Hiring a data analyst
A hundred and seventy applications listing the same six tools in the same order. The pile contains two genuinely different populations, and nothing in the CV format distinguishes them.
SQL, Python, Tableau, Power BI. Everyone. Every time. The tooling section of a data CV has become pure noise — it is the same list on almost every application, and it says nothing about whether someone can be trusted with a question. The two populations hiding behind it are the analyst who receives a request and returns a number, and the analyst who works out that the request is the wrong one; the second is worth several of the first and writes an almost identical CV, because the achievements section rewards volume of output. The classic false positive is a long list of dashboards built. Nobody ever writes down how many of them are still opened.
Who brings the questions
Selene goes after where the questions come from, because an analyst who receives them and an analyst who forms them are different hires with the same job title.
Requests, stakeholders, and how raw the data is
For an analytics role the shape of the work is set by two things: whether questions arrive fully formed or need shaping, and how much of the job is cleaning before any analysis starts. Managers rarely volunteer either, and both change the criteria substantially.
One sentence changes the guide
This is the cheapest possible test of the distinction that matters and almost nobody screens for it, because it does not appear anywhere on a standard CV. It has to be asked for — which is exactly what the discovery conversation is for.
Exceptional, not just good —
“They ask what decision it’s for before they start querying.”
→ Reframing a request became a must-have; dashboard volume stopped being counted at all.
What the guide ends up measuring
Six criteria, one of which names a technology. The rest are about what someone does before and after the query — which is where the two populations in this pile actually separate.
No gates, and the number that gets misused instead
Nothing here is a yes-or-no fact. The proxy teams reach for is years of experience, and on this role it is particularly poor: the difference between the two populations in the pile is a habit of mind that some people arrive with in year one and others never acquire in ten.
Scored on the question, not the query
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.
She reads for the moments where a candidate describes deciding what to measure, which is the part of this job that never fits in a bullet point.
How the blind scoring works →What stays with you
Selene's half
- Establishing whether questions arrive formed or need shaping, and weighting accordingly
- Reading all 170 against the same guide, blind, with the evidence attached
- A ranked shortlist that separates the two populations hiding behind the same tool list
- An interview brief with the reframing question aimed at each candidate’s own examples
Yours
- Getting analysts to apply — Selene doesn’t post to job boards or source candidates
- Any SQL test or case exercise; she reads what people wrote, not what they can do live
- The interview, and the judgment about how they think out loud
- The offer, and every call that matters
Say where the questions come from
That answer alone separates a reporting hire from a thinking one — describe yours and watch the guide follow.