP criterion, ask these questions. Any "yes" is a candidate for an edit.

• Demand drift - has hiring volume in this segment risen or fallen versus last quarter in your own briefs and in the market [1]?
• Fee drift - has the average fee percentage or the share of discounted terms moved by more than a couple of points?
• Buyer drift - are briefs now arriving from a different job title than the one you target?
• Trigger drift - are the events that precede briefs (funding, leadership change, expansion, job-posting spikes) still the same ones?
• Competitor drift - are you losing more to in-house teams, a new PSL or a specific rival?
• Delivery drift - where are your best recent placements, regardless of the ICP slide?
• Data drift - how much of the target list has changed contacts, headcount or status since you last cleaned it [2]?
• Scoring drift - do the highest-scoring accounts still convert best? If not, revisit the weights in your lead scoring model.

Step 4: Decide, then change only a few things

Sort every proposed change into one of four buckets:

Decision	When to use it	Example
Keep	Evidence still supports it	Fintech Series A-C stays core
Tighten	Fits on paper, converts poorly	Add "has hired externally in the last 12 months"
Expand	Winners appear outside the ICP	Add climate-tech as a test segment
Retire	Repeated losses or fee erosion	Drop sub-£10m revenue companies

Pick at most two or three changes. Write each as a dated, testable claim with a success measure for next quarter.

Step 5: Push the change into every system that uses it

• Targeting - update search filters and company lists.
• Scoring - update the weights so priority ranking reflects the new profile.
• Messaging - update the angles and proof points per segment.
• Signals - update which hiring signals matter for the new trigger mix.
• Onboarding - update what new consultants are told on day one.

This is where most refreshes fail. The ICP is rewritten in a slide, the slide goes into a shared drive, and the consultants carry on with the old filters.

Manual Refresh vs. a Company Brain Workflow

Part of the refresh	Manual (spreadsheet and slide)	With boilr
Finding what converts	Hours of CRM exports and pivot tables	The Company Brain tracks which ICP segments produce replies and meetings [6]
Spotting new triggers	Anecdotes from the Monday meeting	Signals show funding, exec moves, expansions and job-posting velocity as they happen
Applying the new ICP	Email to the team, hope for the best	Scoring and prioritisation use the updated profile
Onboarding a new hire	Share an out-of-date deck	The new consultant starts with the agency's accumulated knowledge [6]
Keeping history	Lost when the file or the person goes	Stays with the agency, not the individual

KPIs: How to Know the Refresh Worked

Measure at the segment level and compare quarter to quarter. Do not judge a change on fewer than a few dozen contacts in a segment.

KPI	What it tells you	Direction after a good refresh
Reply rate by segment	Message and market fit	Up in core, flat or down in retired segments
Meeting-to-brief rate	Whether you reach buyers with real need	Up
Average fee percentage	Economic quality of the ICP	Stable or up
Time to first brief	Urgency of the segment	Down
Repeat-client rate	Long-term fit	Up
Score-to-conversion correlation	Whether scoring still predicts wins	Stronger
Share of BD on ICP-matched accounts	Whether the team actually follows the ICP	Up
ICP last-reviewed date	Process discipline	Within 90 days

Why Consultant Churn Makes ICP Decay Worse

Sales-team data suggests average rep tenure is around 18 months and annual turnover around 35% in the US, according to one compilation of HubSpot and Xactly figures [5]. Recruitment-specific numbers vary by agency, but the pattern is familiar to any owner: consultants move on, and their sense of who the good clients are goes with them.

• The ICP lives in people. The best-performing biller "just knows" which companies to call. That knowledge is never written down.
• New hires restart the experiment. They prospect the segments the previous consultant already found to be dead ends.
• Refreshes lose their lineage. Without records of why a criterion changed, the next review argues the same points again.
• Desks diverge. Each team keeps its own version of the profile, so the agency has three ICPs and none is current.

The Company Brain is built for this problem. According to boilr, it learns passively from every outreach a consultant verifies and sends, with no tagging or training uploads, and the knowledge belongs to the agency and stays when people leave [6]. For the loss-reason side of the same problem, see our guide to win-loss analysis for recruitment agencies.

How boilr Supports the Quarterly Refresh

boilr is an AI sales employee for recruitment, one per consultant. It does the research and drafting; the consultant verifies and sends. These are the modules that matter for an ICP refresh:

• Company Brain - the shared memory of which ICPs, messages and signal patterns work for your agency [6].
• Companies - finds and enriches client companies that match the current profile.
• Candidates - sources candidates alongside the company side, so a new segment can be tested on both sides.
• Signals - detects funding, exec moves, expansions and job-posting velocity, often before job boards, which lets you test a new trigger quickly.
• ICP scoring - ranks accounts against your ICP so consultants work the best fit first.
• Tasks - drafted outreach waits for the consultant to verify and send, and those verified sends feed the learning loop.
• Integrations - connects with Bullhorn, RecruiterFlow, Spott, CRMs, calendars and email, so placement and fee data can stay in your system of record.

