Hiring practice3 min read
Screening hundreds of applications without reading every CV twice
When a role draws four hundred applicants, the reading is not the hard part. Staying consistent from the first CV to the four-hundredth is.
Hiring practice3 min read
When a role draws four hundred applicants, the reading is not the hard part. Staying consistent from the first CV to the four-hundredth is.
Put this into practice on your next role.
Free for your first role. No card.
Treat any automated score as a ranking with reasons, then check a sample of what it put at the bottom.
Answer everyone. At volume, silence is the default, and it is what people remember.
A well-written role for a graduate programme or a customer support team in Lagos can easily draw several hundred applications in a week. Nobody reads four hundred CVs with the same attention. By the two-hundredth, the reader is tired, the bar has moved, and the order in which applications arrived has become part of the decision.
Volume does not need a different philosophy from any other hire. It needs the same one, applied without drift.
The single most useful thing you can do is write down, before the role opens, what moves an application forward. Once you have seen the applicants, the rules start bending to fit the ones you liked.
| Kind of rule | Example | How to apply it |
|---|---|---|
| Hard requirement | Right to work in Nigeria | A yes/no screening question. Automatic. |
| Must-have skill | Has handled customer tickets in writing | A short screening question, then a test. |
| Preference | Speaks Yoruba or Hausa | Noted, never used to reject. |
?Quick check
You are hiring delivery riders. Which of these belongs in a knockout question?
Keep knockout questions to real, legal, binary requirements: a licence, a location, the right to work, availability on the start date. Anything that needs judgement is not a knockout question.
Automated screening is useful at volume for one reason: it reads the four-hundredth application with the same attention as the first. That is its whole value, and it is only worth having if you can see its reasoning.
A score on its own is a verdict you cannot argue with. A score with reasons (“meets three of four must-haves; no evidence of written ticket handling”) is advice you can check. Insist on the second kind.
Whatever ranks your applications, a person or a model, it will make mistakes, and the costly ones sit at the bottom where nobody looks. Build in a small audit.
Take a random sample
Ten or twenty applications from the bottom third, chosen at random, not by skimming.
Read them properly
Would any of them have passed if a careful person had read them first?
Fix the rule, not the person
If a good candidate was ranked low, find out why. Usually a must-have was worded so that only one phrasing of it counted.
This takes twenty minutes and it is the difference between using a tool and trusting one.
At volume, the screen should be generous and the test should be decisive. It is far cheaper to send a forty-minute skills test to eighty people than to interview forty, and the test tells you more. Everyone gets the same questions, marked the same way.
Screen generously, test everyone the same way, interview few.
The candidate experience at volume is mostly silence. Hundreds of people applied, a handful heard back, and the rest are left refreshing their inbox. Every one of them will remember your company’s name.
A short, specific message closes the loop: we have filled the role, or you did not meet the requirement for X. It costs you minutes to set up once. For the person reading it, it is the difference between a rejection and being ignored.
Hiring on evidence: 0 of 5 read
Up next · Part 4
Skills tests that candidates don’t resent
A good test is short, clearly about the job, and honest about what it measures. A bad one is unpaid work with a timer on it. Candidates can tell the difference immediately.
Read next · 3 minWas this useful?
One tap. It decides what we write next.
Get the next one by email
One short email when something new is worth reading.
A good test is short, clearly about the job, and honest about what it measures. A bad one is unpaid work with a timer on it. Candidates can tell the difference immediately.