Conversion Rate Optimisation Jobs: Skills, Tools, Salaries

Conversion rate optimisation jobs cover roles from CRO Analyst to Experimentation Engineer. Expect to demonstrate GA4, GTM, testing tools (VWO/Optimizely), SQL, and UX research. In the UK, typical salaries range from £28k–£80k, with common interview tasks including funnel audits, test plans, and simple SQL or GTM exercises.

If you’re moving into CRO or levelling up, this guide shows the exact skills, tooling, salaries, and interview patterns we see when hiring and delivering client work.

What CRO roles exist (and what they actually do)

Most job ads bundle responsibilities. Here’s how the work usually splits on real teams:

• CRO Analyst / Strategist: Owns research, problem framing, and prioritisation. Designs test hypotheses, writes briefs, and reports impact to the business. Liaises with product, marketing, and dev.
• Experimentation Engineer / Developer: Builds A/B/n tests, feature flags, and personalised experiences. Works in JavaScript/TypeScript and HTML/CSS, sometimes React. Knows QA and performance.
• UX Researcher (Quant/Qual): Runs user interviews, surveys, and usability tests. Triangulates with analytics and session recordings.
• Analytics & Tracking Specialist: Implements GA4, GTM (web and server-side), models events, maintains dataLayer, and validates data quality.
• Data Analyst (Experimentation): Designs sample size, segments, and power analysis. Crunches results in SQL/Python, maintains dashboards in Looker Studio or Metabase.

Titles vary. In smaller teams, one person covers strategy + light implementation. In larger orgs, roles are specialised and sit within Product or Growth.

Core skills and tools hiring managers test for

Expect take-homes or live exercises that use these stacks (we use most of them in production):

• Analytics & tracking:
  • GA4 (events, conversions, Explorations, consent mode v2), GTM web + server-side, dataLayer design.
  • PostHog or Mixpanel for product analytics.
  • BigQuery or Postgres for raw data. SQL fluency is a differentiator (CTEs, window functions).
• Experimentation:
  • VWO, Optimizely, AB Tasty, Convert.com, or PostHog Experiments. Understand Stats Engine vs frequentist, MDE, power.
  • Feature flags with LaunchDarkly or GrowthBook (open-source) for server-side tests.
• UX research & behaviour:
  • Hotjar or Microsoft Clarity for recordings, heatmaps, and in-page surveys.
  • UsabilityHub, UserTesting or Maze for remote tests.
• E‑commerce context:
  • Shopify Admin API and GraphQL for custom data pulls; Shopify theme/section updates (Liquid, JSON templates).
  • Klaviyo flows for post-test follow-up and lifecycle uplift.
• Data pipelines & reporting:
  • n8n or Zapier for light automation; Airbyte for ELT; dbt for modelling.
  • Looker Studio or Metabase for stakeholder-friendly dashboards.
• QA & performance:
  • BrowserStack for cross‑device testing, Percy or Chromatic for visual diffs, Lighthouse/WebPageTest for speed.

Soft skills matter, but they’re demonstrated through outputs. Bring test specs, dashboards, and annotated code/screenshots rather than adjectives on a CV.

Real metrics: what good looks like

Hiring managers want signal you can separate noise from impact. Use specific, defensible numbers and context:

• Baseline and segmentation: “PDP conversion was 2.1% (mobile 1.6%, desktop 2.9%). Bounce 58%. 65% mobile traffic.”
• Hypothesis clarity: “If we compress hero images and defer non-critical scripts, LCP improves ~600ms, lifting mobile CR by 5–8% relative.”
• Sample size: “At 2.1% baseline, MDE 10% relative (2.31%), 80% power, α=0.05, needs ~85k sessions per variant.” Rough is fine if it shows method.
• Test outcomes with sanity checks: “Variant B reduced PDP image weight by 420KB, mobile LCP −780ms; added-to-cart +8.4%, CR +0.6pp (2.1% → 2.7%), revenue/session +9.1%, sample ratio OK, device-balanced.”
• Secondary effects: “No negative impact on return rate or NPS. Email sign-ups +12% from sticky footer variant.”

