How to hire great people
Great people are rare, and most hiring is built to find adequate ones.
Great is not the same as not bad
Avoiding a bad hire and finding a great one are different jobs. In most work, performance is lopsided: a small group of people produce far more than the rest. Aguinis, O'Boyle, Gonzalez-Mulé and Joo looked at about 625,000 people across 229 datasets in 2016. Stars appeared where the job had freedom, complexity and no ceiling on output.
Where output is capped, like a checkout or an assembly line, people sit closer together. Most professional jobs are not capped. That is where finding the right person is worth the most.
Step 01start from what the person must deliver
Write down what the best person in the role gets done in their first 12 months, and how. Not a list of skills, but outcomes: the result, the obstacle and what success looks like to the manager. Then name the hard part of the job, the thing that makes good people struggle.
This is what Speak to Kit asks a hiring manager first. With an ATS linked, the manager gets a top performer profile already written for the role and refines it.
Step 02write one standard, and use it for everyone
Kuncel and colleagues compared fixed scoring rules with experts weighing the same information in their heads. Across 25 samples, the rule did more than 50% better at picking good performers, .44 against .28.
A free-form chat can make things worse. In a lab study by Dana, Dawes and Peterson, interviewers given random answers felt just as sure of their judgement. So write each question with a good answer and a great one, and score every person against the same words.
Step 03speak to everyone, not only the best CVs
A CV is a weak guide. Years of experience barely relate to performance and university grades only a little. The person who would be best at the job may have a career break, an odd layout or an unusual path.
Speak to Kit speaks to every applicant, for as long as they need, so nobody is left out because of how their CV looks. At Rhaelm, one woman who had spent five years at home with her children was ranked 100th by her CV. Kit put her first, and she got the job.
Find the best from every source, then assess them all the same way
The best person for a job may not be in the pile of applicants. Some are applicants. Some are off market: in work, doing well, and not looking. Some come through a referral from someone who has worked with them. Some are found by a sourcing tool. Hiring well means looking in all four places, then giving everyone the same test.
Applicants come through job boards and the company's own careers site. Speak to Kit speaks to every one of them as part of applying, by phone or in writing, for 15 to 20 minutes, whenever suits them. 276 in every 1,000 Kit calls start between 7pm and 7am, so people in work can take part after hours.
Off-market people and sourced candidates are found by the team. Juicebox describes itself as "the AI Recruiting Platform that learns how you search, so you source better talent, faster". LinkedIn says Recruiter lets a team "Reach candidates directly on LinkedIn with Recruiter's built-in messaging". Speak to Kit works alongside both. It does not connect to the sourcing tool or to LinkedIn. The team sends the people it finds the job's link, they apply, and they speak to Kit like everyone else.
Referrals are worth the effort. People referred by staff quit less, in a study of nine large firms by Burks, Cowgill, Hoffman and Housman (2015). A referral speaks to Kit as part of applying, against the same standard as everyone else.
Every person, whatever the source, is asked about the same things: what the hiring manager says a top performer delivers in their first 12 months, and the hard part of the job. Each is scored against the same words, so the best person wins rather than whichever channel shouted loudest. At one large tech employer, applicants a computer picked from CVs after human screeners had turned them down were 14% more likely to get an offer and more productive once hired, and the gains came mostly from people without a referral, an elite school or a typical CV (Cowgill, 2020, working paper).
Then the sources can be compared. On the Talent and Enterprise plans, where Kit checks in on every hire at 30 days, 90 days and a year, a team can see which source brought the people still delivering a year later: job boards, referrals, sourcing tools or agencies.
“We analyze every candidate that comes in from every source, be that LinkedIn, job adverts, company website, Juicebox, or any hiring AI agent.”
Step 04check, and learn
At 30 days, 90 days and a year, ask how each hire is doing: are they doing well, did they stay, did they grow. Feed the answers back into the next profile for that team. Most employers skip this step, so their hiring never learns.
- 01Define the outcome over 12 months, with the hard part named.
- 02Write one standard with a good and a great answer.
- 03Speak to every applicant against it.
- 04Let people make the decision, with the evidence in front of them.
- 05Check every hire at 30 days, 90 days and a year.
What great hires have in common
When hires fail, managers mostly blame attitude, not skill. A Leadership IQ study in 2005 asked managers why their hires had not worked out: 89% said attitude, such as taking feedback, motivation and temperament, and 11% said skills. It is managers' own view, not peer-reviewed, but it matches what recruiters hear.
So a good conversation asks about the way someone works. How they took hard feedback. What kept them going when a project stalled. Which kind of team brought out their best. Those answers do not appear on a CV.
A fair warning from the research
Not every researcher agrees on how lopsided performance is. Beck, Beatty and Sackett found in 2014 that, measured within one job by ratings or capped output, performance looks closer to a bell curve. Both sides agree it is lopsided in work that is free, uncapped and complex.
For a hiring manager the practical lesson is the same. The more freedom and complexity a role has, the more it is worth finding the outstanding person, and the more careful the search should be.
What it is worth
A great hire is worth more than it costs to find. A failed one costs about £51,000 on an £80,000 salary once the replacement is up to speed, and in UK small firms 16 in 100 new hires leave within their first year. "Quality of hire is worth more than cost of hire, far more than cost of hire." For example, a phone leasing company hiring about 43 people a year has about 7 first-year leavers, costing about £251,000 a year. Halving that would be worth about £126,000 a year (estimate).
How one failed hire costs about £51,000
“Every applicant is asked about the same things in a way that's contextual to their experience. It is a dynamic conversation based upon a static scorecard.”