The Executive Search Process, Step by Step
Descriptions of the executive search process tend to be written for clients, which makes them accurate and useless: intake, research, longlist, shortlist, interviews, offer. True, and it tells you nothing about why searches go wrong.
The more useful version names, for each stage, what the stage is actually for and how it fails — because most search failures are not visible in the stage where they happen. They surface two stages later, disguised as something else.
What follows is the sequence as we run it. Firms order and name these differently, and the labels matter far less than whether each job actually gets done by someone.
1. Intake and brief
What it is for: turning a client’s description of the role into criteria that research can be run against.
This is not a job-description exercise. The output of intake is a rubric: a set of requirements each of which can be held against evidence and produce a result. “Strong commercial instincts” is not one of those. “Has carried a P&L of €50M or more” is.
Intake is also where you find out which stated requirements are real. Almost every brief contains at least one criterion that is in the document because it was in the last document. One question separates the must-haves from the inherited preferences, and it is the most valuable thing said in the meeting: would you still want to meet someone who cleared every other bar and missed this one?
How it fails: vague criteria get accepted because everyone in the room understands what is meant. Six weeks later, nobody can reconstruct what was meant, and the shortlist gets argued about instead of assessed. Getting the brief specific enough is upstream of everything else.
2. Market mapping and discovery
What it is for: assembling the pool of people who could plausibly be considered, before anyone is evaluated.
A first pass worth the name casts wide — wide enough that a plausible candidate is not lost because the framing was too narrow at the start. How wide depends on how well-mapped the sector is. It draws from several source layers — professional profiles, company org structures, appointment announcements, filings, published work, and deliberate mapping of adjacent sectors where the operating constraints are similar.
The discipline here is holding off on judgment. Discovery and evaluation are different activities, and running them together produces a narrow pool, because you calibrate to the type of person you have already been finding.
How it fails: quietly. A candidate who was never entered is never missed, and there is no downstream stage that goes looking for them. Most of the ways a pool ends up too narrow are invisible from inside the search.
3. First-pass filter
What it is for: removing, quickly, the people who could not qualify no matter what deeper research found.
Function, seniority band, sector, geography, hard disqualifiers. Signals that are quick to check and require no interpretation. Most of the pool leaves at this stage, and it should.
How it fails: the filter starts doing work it is not equipped for. Someone begins assessing quality here rather than plausibility, and cuts get made on a title or a company name, at a glance, with no record of the reason. Staging the filter properly is what keeps this stage honest.
4. Deep research on the survivors
What it is for: establishing what each remaining candidate has actually done, against each criterion in the rubric individually.
Scope rather than title. Sources rather than self-description. Every confirmed criterion carries a specific, locatable citation, and every criterion returns one of three verdicts: confirmed, unconfirmed, or flagged as ambiguous with the reason stated.
This is the expensive stage in time and attention, which is exactly why it runs on a filtered pool rather than the raw one. Deep research on the wrong fifty people produces a thorough document about the wrong fifty people.
How it fails: inference gets recorded in the same voice as evidence. “Was in the role while the acquisition closed” becomes “led the integration.” Both sentences read the same on the page, and only one survives a client checking it. It is one of the failures our own QA pass exists to catch. Reconstructing scope properly is the whole craft of this stage.
5. The lateral pass
What it is for: returning to the people who failed exactly one criterion and asking a second question — would this person do the job well anyway?
The plant director whose title lagged the job. The operator whose sector label differs but whose daily constraints do not. The candidate from a smaller company who was running more than the org chart suggested.
This stage exists because a search that only surfaces people who cleared every stated box has produced something a database query could have produced. Laterals reach the client with the miss named and the reasoning attached, so the trade-off is theirs to weigh.
How it fails: it gets skipped. It is the easiest stage to drop under deadline because nothing visibly breaks when you do.
A client can see the shortlist and the interviews. Everything between them is invisible from outside, which is exactly why it is where the corners get cut.
6. Quality assurance
What it is for: finding what the previous stages got wrong, before the client sees it.
The requirement that makes this work is separation. A review conducted by the person who did the research, in the same sitting, inherits the assumptions that produced the errors. A genuine QA pass starts cold and goes looking for specific failures: claims cited to sources that do not support them, merged identities, inference stated as fact, sources too vague to locate.
Merged identities deserve particular attention. Two executives with the same name in the same sector produce one record with a remarkable career, and it passes every check that is not specifically looking for it. It is the failure we built our own QA stage hardest against, and the one worth asking any firm how they guard against.
How it fails: it becomes a proofread. Formatting gets checked; attribution does not.
7. Shortlist delivery
What it is for: handing over a document that answers the client’s questions before they ask them.
Two sections, not one ranked list: on-spec matches with per-criterion evidence, and strong laterals with the trade-off stated. Each entry carries confirmed role history with sources, the criteria mapping, the context signals relevant to this role, and contact data.
Length follows the evidence rather than an expectation. If five candidates confirm, the shortlist has five. Padding to a round number means some entries exist to fill space, and the client will identify which ones. What each entry owes the reader is worth being deliberate about.
How it fails: the document ranks. A shortlist that orders candidates by predicted fit has made a judgment call using none of the context that judgment requires.
8. Interviews, references, close
What it is for: the client’s decision, supported by everything upstream.
This is where fit gets assessed, and it is properly the recruiter’s and client’s territory. The research stages exist to make sure this conversation happens about the right people, on accurate information.
How it fails: a claim that was never confirmed surfaces in a reference call. The cost is rarely one candidate — it is the credibility of every other entry in the document.
Why the sequence matters
Each stage assumes the one before it did its job. Vague criteria produce an undisciplined filter. A narrow pool produces a thin shortlist that looks like a market problem. Skipped QA produces a mis-attributed claim that surfaces in front of the client and retroactively puts every other claim in question.
Which is also why the research stages — two through six — are where the process gets compressed when time is short. They are the least visible to the client and the most expensive in hours, so they absorb the pressure. What that compression actually costs a firm is the part that rarely gets measured.
Related reading: how to tell search firms apart, what a retained executive search firm actually does, and how Cerna handles the research stages.
Frequently asked questions
How many stages does a typical executive search process have?
Eight: intake and brief, market mapping and discovery, first-pass filter, deep research, the lateral pass, quality assurance, shortlist delivery, and interviews through close. Firms label and order these differently, but each job still has to get done by someone.
What is the difference between market mapping and the first-pass filter?
Market mapping assembles everyone who could plausibly be considered, before anyone is evaluated. The first-pass filter then removes people on quick, checkable signals like function, seniority, and geography. Judgment about quality belongs to neither step.
Why does the executive search process include a dedicated lateral pass?
It returns to candidates who missed exactly one requirement and asks whether they would do the job well anyway. It is easy to skip under deadline because nothing visibly breaks when it is dropped, which is exactly why it gets cut first.
What does quality assurance actually check in an executive search process?
A separate, cold review that looks for claims cited to sources that do not support them, two people's histories merged into one record, and inference written in the voice of confirmation — not a proofread, which only checks formatting.
Which stage of the executive search process gets compressed first under deadline pressure?
The research stages, from market mapping through quality assurance. They are the least visible to the client and the most expensive in hours, so they absorb the pressure first while nothing about the compression looks like a visible failure.