Sourcing a Functional Leader Who Scaled Through Your Stage

This guide walks one real stage-fit executive search from target-company selection to a ranked, reachable shortlist. It is for in-house recruiters, sourcers, talent leaders, and founders filling a first-or-next functional-leader role, who need someone who genuinely built and scaled that function at their company's current stage, not someone who held the title at a company ten times larger. By the end you can reconstruct what a candidate's past employer looked like during their exact tenure, screen out title-only and wrong-stage leaders, and rank the survivors by how tightly their proven stage matches yours.

The worked example running through this guide: a Series A B2B SaaS company, 20 people, $2M ARR, hiring its first Head of Sales. I carry that one case all the way, including the wrong turns.

Why stage-fit is the whole game for a leadership hire

The single largest cause of senior-hire failure is stage mismatch, not lack of ability, and it shows up on an 18-month clock. Four independent studies converge on the same band, and the practitioners behind them attribute the churn to fit, not talent.

Here is what the evidence says. HubSpot's own data, relayed through Sequoia, found about 40% of senior hires churn within 18 months and 60% stick. Broader executive research puts failure at 50% to 70% inside 18 months, with the cost reaching 400% of salary. For sales specifically the numbers are worse and remarkably consistent.

The mechanism matters. People self-select for stage, and the risk-seeking leader who thrived building from chaos is a different person from the one who fine-tunes a mature process. A screen for stage overlap attacks the dominant failure cause head-on. That is the bet this guide makes: if you can prove a candidate lived through your exact stage and built through it, you have removed the largest single reason the hire fails.

Two forces make this harder than it sounds. First, senior candidates interview brilliantly, and interviewers overestimate their own read on them, so the resume and the conversation both lie in the candidate's favor. Second, the public data you would use to check them drifts. The rest of this guide is the discipline that survives both problems.

The stage-fit clock is a fit clock, not a talent clock

The 18-month failure window is best read as a countdown to the moment a wrong-stage leader hits the phase they were never built for. The timing data supports this. VP Sales failures cluster: 23% in months one to six, 28% in months seven to twelve, and 16% in months thirteen to eighteen. The middle band is the largest, which is when a builder hired into an optimizer role, or an optimizer hired into a builder role, meets work their instincts do not fit.

This reframes the whole search. You are not looking for the most impressive leader. You are looking for the leader whose last proven stage matches your next twelve months. The Felicis executive-bench study, built from over 300 hours of research on 30 outlier companies, makes the function-by-function version of this concrete: 90% of Series C sales leaders were replaced or leveled before IPO across 21 public companies, and 100% of companies that had a C-level sales exec before Series B replaced that person by Series C. Sales leaders almost never survive their own scaling. Technology leaders, by contrast, were the earliest to have a C-level exec, usually a founder, and the most likely to scale through rounds.

For my worked example, this is decisive. I am hiring a first Head of Sales at Series A, which is the most stage-brittle function of all. The screen has to be strict, and it has to prove tenure sat inside the builder band before I trust anything else on the resume.

Fixing your own stage baseline and the function profile

Before you can match a candidate to your stage, you have to state your stage in one falsifiable line. Everything downstream is a comparison against that line, so vagueness here poisons the whole search.

For the worked case the line is: Series A, 20 people, $2M ARR, last round closed 18 months ago. That single sentence sets the target band the candidate's past tenure must overlap. Now translate it into a function profile using ARR bands rather than titles.

A builder VP creates process from chaos, a scaler VP systematizes what works, and an optimizer VP fine-tunes mature processes. At $2M ARR I need a builder. This is the fork where many searches quietly go wrong: a founder sees an impressive optimizer VP from a $200M ARR company and assumes bigger is better. It is not. Just because someone had success at Salesforce or HubSpot does not mean they will thrive at a $2M ARR startup, where there is no team to inherit and no playbook to tune.

The profile is set when it names one stage band, not a title. For the worked case: builder VP, 0 to $5M ARR, tenure that started under roughly 40 people and grew the function from near zero.

Building the target-company list from firms a few steps ahead

Build the list from companies that were at your stage two to four years ago and have since grown one to two stages, not from the biggest names you admire. This is the "few steps ahead" rule, and it beats the prestige rule for a specific reason: impedance mismatch.

Hires from companies orders of magnitude larger, the Googles and Microsofts, deal with different issues at a different scale, and dropping them into a 20-person company creates friction on every axis. Hires from admired companies just a few steps ahead work better because the problems they last solved are the problems you are about to have. So the target list for my Series A search is companies that were Series A two to four years ago and are now Series B or C.

Supply is not evenly distributed, and geography is a strategic lever you should set consciously before you start pulling names.

In Refolk's index there are 19,115 US VP or Head of Sales profiles against 352 in the UK, roughly 54 to one. If my Series A company were based in the UK, a stage-fit sales search would run into a wall almost immediately, and I would have to widen filters, accept relocation, or loosen the stage band. In the US the same search is comfortable, but a US finance search at the same seniority is tighter than sales by about 2.2 to one, so a finance leader search needs a wider net from the start. Knowing the supply shape before you build the list prevents a search that runs dry at step five.

For the worked case I ended with 28 companies whose growth window I could verify against funding dates. The list is done when every company on it has a growth trajectory you can check, not a name you recognize.

Reconstructing an employer's stage during the exact tenure window

This is the core skill. You reconstruct what a candidate's past employer looked like during their exact start-to-end dates by overlaying three dated sources, and you only trust the result when two of them agree. A single snapshot lies, because private-company data drifts.