Pitfalls you should call out unprompted:

• Peeking and optional stopping inflate false positives. Use sequential methods or discipline. Understand Optimizely Stats Engine and VWO’s SmartStats.
• Sample ratio mismatch indicates bucketing or tag issues. Check redirects, JS errors, and consent gating.
• Seasonality and promos distort baselines. Avoid Black Friday tests unless you’re testing promo mechanics.
• ITP and consent: Safari traffic will undercount without server-side tagging and proper consent integration.

If you can talk to these, you’ll stand out.

How to build a CRO portfolio without big-brand logos

You don’t need Nike on your CV. You do need artefacts that show your thinking and execution.

• Run your own tests:
  • Use PostHog or GrowthBook on a personal site (Next.js + Vercel works well). Drive ~1–2k sessions/month via content or small PPC spend (£150–£300) to get a few micro‑tests live.
  • Example: headline copy test, sticky add-to-cart on mobile, or image compression vs quality. Document hypothesis, MDE calc, variant code snippet, QA steps, and results.
• Do structured teardowns:
  • Pick 3 Shopify stores in a niche. Audit PDP, PLP, and checkout. Provide 5 prioritised recommendations using ICE or PXL. Include Shopify Liquid change examples.
• Publish a tracking plan:
  • Draft a GA4 event schema + dataLayer for an e‑commerce flow (view_item, add_to_cart, begin_checkout, add_payment_info, purchase). Show GTM screenshots and a Looker Studio funnel.
• Show simple data work:
  • A Metabase chart pack or a short SQL notebook in BigQuery/Postgres calculating CR by device/source with confidence intervals.

Package everything into a clean Notion or GitHub Pages site. Hiring teams want to scan, click, and see proof in 5 minutes.

Where to find conversion rate optimisation jobs in the UK

You’ll find roles scattered across Marketing, Product, and Data. Cast a wide net but use the right filters.

• Job boards and communities:
  • LinkedIn Jobs (use keywords: “CRO”, “Experimentation”, “A/B testing”, “Optimisation”, “Growth”).
  • Otta and WorkInStartups for product-led startups.
  • Indeed for broader market coverage; set alerts by city and remote.
  • Experiment Nation job board; MeasureCamp Slack #jobs; GrowthHackers community.
• Agencies and consultancies: Look at UK agencies with CRO/experimentation practices. Even if there’s no live ad, send a tight portfolio and a 5‑slide teardown.
• Direct-to-brand: Retail, travel, fintech, and SaaS are strong. Shopify Plus merchants often hire hybrid CRO/Shopify roles.

Salary guide (rough UK ranges, permanent):

• Junior/Associate: £28k–£35k
• Mid-level CRO / Experimentation Analyst: £40k–£55k
• Senior CRO / Experimentation Lead: £60k–£80k
• Contractor day rates: £350–£650 depending on stack and delivery (server-side testing, analytics migrations, or Shopify theme dev trend higher)

London is 10–20% higher. Hybrid is common. Fully remote is credible if you show strong async documentation.

The interview loop: tasks, take‑homes, and how to prepare

What we and peers typically do:

• Screen (30 mins): Walk through two projects. Expect “what was the MDE?” and “how did you QA?”
• Take‑home (2–4 hours):
  • Audit a mobile PDP and propose 3 tests with hypotheses, impact/effort, and success metrics.
  • Write a basic test spec (brief, targeting, guardrails, metrics, launch/stop rules).
  • Implement or pseudo‑code a dataLayer for checkout steps.
  • Short SQL task: calculate CR and revenue per session by device and source over 8 weeks (handle bots and missing values).
• Technical deep dive (60 mins): Review your work. Might include debugging a sample ratio mismatch or outlining server-side testing on Shopify via a proxy/app embed.
• Stakeholder round (30–45 mins): Communicating trade-offs, reporting results, dealing with HiPPO-driven ideas.

Preparation checklist:

• Refresh GA4 Explorations, VWO/Optimizely interface, and Looker Studio model building.
• Practice a power/MDE calc and know when a test is underpowered.
• Have a crisp view on consent, ITP, and server-side tagging.
• Prepare one failure story (bad test, wrong metric) and what you changed afterwards.

