How do I know whether the fix actually worked?

You need before-and-after evidence in the browser and in the downstream report. A clean-looking dashboard without validation is not enough.

When should this become a full GA4 audit instead of a quick fix?

If the issue touches attribution, consent, revenue, campaign quality, or data trust for more than one workflow, it is usually safer to audit the surrounding implementation than patch only the visible symptom.

How to Hire a GA4 Analyst or Consultant: Skills, Red Flags, and Interview Questions (2026)

What skills should a GA4 analyst have in 2026?

A competent GA4 analyst in 2026 must have three skill layers: (1) Platform knowledge (GA4 event model, custom dimensions, Explorations, BigQuery export, Consent Mode V2), (2) Implementation capability (GTM, dataLayer, event specification, validation via DebugView), and (3) Commercial application (translating data into budget, CRO, and product decisions). The minimum viable skill set for a mid-market GA4 analyst: Able to build a measurement plan, implement custom events in GTM, configure Consent Mode V2, validate implementation in DebugView, build Looker Studio dashboards from GA4 data, and run basic BigQuery SQL queries.

GA4 analyst vs GA4 consultant: the difference

GA4 analyst (in-house role):

Maintains ongoing property health (auditing, annotation, configuration)
Builds reports and dashboards for stakeholders
Interprets data and makes recommendations
Implements tracking changes via GTM
Manages GA4 access and governance

GA4 consultant (external specialist):

Audits and fixes existing implementations
Designs measurement plans for new projects
Implements complex tracking (server-side, cross-domain, app+web)
Provides training and documentation
Works on a project/retainer basis, not ongoing maintenance

The gap many businesses fall into: Hiring a junior "analyst" who can build dashboards but lacks implementation capability, then paying expensive consultants for every tracking change.

The 5 red flags in GA4 candidates

Red flag 1 — "I'm Google Analytics certified"

GA4's certification from Google is a basic multiple-choice test covering surface-level platform knowledge. It does not assess implementation capability, BigQuery SQL, Consent Mode, or the analytical judgement that makes a hire valuable. Certification is a starter credential, not a differentiator. Treat it as table stakes, not evidence of expertise.

Red flag 2 — Cannot explain Consent Mode V2

In 2026, a professional GA4 analyst who cannot explain what Consent Mode V2 is, why it's required for EU/UK Google Ads, and what the gcd parameter indicates is not current. This is a fundamental operational requirement — not advanced knowledge.

Red flag 3 — "I can see the data in GA4" but cannot explain why it might be wrong

Sophisticated GA4 analysts maintain professional scepticism. When asked "why might your GA4 conversion data be wrong?", a strong candidate lists: consent rejection, duplicate transactions, wrong attribution model, missing cross-domain config, bot traffic. A weak candidate says "the data should be right."

Red flag 4 — No BigQuery experience for a senior role

Any senior GA4 analyst (5+ years experience) hired in 2026 should have working SQL proficiency and be comfortable querying the GA4 BigQuery export. "I work only in the GA4 UI" is a significant limitation for a senior hire.

Red flag 5 — Cannot explain the difference between a session and a user

This is genuinely foundational. If a candidate conflates sessions and users — or can't explain why GA4 shows fewer sessions than UA — they lack the analytical foundation to produce trustworthy reports.

The 10 interview questions

Technical knowledge (4 questions)

Q1: A client's GA4 shows only 2 months of Exploration data. What happened and what should have been done?

Strong answer: Data retention was left at the default 2 months. Should have been changed to 14 months immediately when the property was created. Not retroactive — historical data is permanently gone.

Need a faster way to turn GA4 problems into a client-ready audit workflow?

Start free audit

Q2: Explain Consent Mode V2. What are the four parameters and what happens when analytics_storage is denied?

Strong answer: Four parameters: analytics_storage, ad_storage, ad_user_data, ad_personalization. When analytics_storage is denied, GA4 sends a cookieless ping (in Advanced mode) rather than a full tracking hit. Sessions are not recorded in standard reports for non-consenting users, but the ping enables Google to model aggregate conversions.

Q3: A client's GA4 revenue is 35% higher than Shopify's actual revenue. What are the most likely causes?

Strong answer: Duplicate transactions (most likely — order confirmation page viewed multiple times), refund events not being tracked, or a purchase event firing on page load of the thank-you URL rather than on actual order completion.

Q4: What is the `gcd` parameter in a GA4 network request and what does `gcd=11l` mean?

Strong answer: The gcd parameter encodes consent state at the time of the hit. gcd=11l means consent was denied but Consent Mode Advanced mode is active — a cookieless ping is being sent to allow Google to model aggregate behaviour.

Implementation capability (3 questions)

Q5: Walk me through how you'd implement a purchase event on a React SPA.

Strong answer: Describes using GTM with the History Change trigger for SPA page views, pushing purchase data to the dataLayer from the React component's order confirmation state (not via page URL matching), using a Custom HTML tag or GTM Community Template for GA4 event. Mentions validation in DebugView.

Q6: How would you prevent the same purchase from being tracked twice in GA4?

Strong answer: Describes sessionStorage-based deduplication: check if transaction_id is already stored before pushing to dataLayer. Mentions server-side approaches (backend flag like _ga4_tracked meta on the order). For Shopify, uses checkout_completed Custom Pixel event.

Q7: A client says their Paid Social traffic shows in (Other) in GA4. What do you check first?

Strong answer: utm_medium value. Likely paid_social (with underscore) instead of paidsocial (without). GA4's Paid Social channel group requires the medium to match specific patterns that don't include underscored variants.

Commercial application (3 questions)

Q8: A client's GA4 shows 45% mobile traffic but mobile converts at 1.1% vs 4.2% for desktop. What do you recommend?

Strong answer: Identifies this as a CRO opportunity — the mobile/desktop CVR gap. Recommends: mobile funnel analysis in GA4 Explorations to identify specific drop-off points, page speed audit on mobile, UX review of mobile checkout flow. Quantifies the revenue opportunity: if mobile CVR improved to 2%, estimated additional revenue is X.

Q9: You're presenting GA4 data to a CMO who says "your numbers look lower than what we had in Universal Analytics." How do you respond?

Strong answer: Explains the structural reasons GA4 shows fewer sessions (no midnight reset, no new session on campaign change, consent mode exclusions). Frames it as methodological improvement rather than data loss. Notes that trend direction is still valid; absolute counts differ.

Q10: The marketing team wants to track every click on the site as a key event to "see which elements get the most attention." How do you respond?

Strong answer: Explains that marking click events as key events inflates key event count to the point of meaninglessness, corrupts Smart Bidding by giving Google Ads the wrong conversion signal, and doesn't answer the actual question (which elements drive conversions, not which get clicked). Recommends custom click events with specific element parameters, used in Free Form Exploration for analysis — not marked as key events.

Evaluating an agency's GA4 capability

Ask for a sample report from a current client (anonymised). A strong agency's sample report shows: channel quality metrics (not just volume), conversion rate trends with annotations explaining changes, and specific recommendations with ROI estimates. A weak agency's sample shows sessions and bounce rate.

Ask "what was the last GA4 property issue you diagnosed and fixed?" Strong agencies describe specific technical problems (duplicate transactions, broken cross-domain tracking, Consent Mode V2 failure). Weak agencies describe setting up a dashboard.

Ask "do you implement Consent Mode V2 for all EU/UK clients?" The answer must be yes with a clear description of the implementation approach. If they say "Google handles that" or show uncertainty, walk away.