Hiring a customer support agent
Three hundred applications for one seat, most of them two lines long, and the qualities that actually predict success — patience, clarity, composure with someone who is furious — appear nowhere on a CV. This is a volume problem and a signal problem at the same time.
The CV is the least informative thing in the application. Support CVs are short by nature and near-identical: a list of employers, a line about being a people person. A previous job title tells you almost nothing, because "customer service" spans a supermarket till and a technical escalation queue. Meanwhile a large slice of the pile fails on shift availability or language before anyone reads a word — facts no CV surfaces, so hours go into people who were never eligible. What predicts the hire is how someone writes and how they handle pressure, and only one of those is even visible on paper.
One sentence about the queue, and the guide changes shape
Selene asks what a high-volume support role actually turns on: channel, hours, languages, and how much of the job is judgment versus script.
Channel, hours, languages, and how much judgment
The questions here are nothing like an engineering role's. Selene goes after coverage and channel mix, because those decide both the eligibility gates and whether written communication or phone manner carries the weight.
One sentence changes the guide
Every support manager knows this and almost none of them put it in the ad, which is why support ads all ask for "2 years' experience in a customer-facing role" instead. That requirement filters on tenure; this answer filters on temperament — and they rank the pile in very different orders.
Exceptional, not just good —
“The great ones make the angry ones calm down. You can't teach that.”
→ De-escalation evidence became a must-have; product familiarity dropped to a nice-to-have, because it can be taught.
What the guide ends up measuring
Seven criteria. Three are yes-or-no facts that belong on the application page rather than in the scoring, and the rest lean on how someone writes — which on this role is the closest thing to a work sample you get for free.
Resolve the rota before you read anything
On a role this size the eligibility questions do more work than the scoring does. Availability and right to work are facts, not judgments — asking them on the application page means the answer arrives with the application instead of surfacing on a phone call nine days later. On a 320-applicant pile that regularly clears a third of it before any AI runs, and none of that is billed as screening.
What she found, and what she couldn't
Selene never receives the candidate’s name or the raw CV. She scores a blind, structured profile, with any stated age, sex, nationality and religion stripped out before scoring runs.
On a role where written English is itself a criterion, the separation matters more than usual — the guide asks how clearly someone writes, not who it sounds like they are.
How the blind scoring works →What stays with you
Selene's half
- Turning the brief into requirements, and the eligibility ones into application-page questions
- Reading all 320 against the same guide, blind, with the evidence attached
- A ranked shortlist, and a decline reason for everyone not on it
- Drafting the declines, so 300 people get an answer rather than silence
Yours
- Advertising the role — Selene doesn't post to job boards or sourcing sites
- Any trial shift, written test or role-play; she reads applications, she doesn't assess live
- The interview and the tone-of-voice judgment, with her brief in hand
- Every hiring decision, and any override — stored with your reason
Nearby roles
Put your queue in front of her
Email, chat, phone, or all three at once — describe the queue and the guide comes back weighted for it.