Agency vs in‑house vs freelance: which path fits you

• Agency (our world): Fast cycles, mixed sectors, solid learning curve. You’ll ship. Downsides: context switching, tight timelines.
• In‑house: Deeper product context, richer data, longer experiments, easier to influence roadmap. Downsides: slower pace, stakeholder politics.
• Freelance/contract: Autonomy and day rates. You’ll need a repeatable sales pipeline and ops (invoicing, proposals, insurance). Good for specialists (server-side testing, GA4 migrations, Shopify CRO).

A practical hybrid: full‑time in‑house plus one small retained client on Fridays (with employer’s consent). Keeps skills sharp and de‑risks salary plateaus.

What we look for at Streamline Digital (and how to stand out)

We’re a Bournemouth agency building AI workflow automation, Shopify integrations, custom APIs and SEO/CRO systems. Our CRO work is anchored in data and shipping:

• You can instrument events cleanly in GA4/GTM and PostHog, including consent mode and sGTM.
• You can build lightweight, robust tests in VWO/Optimizely or GrowthBook, with clean rollback and QA.
• You understand Shopify (Liquid, sections, app embeds) and can query Shopify Admin API/GraphQL when analytics is murky.
• You write short, clear specs and follow through with results that finance can trust.

If you’re a merchant considering outside help, our Conversion Rate Optimisation service page explains how we approach research, testing, and engineering. If you’d like to sanity‑check your stack or discuss a role, book a free discovery call.

What makes a CV get a callback:

• Two crisp case studies with baselines, MDE, lift, time to significance, and screenshots or code snippets.
• Evidence of tooling (not just logos): a GTM container export, a GrowthBook config, or a Metabase dashboard link.
• Pragmatism: you know when a test isn’t worth it and ship a high‑confidence change instead (e.g., fixing a broken coupon field).

Practical examples from the field

• Speed → money on mobile PDPs:
  • Context: Shopify Plus retailer, 70% mobile, LCP 3.8s.
  • Work: lazy-load non-critical JS, image CDN params, remove unused app scripts. Measured with Lighthouse and WebPageTest; monitored in GA4 and VWO.
  • Result: LCP −900ms; add‑to‑cart +7.2%; CR +0.5pp (1.9% → 2.4%); statistically significant in 18 days (~120k sessions/variant).
• Bundle builder vs discount banner:
  • Context: AOV stuck at £42.
  • Work: Built bundle widget (React) behind feature flag (GrowthBook), compared to 10% off banner.
  • Result: Bundle variant +12% AOV, no CR dip. Revenue/session +14%. Keep. Flag rolled out to 100%.
• Checkout tracking repair:
  • Context: GA4 purchases undercounted by ~18% on Safari.
  • Work: Implemented sGTM with server-side UA, consent mode v2, and server-to-server webhook from Shopify to BigQuery.
  • Result: Safari delta reduced to <3%. Finance reconciles revenue within ±1.5% weekly.

These are the kinds of outcomes to foreground in your portfolio.

Common mistakes that cost offers

• Treating CRO as just copy tests. Engineering changes (speed, dataLayer, checkout bugs) often move revenue more.
• Overclaiming causality from before/after changes with no controls. If you can’t test, at least provide a guardrail and a pre/post segmentation with seasonality context.
• Ignoring QA. A quick BrowserStack pass across iPhone + Android + Safari desktop + Chrome stops expensive rollbacks.
• No link to the money. “+20% click-through” means nothing without revenue/session or profit impact.

FAQ

What entry-level conversion rate optimisation jobs can I get without agency experience?

Look for Junior CRO Analyst, Optimisation Executive, or Marketing Analyst roles with exposure to GA4 and a testing tool. Bring a small portfolio (even from your own site) and show you can write a test brief, set up events in GTM, and report results cleanly.

Do I need certification for CRO roles?

Certs help a bit (GA4, VWO, Optimizely), but shipped work trumps badges. Two well-documented experiments and a tracking plan will beat three certificates nine times out of ten.

How technical do I need to be for experimentation roles?

Enough to be dangerous. You should read and write basic JavaScript for test variants, edit HTML/CSS, and write SQL to aggregate metrics. For engineer roles, TypeScript, React, and familiarity with feature flags and SDKs are expected.

How long should an A/B test run?

Long enough to hit your pre‑calculated sample size with stable traffic. As a rule of thumb, plan for 2–4 weeks on medium‑traffic sites. Avoid stopping early on a spike; use sequential methods if your platform provides them and sanity‑check for sample ratio mismatches.