How to write a data analyst CV that gets read

What a hiring manager scans for first

Someone screening data analyst CVs is usually doing it fast, often between meetings, and they are looking for two things before anything else: the tools you've actually used, and evidence that you've turned data into a decision someone else made. Not "analysed data to support business objectives" — that sentence tells them nothing and they've read it a hundred times. They want SQL, and which flavour if it matters to them (Postgres, T-SQL, Snowflake). They want to see Python or R named, not implied. They want a BI tool — Tableau, Power BI, Looker — because most analyst roles now sit downstream of a dashboard someone else has to open. If none of that is visible in the first third of the page, the CV gets set aside, not because the candidate lacks skill but because the reader had no time to find out.

The second thing they scan for is scale and stakes: how big was the dataset, how many people used what you built, what changed because of it. A data analyst CV that reads like a data scientist's research summary, all method and no outcome, tends to underperform even when the underlying work was strong.

The vocabulary that has to be there, specifically

Generic CV advice says "use keywords from the job description." For this occupation that advice needs teeth. The words that actually separate an analyst CV from an adjacent one — data scientist, BI developer, data engineer — are things like: SQL joins and window functions, ETL or ELT (and which tools — dbt, Airflow, SSIS), data cleaning and wrangling (pandas, or dplyr if R), A/B testing and experiment design, cohort or funnel analysis, dashboarding and self-serve reporting, and the specific warehouse or platform (BigQuery, Redshift, Snowflake, Databricks). If you use Excel seriously — pivot tables, VLOOKUP/XLOOKUP, Power Query — say so plainly; some employers, especially outside pure tech, still filter on it and dismissing it as beneath the CV is a mistake.

Avoid vocabulary that belongs to a neighbouring role unless you actually did that work. "Built and deployed machine learning models" is a data scientist or ML engineer claim; if what you did was build a logistic regression in a notebook to inform a report, say that instead. Hiring managers in this field can usually tell the difference within a sentence, and an inflated claim costs more credibility than a modest true one gains.

Qualifications and certificates: what actually carries weight

There is no licence to hold for this job and no professional register to belong to — unlike, say, an accountant or an engineer, a data analyst's legitimacy rests entirely on demonstrated work, not credentials. A degree in a quantitative subject (statistics, economics, maths, computer science) is common but not universal, and plenty of analysts move in from other backgrounds with a portfolio instead. If you have one, name the subject; "BSc" alone tells the reader nothing.

Certificates are contested territory. Google's Data Analytics Certificate, Microsoft's PL-300 (Power BI), and vendor-specific credentials from Tableau or SQL providers do get mentioned in job adverts, particularly in less senior postings, and putting one on the CV signals you've covered a defined syllabus. What they do not do is substitute for shown work — no hiring manager in this field treats a certificate as equivalent to a dashboard that's still in production use somewhere. If you have limited experience, list the certificate briefly and spend the space you save on a project instead of a longer certificate list.

How experience actually gets evidenced in this field

The convention in this occupation is to evidence experience through outcomes tied to a specific artefact — a report, a dashboard, a model, an experiment — not through a list of duties. "Responsible for weekly sales reporting" says nothing a hiring manager can check or picture. "Rebuilt the weekly sales dashboard in Power BI, cutting manual reporting time from two days to under an hour, adopted by three regional teams" is checkable, specific, and tells them what you're capable of doing again.

Numbers matter here more than in most fields, because the whole discipline is about quantifying things, so a CV that describes analytical work without a single figure looks like a contradiction. But the number needs to be yours to claim — if the dashboard's adoption or the revenue lift was a team result, say "contributed to" honestly rather than implying sole ownership; in a field this small, overclaiming on a specific project is often checkable by a hiring manager who knows someone who was there.

Where formal metrics aren't available — a lot of internal analytics work never gets a clean before/after number — describe the decision the analysis fed into and who acted on it. "Analysis was used by the pricing team to adjust the March promotion" is weaker than a percentage but far stronger than nothing, and it's honest about what you can actually verify.

A link to a portfolio, GitHub, or a couple of anonymised dashboard screenshots does more for this occupation than for almost any other, because the deliverable is visual and inspectable. If you have one, put the link near the top, not buried at the bottom under "references available on request."

What candidates from this background routinely leave out

Three things show up missing again and again on data analyst CVs.

First, the data cleaning and wrangling work itself. Analysts often skip straight to the finding because the cleaning felt like unglamorous groundwork, but a hiring manager knows that most real analyst time goes into getting messy, inconsistent, or incomplete data into a usable state, and they want to know you can do that, not just query a tidy table. A line like "consolidated data from four disconnected CRM exports with inconsistent field naming" tells them more about your practical competence than another dashboard bullet point.

Second, stakeholder communication. This role sits between raw data and people who make decisions without technical background, and the ability to explain a finding to someone in marketing or operations without jargon is a distinct, checkable skill that candidates rarely mention because it doesn't feel like "analysis." If you've presented findings to non-technical stakeholders, written a report a director actually read, or changed how a team measured something, say so directly.

Third, the domain context. "Data analyst" covers people working in marketing, finance, product, healthcare, logistics, and more, and the specific business questions each domain asks are different — churn and lifetime value in a subscription business, conversion and attribution in e-commerce, claims and risk in insurance. Leaving domain context off the CV makes every analyst look interchangeable, when in practice a hiring manager in fintech usually wants to see you've handled financial or transactional data before, not just data in general.

What to do next

Go through your CV and check that the first third of the page names your core tools by their actual names — SQL, the specific BI tool, Python or R — not a paraphrase of them. Then check every bullet point under experience: does it name an artefact (a dashboard, a report, a model, an experiment) and a result someone else acted on? If a bullet only describes an activity, rewrite it or cut it. Add one line about data cleaning or wrangling if it's missing, and one about a non-technical audience you've reported to, since both are commonly assumed rather than stated. If you have a portfolio or dashboard samples, link them near the top.

If you're applying to more than a handful of roles and each one wants the CV framed slightly differently — different tools foregrounded, different domain experience surfaced — that rewriting is where most people run out of time before they run out of applications. jobmarket.pro reads each advert and prepares the application from one profile it can't invent experience into, which is the specific problem this article has been describing.