Analytics
AI GA4 analysis that reviews results and updates your marketing strategy
Analytics Strategist closes OrqLabs' feedback loop: it reads GA4, social and campaign data for the chosen period, explains with evidence what worked against your goals, and proposes measured changes to the strategy document the other agents follow.
- As a workflow node
agent.analyticsStrategist- Default model
- Claude Opus 5
- Economy model
- Claude Sonnet 5
- Tools
- 7
- Report
- Analytics review
What the agent does
For GA4 analysis the agent reads, in parallel, the strategy document, the Google Analytics summary, custom GA4 reports by channel group and landing page, post performance for the strategy's channels and campaign statistics. For the goals you pick it builds a KPI table with the current value, the previous equal period, the change, the target where one exists and a status (on track, at risk, off track), and always gives the base (for example “conversions 64 → 48, −25%”).
It diagnoses which channels, topics, formats, posting times and campaigns drove the change, separates signal from noise such as small bases, seasonality and one-off spikes, and says plainly what the data cannot show. Changes are proposed only where the evidence is clear, at most 3–5 per review: channel weights and posts per week, topic weights and content pillars, best times, KPI targets and new experiments.
With Apply changes automatically off (default) the strategy document is left untouched; proposed changes appear in the report with before/after values and reasons, and you are notified. When on, the changes are applied in one go and the other agents, including Content Creator, pick up the new strategy on their next run. Ready-made flow: the weekly GA4 review and strategy update template.
Inputs / Outputs
Inputs
- Period to review
- Last 7, 28 or 90 days (default 7). Results are compared with the previous equal period; with fewer than 500 visits a week, 28 days is more reliable.
- Goals to focus on
- Choose from sessions, engaged sessions, conversions, leads and engagement rate (default: sessions and conversions). Other KPIs with a target in your strategy are added to the table too; 1–3 goals keep the recommendations sharp.
- Apply changes automatically
- When off (default), changes arrive as proposals only. When on, the agent updates the strategy document without waiting for you; keep it off for the first 4–6 weeks, until you trust its proposals.
Outputs
- review
- A Markdown review: summary, KPI table, what worked, what didn't, diagnosis, proposed changes and the next review.
- strategyChanges
- Proposed or applied strategy changes: field, current value, proposed value and a rationale with numbers. A readable diff is returned as well, handy as the description of an approval step in your workflow.
- kpis
- KPI table: metric, current and previous value, change in percent, target where set, and status (on track, at risk, off track, no target).
Saves a report to the Reports page on every run: Analytics review.
Tools
- Analytics summary (GA4)
get_analytics_summary - Custom GA4 report
get_analytics_report - Post performance
get_post_performance - Campaign stats
get_campaign_stats - Read the strategy
get_strategy - Update the strategy
update_strategy - Notify the owner
notify_owner
Example tasks
- “Review last week against our goals. Back every finding with a number and keep strategy changes to at most 3 points.”
- “Find the content topics that brought online bookings in the last 28 days; reweight topics accordingly, but keep every topic above 10%.”
- “Our discount campaign has ended; compare conversions before and after, and revise the messages for each customer group for the next campaign.”
- “Update the experiments list: mark finished ones as won or lost and propose two new hypotheses.”
Tips for good results
- Keep numeric targets in the strategy document (“4,000 visits a week, 60 bookings a month”); the agent can only answer “are we on track?” when a target exists.
- Set guardrails: “Suggest at most 3 changes; never shift channel weights by more than 15% at once.”
- Ask for evidence: “Put the supporting number next to every finding.”
- If your strategy document is empty, the first run proposes a starting point from your data; review it and copy it into the document.
- Connect its output to Report Writer to share the technical review in language a business owner reads easily.
Model advice
Default model
Claude Opus 5
Per 1M tokens: $5 input · $25 output
Economy model
Claude Sonnet 5
Per 1M tokens: $2 input · $10 output
Because this agent changes the strategy every other agent follows, its default model is flagship tier: separating signal from noise and reading multi-source data consistently pays off most here, and a weekly run keeps the cost reasonable. On a tight budget, or when you only want a KPI summary, the balanced model listed as its economy option works; in that case keep automatic apply switched off.
Templates that use this agent
Frequently asked questions
What data does it read?
Your Google Analytics 4 summary plus channel and landing page reports, post performance from connected social accounts, email, WhatsApp and SMS campaign statistics, and the project's strategy document.
Does it work without GA4 connected?
It works with social and campaign data; on the GA4 side only previously stored snapshots can be used. Traffic and conversion goals can only be judged properly with GA4 connected.
Will it change my strategy on its own?
Only when Apply changes automatically is on. Otherwise proposals arrive in the report and a notification, and updating the strategy document is your call.
Should I run it weekly or monthly?
Weekly is recommended. With little data, choose a 28-day period and still run it every week.
Get started with OrqLabs
Build your first workflow today
Start on the free plan: connect your accounts, pick a template and stay in control with approval steps. No credit card required.