Fit is not likeability, and screening for it by instinct reproduces the founder. Define the role precisely enough that a candidate can be certified against it on knowledge and on skills, screen for the gap in the firm rather than the resemblance to the people already in it, and accept that the ability to recruit formally, at scale, is what separates a firm that can grow from one that cannot.
Founders ask Collective 54 this 21 times in our records. Most are asking how to interview better. The firms that solve it change what they are screening for, and change how they define the job before anyone is interviewed.
Recruiting as a start-up is not a mission-critical task. The needs are basic and most jobs can be filled from personal networks. That is why founders arrive at this question late: the method that worked for the first fifteen hires quietly stops working, and it does not announce itself.
At scale, recruiting is mission critical. The job requirements are far more demanding and the roles often cannot be filled from personal networks at all. Specialists, managers and executives have to be formally recruited. The ability to recruit at scale is one of the clearest things separating market leaders from boutiques, and it is a capability rather than a knack.
Screening problems are usually definition problems wearing a disguise. If the role is not defined, no interview process can assess fit against it, and the interviewer falls back on whether they liked the person.
Sophisticated clients force the definition whether you want it or not. As a firm scales it attracts clients who pay more and expect more, and who are experienced purchasers of professional services. Proposals stop being able to say engagement manager with five years of experience. They have to say engagement manager, ten years in this industry, this segment, this project type. That specificity is the job description. If a client would not accept the vague version in a proposal, you should not accept it in a job posting.
Here is the trap, and it is almost universal in founder-led firms.
The culture of a firm usually originates from the founder, who designed a firm they wanted to work in. Early recruits are the people who get along well with the founder. Those employees perpetuate the culture to the next set of recruits, and on it goes, until the firm has hardened around its culture. Employees who do not gel with it get ejected, the way a transplant is rejected by its host.
That process is not a problem in itself. There is no right or wrong culture, and the type of culture is not a predictive indicator of success. Successful firms exist with many different cultures. The problem is that a founder screening on instinct is screening for resemblance, and resemblance compounds. Fifteen hires later the firm is very good at one thing and structurally incapable of the others.
The founding-team principle is the corrective, and it applies to every hire after the first three. A firm is best started by a team of three: one person great at bringing in clients, one great at servicing them, one great at developing service offerings, with very little overlap in skills. The idea is that one plus one plus one equals ten. When resources are constrained you cannot afford redundant skills.
Apply that to screening. The question is not whether this candidate is like us. It is which of the gaps in the firm this candidate closes, and whether the person closing it will be tolerated by a culture that has hardened around the opposite profile. If the honest answer to the second half is no, that is a leadership problem to solve before the offer goes out, not after.
The same discipline is why the most common founding-team failure keeps repeating: a group of people who all enjoy the technical work and none of whom enjoy marketing and selling it. Once the initial referral stream dries up there is not enough work to survive. That team was not unlucky. It was badly constructed, and it was constructed by people hiring in their own image.
Interviews test whether someone can describe the work. Certification tests whether they can do it.
Employee certification is a defined process proving an employee has reached a level of competency before work is handed to them. There are two things to certify in professional services and both are needed. Knowledge, meaning what the person understands. Skills, meaning what the person can reliably execute.
Build it as the screen and two things happen at once. Hiring stops being an act of faith, which is what founders are actually resisting when they refuse to hand work over. And the same standard that screens candidates becomes the standard that develops the people already there, because the certification is a description of competence rather than a description of a person.
Firms with high turnover usually believe they have a screening problem. Often they have a design problem, and the exit interviews say so.
In one firm with forty per cent annual turnover, former employees consistently cited role corruption, meaning that the job was not clearly defined. Star employees ended up doing the work of their coworkers and burned out. Performance reviews were a compliance exercise rather than real feedback. The benefits package pushed too much cost onto employees. Almost everyone who left landed a job with better pay. The owner was underpaying and the employees knew it.
None of that is fixed by better interviewing. A firm competes in two markets, the market for clients and the market for employees, and as much effort has to go into the second as into the first. The question worth asking honestly before blaming the candidate pool is whether you would want to work for you.
Screening has always been an attempt to predict output. In Era 3 of the Collective 54 Era Framework, a growing share of the repeatable output is produced by software under human supervision, and what remains for people is judgment.
That moves the screen. The scarce capability is no longer throughput, it is the ability to define work precisely enough for it to be produced reliably, and to supervise the result. Those are the traits worth testing for, and they are testable: give a candidate an ambiguous brief and see whether they specify it before executing it.
The definitional discipline is the same one this answer opens with. A firm that cannot define a role well enough to certify a human against it cannot define the work well enough to hand it to a system either. Read the Era Framework in full, and the AI-Native Firm Map for how the nineteen roles inside a firm are organised.
If you are losing people faster than you can hire them, stop working on screening. Fifteen per cent turnover or lower is the benchmark, and at thirty per cent the entire employee base turns over every three years. At that rate the constraint is retention, pay and role clarity, and better candidate selection cannot outrun it.
And if you genuinely are hiring your first handful of people, personal networks are the right tool. Formal recruiting is a scale capability. Building it too early is overhead you cannot afford.
Define the role to the standard a sophisticated client would accept in a proposal. Certify candidates against it on knowledge and on skills instead of interviewing for likeability. Screen for the gap in the firm rather than the resemblance to its founder, because resemblance compounds. And check the exit interviews first: if good hires are failing, the cause is usually role corruption and pay, not selection.
Fit against a defined role, not fit with the people already there. If the role is not specified to the standard a sophisticated client would accept in a proposal, no interview can assess fit against it, and the interviewer defaults to whether they liked the candidate. Definition comes before screening.
Because instinct screens for resemblance. The culture of a firm usually originates from its founder, early recruits are the people who get on well with the founder, and those employees perpetuate the same profile to the next set of hires. Fifteen hires later the firm is very good at one thing and structurally incapable of the others.
Employee certification. Define a process that proves a person has reached a level of competency before work is handed to them, covering both knowledge, meaning what they understand, and skills, meaning what they can reliably execute. Interviews test whether someone can describe the work. Certification tests whether they can do it.
Usually not. The common cause is role corruption, meaning the job was never clearly defined, so star employees absorb the work of others and burn out. Underpaying relative to the market and treating performance reviews as a compliance exercise do the rest. Read the exit interviews before changing the interview process.
Sources: Greg Alexander, The Boutique: How to Start, Scale, and Sell a Professional Services Firm (Advantage, 2020), chapter 22 on recruiting at scale and the shift from generalists to specialists, chapter 9 on founding-team composition and the cost of overlapping skills, chapter 16 on employee certification of knowledge and skills, chapter 37 on how culture originates from the founder and hardens over time, and chapter 35 on role corruption, turnover benchmarks and competing in two markets. Greg Alexander, The AI-Native Boutique Firm (Advantage Books, January 2027), for the Era Framework and the shift from producing output to supervising judgment.
Collective 54 is the private community for founders and executives of boutique professional services firms between $5M and $50M in revenue. Members work these answers against their own numbers.