What to keep human

• The decision to enter or exit a segment - that is a commercial judgement about your desks and your risk appetite.
• Fee and terms strategy - boilr can show you the pattern, but you set the floor.
• Win-loss conversations - the candid client debrief is a human call.
• Final verification of every message - the consultant always reviews and sends.

Want your ICP and the evidence behind it to survive the next consultant who leaves? Book a demo or start with boilr and see the Company Brain on your own agency's data.

Common Mistakes in an ICP Refresh

• Refreshing on vibes - changing the ICP because of one big deal or one lost pitch. Use a quarter of data.
• Changing everything at once - you will not know what worked. Cap it at two or three edits.
• Counting meetings, not fees - a segment that books easily but pays badly is a trap. Weigh fee and repeat rate.
• Ignoring the wins outside the ICP - your best new client in months may not match the slide. Investigate it.
• Letting marketing or the owner write it alone - the people who place candidates know which clients are real.
• Not updating the scoring - a new ICP with an old lead score still ranks the wrong accounts first.
• Treating the refresh as a document - if it does not change filters, scoring and messaging, nothing changed.
• Skipping the log - without a record of what changed and why, the next review starts from zero.

A 90-Day Plan to Install the Habit

When	Action	Output
Week 1	Export 90 days of placements, briefs, outreach outcomes and loss reasons	One dataset per desk
Week 2	Half-day workshop: stated vs. actual ICP, drift checklist	List of proposed changes
Week 3	Choose two or three changes, write them as testable claims, update filters, scoring, messaging	ICP v-next with dates and evidence
Weeks 4-10	Run the new ICP; track the KPIs by segment	Mid-quarter dashboard
Week 11	Debrief the segments: what the data says	Keep / tighten / expand / retire notes
Week 12	Log the lessons in the Company Brain and book the next refresh	Dated change log, next date in the calendar

1. Put the next review in the calendar before you leave the room.
2. Name an owner for the ICP - one person, not a committee.
3. Keep a single change log that every new consultant reads in week one.

Frequently Asked Questions

How often should a recruitment agency refresh its ICP?

Quarterly is the most common recommendation in sales guidance, with an extra review after big changes such as a new market, pricing change or major shift in where briefs come from [3] [4]. Small agencies with few closed deals per quarter can combine a quarterly check with a deal-count trigger.

Is an ICP really wrong by Q3?

Not always. "Wrong by Q3" is shorthand for the idea that a profile written in January is based on January's market. In fast-moving sectors it can drift within months, in stable ones it may hold. The point of a quarterly review is that you find out with data instead of guessing.

What data do I need for an ICP refresh?

Your own placement and fee data, briefs received and won, outreach outcomes by segment, loss reasons and repeat client rates. Market reports such as the KPMG and REC Report on Jobs add context [1], but your CRM or ATS is the main source.

How many criteria should I change per quarter?

Two or three. Fewer changes make it possible to attribute results. If the data demands a larger pivot, treat it as a separate project with its own test plan.

How is an ICP refresh different from win-loss analysis?

Win-loss analysis explains why individual pitches were won or lost. The ICP refresh uses that evidence, plus placement, fee and outreach data, to decide which kinds of client to target at all. They feed each other; start with our win-loss guide.

Does my lead scoring need to change when the ICP changes?

Yes. Scoring is how the ICP reaches daily work. After each refresh, update weights and check whether the top-scoring accounts still convert best. See our lead scoring guide.

How does the Company Brain help with consultant turnover?

According to boilr, the Company Brain stores which ICPs, messages and signal patterns work for the agency and belongs to the agency, so new hires start with that knowledge rather than from zero [6]. Without such a shared record, the ICP tends to live in individual heads and leaves with them.

What if I do not have much data yet?

Run the refresh anyway with what you have, and say clearly where the sample is small. Use the quarterly rhythm to start recording the data you will need next time. A small, honest review beats none.

Sources

Information sourced from public industry reports as of October 2026.

1. KPMG and REC UK - Report on Jobs, August 2026
2. Landbase - B2B data decay rate statistics (vendor compilation)
3. GetAccept - When and how to update your ideal customer profile
4. Prospeo - How to create an ideal client profile (2026 guide)
5. Gangly - Sales rep turnover statistics (citing HubSpot and Xactly)
6. boilr.ai - Company Brain
7. Mind the Product - How revisiting your ICP twice a year can save your B2B product