Crunchbase and PitchBook update a private company's employee count mainly at funding events or press releases, while LinkedIn's count changes gradually as profiles update. So a headcount you read today can misdate a three-year-old tenure by a full stage. The stakes are high because the growth band is steep: Carta's H2 2025 benchmark shows the average venture-backed company grew from about 16 people at Series A to roughly 87 at Series C. One stale number can flip a builder into an optimizer.

Here is the reconstruction for one real candidate from my worked list, "Candidate J," listed as Head of Sales at a company from her start date through 26 months later.

Candidate: J
Employer: [B2B SaaS company on target list]
Tenure: start month -> +26 months

Funding dates (Crunchbase):   Seed at start-1yr; Series A at start+4mo; Series B at start+22mo
Headcount at start (Wayback): careers page capture shows ~18 named team members
Headcount at end (LinkedIn):  ~74 employees at tenure end
Press release corroboration:  Series B announcement cites "growing team of ~70"

Verdict: joined pre-Series-A at ~18 people, left mid-Series-B at ~74.
Overlapped the steep 16-to-87 growth band. BUILDER. Two sources agree at each end.

The order of operations is contested, and it is worth naming. Some practitioners lead with funding dates and infer headcount, others lead with headcount snapshots and infer stage. Both are valid, but you must reconcile both before deciding. When I led with Crunchbase alone on another candidate, the funding timeline looked like a clean builder story, but the Wayback capture of the careers page showed the team already at 60 people when he joined, because the company had raised late relative to its hiring. Funding stage and headcount had diverged. The candidate was an optimizer wearing a Series-A-timed resume.

Budget roughly 20 minutes per candidate for this. It is the most expensive step and the one that earns the shortlist its credibility.

Pulling the raw list by title across GitHub, LinkedIn, and the open web is the part that used to eat a full day of manual searching per pass. Describing the stage window in plain English and getting back people whose tenure already overlaps it is exactly the friction Refolk removes, so the 20-minutes-per-candidate reconstruction is spent on a list that is already close, not on names you will reject at first glance.

How this search goes wrong: the seven failure modes

The most valuable part of this standard is the list of ways a stage-fit search produces a confident wrong answer. Each failure mode has a false positive that looks like a strong signal and a check that exposes it.

1. The title-only leader, the "glorified AE." The false positive is strong personal quota numbers and big logos closed. Smashing quota and closing big deals does not mean a candidate can scale a team; the glorified AE is an individual contributor with a VP title who can sell but cannot hire, coach, or enable others. The check is the build-verification interview in the next section: ask how many people they hired and how many hit target.

2. Pedigree dazzle. The false positive is Salesforce, Google, or HubSpot on the CV. The check is the stage band: success at a scaled company says nothing about thriving at $2M ARR. Reconstruct the stage they built through, not the stage of the logo.

3. The post-scale joiner. The candidate joined after headcount crossed roughly 87, the Series C band, and inherited a built function. The check is the tenure start date against a dated headcount snapshot. This is the failure the reconstruction step exists to catch.

4. Headcount snapshot drift. The false positive is today's Crunchbase number applied to a three-year-old tenure. Sources disagree until updated to the same effective date, and a snapshot today can differ from one six months ago. The check is reconciling two dated sources to the tenure window, never one number to the present.

5. The single-source stage claim. The check is cross-referencing against two independent sources. If one headcount contradicts another, dig deeper before using either. My near-miss with the late-raising company was exactly this: funding stage and headcount had to be reconciled, not trusted individually.

6. Wrong function timing. Rejecting a People leader for lacking public-company experience misapplies the CFO rule. Per Felicis, People leaders scale and are promoted internally, while first CFOs join around Series C and are extremely unlikely to be promoted in. Match the rule to the function, and never carry the sales screen wholesale onto engineering or people.

7. Interview overconfidence. The false positive is a candidate who interviews brilliantly. Senior candidates are very good at interviewing, and interviewers overestimate their own skill at reading them. The check is to weight reconstructed evidence over interview impression, and to make the build questions specific and countable rather than open-ended.

A famous logo tells you the stage a candidate left, never the stage they built through. </pull>

The procedure, end to end

Run these eight steps in order. Each has an owner and a definition of done, and the whole sequence takes the worked example from 28 target companies to a ranked shortlist of six.

Step seven deserves its own attention, because it is where the reconstruction gets confirmed against the human. Dig into their history of building teams: how many reps they hired, how many hit quota, and their approach to onboarding and coaching. Standardized onboarding correlates with 62% higher new-hire productivity and 50% better retention, so a builder who cannot describe their onboarding system is a builder in title only. The reconstruction tells you they were in the room during the hard growth; these questions tell you they drove it rather than watched it.

Ranking, reachability, and keeping the search current

Rank the survivors by how tightly their proven stage matches your next twelve months, then attach a contact path to each. Tightness beats breadth: a leader who built exactly from your ARR to one stage ahead outranks a broader operator who spanned three stages but only optimized the last one.

Use a simple two-axis judgement to sort the final list.

Two things go stale and need re-checking rather than trusting once. First, public data sources shift: Crunchbase suffered a cyberattack on January 28, 2026, with over two million records reportedly stolen, so treat any single funding record as provisional and corroborate against a company blog or the Wayback Machine. Second, supply moves. The 54-to-1 US-to-UK sales-leader ratio and the 2.2-to-1 sales-to-finance ratio in Refolk's index are the shape of the market today, not a constant. Re-run the supply read at the start of each search, because a search that was comfortable in the US last quarter can run dry if you shift function or geography.

Keep the reconstruction worksheet from step five as your audit trail. When a hire works or fails, go back to the worksheet and check whether the stage call was right. Over a few searches you will calibrate your own read of where the builder band ends and the optimizer band begins, which is the one judgement no benchmark can make for you.