Introduction
Two candidates apply for a data analyst role. Both list SQL, Python and dashboard experience. Both describe projects that improved business performance. One CV is more polished, but neither tells you how the candidate would handle an unexplained sales decline tomorrow.
A CV helps establish relevant experience. Choosing someone to solve business problems requires further evidence: how they handle data, question assumptions and explain what a business should do next.
CV screening, practical skills challenges and structured interviews each contribute something different. The useful question is how to combine them around the work this analyst will actually perform.
Start With the Work the Analyst Will Actually Do
Before reviewing applications, define the decisions the analyst will support.
For a commercial analyst, that might mean checking sales data, investigating revenue changes and explaining findings to regional managers. A role focused on reporting automation will need a different assessment.
Translate the job description into observable requirements. Strong SQL skills might mean combining order and customer tables without duplicating revenue, handling missing values and producing a result another analyst can reproduce.
Separate skills needed on arrival from those the team can teach. Familiarity with your particular dashboard software may be trainable; independently validating a revenue calculation may be essential.
Agree these priorities before seeing candidates. Otherwise, an impressive background can quietly change what the team considers important.
Use the CV to Identify Relevant Experience and Questions
A CV provides context about responsibilities, industries, tools and claimed achievements. Use it to identify evidence worth exploring.
Consider the statement: “Built a sales dashboard that improved decision-making.”
It leaves several questions unanswered:
- Which parts did the candidate personally deliver?
- How did they check the underlying data?
- What decision changed because of the analysis?
- How was the claimed improvement measured?
A candidate who maintained an existing dashboard may have valuable operational experience. Someone who designed its calculations may demonstrate a different capability. The interview should establish that distinction.
Avoid treating employer names or years of experience as substitutes for relevant evidence. Equally, a less polished CV should not automatically outweigh a clearly described project that matches the role.
The output of screening should be a shortlist with specific questions to investigate.
Use a Practical Challenge to Observe the Work
A practical assessment gives candidates a defined problem and asks them to produce something inspectable.
For this fictional hiring example, both candidates receive the same synthetic sales dataset and brief:
Sales have declined in one region. What changed, what would you validate, and what should the business investigate next?
Ask for their calculations or query, a short explanation of data checks and a recommendation for the sales manager.
Use a manageable time limit and explain permitted tools, including whether AI assistance is allowed. Provide a way to request accessibility adjustments. Keep the task focused on assessment rather than unpaid production work.
If a candidate brings results from data analytics challenges, inspect the task format and available work. A multiple-choice result and a reproducible analysis reveal different things. Check what was assessed and under what conditions.
What the Same Dataset Can Reveal
In our fictional dataset, regional revenue falls from £500,000 to £420,000, a 16% decline. Completed orders fall from 5,000 to 4,000, while average revenue per order rises from £100 to £105. Both periods use the same revenue definition.
Candidate A calculates the revenue decline correctly and produces a clear dashboard. They recommend more discounting to stimulate demand, but provide limited checks and do not investigate order volume.
Candidate B checks duplicate records, missing region labels and period completeness. They reconcile revenue to completed orders and show that fewer orders account for the decline despite higher revenue per order.
Candidate B has explained the arithmetic. They have not established why order volume fell. Their next recommendation is to compare the change by product and customer group, then investigate stock availability and acquisition activity.
That distinction matters: a calculation can identify the component that changed without proving its underlying cause.
Candidate A shows stronger visual presentation. Candidate B shows stronger validation and diagnostic reasoning. Those differences give the interview a clear purpose.
Agree the Scoring Criteria in Advance
Assess the work against four areas:
- Accuracy: Are the calculations and metric definitions correct?
- Validation: Were relevant data problems checked and their effects explained?
- Reasoning: Do conclusions follow from the evidence, with uncertainty acknowledged?
- Communication: Can a manager understand the finding and proposed next step?
Define what acceptable performance looks like before reviewing submissions. For validation, “mentions checking data” is weaker evidence than demonstrating a duplicate check and explaining its result.
Have reviewers record evidence against each criterion before discussing an overall impression.
Use a Structured Interview to Explore the Reasoning
The interview should clarify how the candidate reached their conclusions and whether they can adapt them.
Ask every candidate the same core questions, in the same order, and assess answers against the same criteria. Use focused follow-ups to clarify their particular submission.
For the sales example, useful questions include:
- Which finding are you most confident in, and why?
- What could make your revenue comparison misleading?
- What evidence would change your recommendation?
- How would you explain the result to a sales manager who disagrees?
Candidate A might recognise that their discount recommendation was premature and explain how they would investigate order volume. Candidate B might struggle to translate a technically correct finding into a practical conversation with a manager.
Record those observations separately from the original submission. A helpful explanation adds evidence; it does not retroactively demonstrate checks missing from the work.
Also revisit a CV project. Ask about personal contribution, a difficult decision and something that went wrong. This connects claimed experience with the reasoning demonstrated in the exercise.
Bring the Evidence Together Before Making a Decision
Each method should answer a defined question.
| Method | What it contributes | What still needs checking |
| CV screening | Relevant background, responsibilities and project claims | Personal contribution, depth and current ability |
| Practical challenge | Observable calculations, checks and recommendations | Ownership, reasoning and performance beyond one task |
| Structured interview | Explanation, judgement and responses to follow-up questions | Whether these are supported by demonstrated work |
Skill verification is useful when it connects a specific capability to inspectable evidence. A label or total score alone leaves important questions unanswered.
For the fictional commercial analyst role, suppose reliable diagnosis was agreed as essential and presentation polish as developable. Candidate B currently provides stronger evidence against that requirement, provided their communication meets the role’s minimum standard.
That is a role-specific judgement, not proof that Candidate B is universally the better analyst.
When evidence conflicts, identify the precise gap. A short walkthrough of an unfamiliar query may resolve an ownership question. Another broad interview may simply repeat what the team already knows.
Keep unresolved uncertainty visible rather than averaging it away in a total score.
Assessing Data Skills With AuthenX and CompeteX
PangaeaX offers two products relevant to different parts of this process.
CompeteX is a data competition platform covering areas such as SQL, Python, business intelligence and analytics. Relevant challenge work can give hiring teams material to examine, with its usefulness depending on the task, format and available evidence.
AuthenX is PangaeaX’s skill authentication platform. It uses CV or portfolio screening and AI-led interviews to assess professional skills. Its process provides a different perspective from completing a practical work sample.
Employers should review what each assessment actually covers and relate it to their own requirements. Neither a competition result nor an authentication outcome establishes every aspect of suitability for a job.
For your next analyst hire, start with one business problem the person will need to solve. Use the CV to understand their background, practical work to examine their approach and structured questions to explore their judgement. Then make the decision against the requirements you agreed at the beginning.