01Owns
Metric definitions
Writes down exactly what counts as an active user, a churned account or a qualified lead, including edge cases, and keeps one agreed version in the semantic layer.
Hire · Data and analytics
Hire Data Analysts
Senior data analysts who sit inside your team and turn questions from product, growth and finance into trustworthy numbers. Shortlisted analysts have written SQL against a messy sample warehouse, defined a metric from a vague brief and presented an A/B readout to a non-technical reviewer.
SQL exercise on a deliberately messy dataset
Senior and lead analysts only
Works in your warehouse and BI tool
On your team
What a data analyst owns on your team.
02Owns
Recurring reporting
Maintains the weekly and monthly views leadership reads, retires charts nobody opens and annotates any figure that moved for a known reason.
03Owns
Ad hoc questions
Takes a loose request such as “why did sign-ups dip in March?”, narrows it into something answerable and replies with the query, the caveats and a plain-language conclusion.
04Owns
Experiment readouts
Checks sample ratio, guardrail metrics and pre-registered success criteria before calling a test, and writes a readout the product owner can act on.
05Owns
Data quality questions
Notices when event counts, joins or late-arriving rows look wrong, traces the cause with engineering and flags affected reports instead of quietly patching them.
The analyst you need depends on your questions
Before writing a job post, list the last ten questions leadership asked that went unanswered or got two different answers. If most are about funnels, retention and campaign results, you need an analyst with strong SQL and business sense. If they are about predicting who will leave or what to stock next month, that is a modelling role and belongs on a different brief.
Then ask candidates about a metric they had to redefine. Good analysts can describe the argument, who disagreed, and how the new definition was rolled out without breaking historic comparisons.
Vetting
Skills we test before a data analyst joins.
01Tested
Analytical SQL
Window functions, CTEs, correct handling of NULLs and duplicates, and choosing the grain of a table before joining anything to it.
02Tested
Semantic modelling
Builds or extends dbt models and metric layers so that revenue, retention and conversion are computed once rather than in every dashboard.
03Tested
Dashboard design
Picks the chart that answers the question, labels units and date ranges, and designs views in Looker, Power BI, Tableau or Metabase that survive a change of audience.
04Tested
Experiment interpretation
Understands confidence intervals, peeking, novelty effects and why a flat result is still a result worth reporting.
05Tested
Spreadsheet fluency
Still reaches for a spreadsheet when finance needs to audit a number, and keeps the logic traceable back to source queries.
06Tested
Communication under ambiguity
Asks what decision a number will inform before writing the query, and says clearly when the data cannot support the question.
Work sample
How the data analyst work sample runs.
- 01Step 1
Untangle a broken query
The analyst receives a revenue query that double-counts because of a one-to-many join, explains the fan-out and corrects the grain.
- 02Step 2
Define a metric from a brief
From a two-line request for “weekly engaged users”, they write a definition, list the edge cases they chose to exclude and build the query.
- 03Step 3
Read out an experiment
Given raw assignment and conversion tables, they check for sample ratio mismatch, compute the effect with an interval and recommend a decision. We pay for the hours.
- 04Step 4
Present to a stakeholder
They talk a marketing or finance lead through one chart in ten minutes, handling a sceptical follow-up question without retreating into jargon.
Levels
Data Analysts: what changes at each level.
01Level 1
Senior
Answers questions independently across one or two business areas, maintains their own models and dashboards, and reviews colleagues’ SQL.
02Level 2
Lead
Owns the metric catalogue, sets review standards for queries and dashboards, and decides which requests the team takes on.
03Level 3
Analytics lead
Shapes how the company measures itself: north-star metrics, experimentation policy and the split of work between analysts, analytics engineers and data scientists.
Matching seniority to the state of your data
A senior analyst is productive quickly when a warehouse, a BI tool and some trusted models already exist. If every team computes revenue differently and nobody owns the definitions, hire a lead who will build the catalogue first. An analytics lead makes sense once several analysts, an experimentation programme and competing priorities need one person setting direction.
Other options
When to hire someone other than a data analyst.
01Consider
You need forecasts or models
Churn prediction, demand forecasting or recommendation work calls for a data scientist rather than an analyst.
Open register02Consider
Events are missing or unreliable
If tracking and pipelines are broken at the source, an engineer who can repair ingestion is the first hire.
Open register03Consider
Reporting inside an ERP
If most of your numbers sit in SAP modules and their reporting stack, an SAP specialist will reach them faster.
Open registerMessy data reveals more than tidy puzzles
Interview questions with clean tables reward memorised syntax. Real warehouses have duplicated events, timezone drift, test accounts and joins that silently multiply rows. Giving a candidate a dataset with those flaws shows whether they check row counts and grain before trusting a result, which is the habit that protects your board pack.
Early wins for a new analyst
Pick one recurring report that people distrust and ask the analyst to rebuild it from source, documenting each definition as they go. It forces access to every system, exposes data quality gaps and produces something visible within a fortnight.
Agree a lightweight intake process at the same time. Without one, a good analyst drowns in small requests from whoever asks loudest, and the deeper questions never get answered.
FAQ
Questions to resolve before the build.
01Which BI tools do your data analysts use?
Most have shipped work in at least two of Looker, Power BI, Tableau and Metabase, and all write SQL directly against the warehouse. We match candidates to the tool you already pay for rather than asking you to switch.
02Is a data analyst enough, or do we need a data scientist?
If the questions are about what happened, why and whether a change worked, an analyst is the right hire. If you need predictions, personalisation or models running in the product, start with a data scientist.
03Can an analyst help us fix conflicting numbers between teams?
That is often the most valuable early project. A senior analyst audits where each team’s figure comes from, agrees one definition with the owners and moves it into a shared model so the disagreement does not return.
Works alongside
Roles a data analyst works best alongside.
01Pair with
Data scientists
When a question needs a predictive model, causal method or forecast rather than a well-built query and chart.
Open register02Pair with
Python developers
To fix tracking at the source, build ingestion jobs and expose clean tables the analyst can rely on.
Open register03Pair with
Salesforce developers
When pipeline and revenue numbers live in the CRM and need cleaner objects, fields and exports before analysis.
Open registerProduct delivery
Need the whole product built instead?
Primary sources
References behind this page
Dated official documentation, standards, and research that support the factual claims on this page.
- 01dbt documentation: About MetricFlow
How dbt defines metrics once in a semantic layer for consistent reuse.
- 02PostgreSQL documentation: Window Functions
Reference tutorial for the window functions analysts rely on daily.
How to read this page
Scope, cost and timeline figures are planning guidance until confirmed in a written proposal. Diagrams and screens are illustrations, not client deliverables, and source-code rights and handover follow the signed agreement. See the delivery process and the full disclaimer.
Next decision
Turn the brief into an accepted product scope.
Define outcomes, constraints, evidence, rights and handover before delivery begins.
Commercial rights, repositories, environments, documentation, acceptance and handover remain contract-defined.