# Example: Generate Targeted Follow-up Questions

Fictional company, role, candidate and people.

## What the user pasted

> Data Analyst, Kestrel Mobility. Criteria: 1) SQL and data modelling (must-have), 2) Explaining findings to non-technical people (must-have), 3) Prioritising requests, 4) Working with stakeholders. Want a 30 min follow-up call with D.O. next week, Lena (hiring manager) will run it. Also might do references.
>
> Lena's notes: talked well about the pricing dashboard, "we rebuilt the whole data model" but unclear what D.O. did herself. SQL questions went fine but only basic joins. Explained churn findings to me really clearly, used a simple chart. 4/5.
> Jon (ops lead) notes: good communicator. Asked how she handles 10 requests at once, said "I just work late". Didn't get to stakeholders, ran out of time. Explaining: 3/5, "a bit too detailed for ops people".

## What the skill produced

# Follow-up questions: D.O. for Data Analyst

**Next step:** Follow-up call, 30 min · **Asked by:** Lena · **Based on:** Lena's and Jon's interview notes

## Where the evidence stands

| Criterion | Status | Gap to close |
|---|---|---|
| SQL and data modelling | Open | Own part in the data model unclear; only basic joins seen |
| Explaining findings | Open | Lena rated 4, Jon 3; different audiences |
| Prioritising requests | Open | One answer ("I just work late"), no example |
| Working with stakeholders | Not assessed | Ran out of time |

## Timing and consistency

- **Fits the time:** 30 minutes holds about 2-3 questions at 8-10 min each, with a few minutes to open and close. Core: 1-2 · If time allows: 3 · Moved to another step: 4 (reference check, or extend the call to 45 min [TBC])
- **Other finalists:** If other candidates are still in play, ask each of them questions 1-3 too where possible, so answers can be compared fairly. [TBC: other finalists?]

## Questions, in priority order

### 1. SQL and data modelling

- **Why it matters:** Must-have. Notes say "we rebuilt the whole data model", but D.O.'s own part is unclear.
- **Question:** "In the pricing data model rebuild, which parts did you design or build yourself?"
- **Probes:** "What trade-off did you make, and why?", "What would you change now?"
- **A strong answer would show:** Specific tables or logic she built, a design choice she can explain, and its effect.
- **Asked by / when:** Lena, follow-up call

### 2. Explaining findings to non-technical people

- **Why it matters:** Must-have. Interviewers differ: clear for Lena, "a bit too detailed for ops people" for Jon.
- **Question:** "Tell me about a time you had to explain a finding to people on the ground, like drivers or ops staff. How did you adapt it?"
- **Probes:** "How did you know it landed?"
- **A strong answer would show:** Changing the message for the audience, cutting detail, and a sign it was understood or used.
- **Asked by / when:** Lena, follow-up call

### 3. Prioritising requests

- **Why it matters:** Only one general answer so far ("I just work late").
- **Question:** "Think of a week when you had more requests than time. How did you decide what to do first?"
- **Probes:** "Who did you tell, and what got pushed back?"
- **A strong answer would show:** Clear criteria for priority (impact, deadline), agreeing trade-offs with requesters, not just extra hours.
- **Asked by / when:** Lena, follow-up call, if time allows

### 4. Working with stakeholders

- **Why it matters:** Not assessed yet.
- **Question:** "Tell me about a stakeholder who disagreed with your analysis. What did you do?"
- **Probes:** "What was the outcome?"
- **A strong answer would show:** Listening to the objection, checking the data again, and reaching an agreed next step.
- **Asked by / when:** Reference check (see below), or Lena if the call is extended to 45 min [TBC]

## Better tested another way

- SQL and data modelling: a call will not show SQL depth beyond basic joins. Consider a short take-home or live SQL exercise. [TBC]

## Reference-check version

*Use only with the candidate's agreement. Ask about what the referee saw at work, at first hand.*

**Start with:** "How did you work with D.O., and for how long?"

| Criterion | Question for the referee | Listen for |
|---|---|---|
| SQL and data modelling | "What data work did D.O. own herself? Can you give an example?" | Her own contribution, named projects |
| Explaining findings | "Can you describe a time she presented results to a non-technical group?" | How she adapted, what happened next |
| Prioritising requests | "When she had too many requests, what did she do?" | Priorities agreed with others, not just extra hours |
| Working with stakeholders | "Tell me about a time she disagreed with a stakeholder." | How she handled it, the outcome |

## Open questions

- Is a SQL exercise possible before a decision? [TBC]
- Who would take the reference calls? [TBC]
- Extend the call to 45 min to cover stakeholders directly? [TBC]
