Session 22: Analytics Deep-Dive — GA4, Meta Dashboards & Attribution

Block 4: Email Automation, Influencer Briefs, Analytics & Finalizing A2

View lecture slides: Session 22 Slides

1-Hour Lecture (Theory & Strategy)

Phase 1 introduced GA4/Meta Insights reading. Now: cross-channel attribution — which touchpoint gets credit when a customer sees an ad, later searches organically, then converts? Last-click vs. multi-touch views, and why managers should distrust single-channel ROAS claims in isolation (recall the home services stack case’s “separate layers” lesson).
UTM tagging discipline: every campaign link tagged with source/medium/campaign so channel performance isn’t guesswork.
Read the full theory here: Intro to Web Analytics, Google Analytics Tools, and Key Metrics.
Recap the Phase 1 reference: Analytics & CRO Reference.

flowchart LR
    AD[Paid Ad Touch] --> SEARCH[Organic Search]
    SEARCH --> CONV[Conversion]
    GA4[GA4 Dashboard] --> SRC[Top Traffic Sources]
    META[Meta Insights] --> SRC
    UTM[UTM-Tagged Calendar Links] --> SRC
    SRC --> MEMO[One-Page Insight Memo]
    MEMO --> S23[Feeds Session 23 Report]

Figure 1: Cross-channel attribution and insight memo workflow

Key Concepts in Depth

Attribution decides which touchpoints get credit for a conversion:

Model	Credit goes to	Bias
Last click	Final touch before conversion	Over-credits search and brand terms
First click	First touch	Over-credits awareness channels
Linear	All touches equally	Simple, but ignores differences in influence
Data-driven (GA4 default)	Modelled contribution of each touch	Needs enough conversion volume; a “black box”

No model is “true”. Managers should compare models and ask what would happen if a channel were switched off. Platform dashboards (Meta, Google Ads) each claim credit for the same sale, so their totals add up to more than real conversions.

UTM discipline makes this possible. Every campaign link carries utm_source, utm_medium and utm_campaign, using one naming convention: all lowercase, underscores, no spaces, and consistent medium names (paid_social, cpc, email, influencer).

An insight memo has four parts: what happened (data), why (diagnosis), what to do (one recommendation), and expected impact (a KPI with a number).

Worked Example: Pathshala Tutors (Running Case)

UTM convention:

Link	utm_source	utm_medium	utm_campaign
Meta mock-test ad	facebook	paid_social	ielts_mocktest_2026_11
Welcome email Day 5	newsletter	email	welcome_flow
Influencer Story	instagram	influencer	riya_mocktest_2026_11

One journey, three views. Sita sees a Meta ad (Monday), searches “pathshala tutors” organically (Wednesday), then clicks a Google ad and books (Friday).

Model	Meta	Organic	Google Ads
Last click	0	0	1
First click	1	0	0
Linear	0.33	0.33	0.33

Under last click, Meta looks worthless and might be cut, even though it started the journey.

Insight memo (extract): “Google Ads drives 50% of last-click leads, but 40% of those leads had a Meta touch earlier. Recommendation: keep Meta at NPR 5,000 and shift NPR 2,000 from broad-match search to brand and remarketing. Expected impact: CPL down from NPR 500 to about 430 within 6 weeks.”

1-Hour Lab (Agency Execution)

Using your Phase 1 GA4/Meta data (or the sample dashboard data provided by your instructor), identify the top 3 traffic sources and their conversion rates.
Build a simple UTM naming convention table for your Session 13 content calendar links (source, medium, campaign name).
Write a one-page insight memo: what’s working, what’s underperforming, and one specific, evidence-based optimization recommendation with an expected KPI impact.
Local data baseline: Compare your findings against the Education PPC case for what a real optimization cycle looked like.

Common Mistakes

Adding up platform-reported conversions and presenting the total as real results.
Inconsistent UTMs (Facebook, facebook, fb) that split one channel into three in reports.
Cutting awareness channels because last-click reports undervalue them.
A memo with observations but no recommendation or expected impact.

Deliverable: Insight Memo (feeds Session 23 report)

A data-driven insight memo — direct input into your Session 23 stakeholder report.

Summary

Cross-channel attribution assigns credit for conversions; each model has a bias, so compare models instead of trusting any single one. Consistent UTM tagging makes channel data reliable. The insight memo turns data into a diagnosis, one recommendation and an expected, measurable impact.

Key Takeaways

Every attribution model is a lens, not the truth.
Platforms over-claim; reconcile them with GA4 or your CRM.
One UTM convention, applied everywhere.
Insight = data + diagnosis + recommendation + expected impact.

Review Questions

Short answer

Compare last-click and first-click attribution.
What are the three required UTM parameters?
Why can the conversions reported by Meta and Google Ads add up to more than actual sales?

Applied (case) questions

A customer clicks an Instagram ad, opens two emails, and converts via an organic search. Allocate credit under last-click, first-click and linear models, and discuss which channel is at risk of being under-funded.
Write a 150-word insight memo for a restaurant whose online orders rose 20% while Meta CPC rose 35%.

Session 22: Analytics Deep-Dive — GA4, Meta Dashboards & Attribution