Chapter 28 — Exercises

Work these in order; they graduate from recall to judgment. Items marked † have full worked solutions in the answers appendix. No answers here — struggle first, then check the quiz and the appendix. Everything below uses free tools (GA4, DebugView, the Explore section) and constructed data; every dollar or count is illustrative.

A. Recall and definitions

  1. † In one sentence each, define: event, engaged session, engagement rate, key event, and attribution.
  2. What product did GA4 replace, and on roughly what date did Google stop processing data in standard properties of the old product?
  3. † State GA4's exact definition of an engaged session (all three conditions), and give the formula relating bounce rate to engagement rate in GA4.
  4. Name the four tiers of GA4 events by setup effort, from "no work" to "full implementation," with one example of each.
  5. † GA4 renamed a concept in 2024. What was it called in Universal Analytics, what did early GA4 call it, and what is it called now?

B. Applied reasoning

  1. † Explain the difference between the Traffic acquisition and User acquisition reports, and say which you'd use to answer "how much did organic search drive this month, and how did those visits behave?" — and why.
  2. Why does this book call Direct traffic a "junk drawer," and why might a rise in Direct actually be good news for an SEO?
  3. † A colleague says, "Our engagement rate went from 55% to 68% after the redesign, so our rankings should climb." Identify the reasoning error, then state the legitimate value of that engagement improvement.
  4. Explain why last-click attribution tends to make a genuinely good SEO program look like a poor investment. Use the words first touch and branded search.
  5. † Why can't any tool — GA4, Search Console, or a paid platform — join a specific organic query to a specific person's specific conversion? Name the historical change that created this limit.
  6. GA4 reports a booking as a "conversion." List three distinct real-world things that a booking-request key event does not prove.
  7. † A low-traffic site switches to data-driven attribution and gets erratic, hard-to-explain channel credit month to month. Give the most likely reason, drawn from how DDA works.

C. Read this report

  1. † Read the report. Organic landing pages, last 28 days (constructed):

    Page Sessions Engagement rate Avg engagement time Key events
    /blog/furnace-short-cycling 1,100 74% 3m 10s 2
    /ac-repair 400 62% 1m 40s 21
    /locations/millhaven 300 29% 0m 18s 1
    /book 70 55% 0m 58s 33

    For each page, say in one line what it's doing and what (if anything) you'd change. Which page is the alarm, and which would a traffic-only report wrongly call your worst performer?

  2. In the table above, the /book page has the fewest sessions but the most key events. Explain why that is exactly what you'd expect, and why it would be a mistake to "fix" its low traffic.

  3. † You open a GA4 report and several rows you expected are simply missing, with a small icon noting data was withheld. Name the GA4 feature responsible, explain why it exists, and say whether this means your tracking is broken.

  4. Your organic Landing page report shows 22% of sessions under a row labeled (not set). Is a small amount of this normal? What does a large share indicate, and what would you check?

D. Diagnose the number

  1. † Rivertown's office logged 140 phone calls last month; GA4 shows 30 phone-call key events from organic. The owner says the tracking is broken. Give two reasons the GA4 number can be genuinely and correctly lower, and state where the missing calls can actually be seen.

  2. A page's organic sessions doubled month over month, but key events stayed flat. Give two non-alarming explanations and one that would warrant investigation.

  3. † GA4 shows 1,284 organic sessions this month vs. 900 last month. Which part of that comparison should you trust, which part should you not over-read, and how should you phrase it in a report?

  4. Someone reports "our SEO drove exactly 52.0 conversions and 1,284 sessions." Rewrite the sentence so it is honest about GA4's precision, without making the data sound worthless.

E. Audit your own site (free tools)

  1. † In your own (or a demo) GA4 property, open Reports → Acquisition → Traffic acquisition, filter to Organic Search, and record Sessions, Engaged sessions, and Key events for the last 28 days. Then write one sentence stating what this baseline does and does not tell you.

  2. Using Admin → Events, identify one event that represents business value on your site and mark it as a key event. Then open DebugView, perform the action on your own site, and confirm the event fires. Note what you'd do if it didn't appear.

  3. † Build an organic segment in the Explore section ("sessions where Session default channel group = Organic Search"). Apply it to a free-form exploration broken down by landing page. What can you now see that the standard reports made awkward?

  4. Add Session source / medium as a secondary dimension to your organic view and split it into google / organic, bing / organic, etc. Report the share from each, and note any source that looks like a bot rather than a real search engine.

F. Spot the myth

  1. † For each statement, say whether it's a myth or true, and correct the myths in one line: (a) "A high bounce rate in GA4 drags down your Google rankings." (b) "GA4 shows you the keywords that brought your organic traffic." (c) "Data-driven attribution tells you each channel's true contribution." (d) "In GA4, a pageview is a special kind of hit, separate from events." (e) "Google uses your Analytics engagement data as a ranking signal."

  2. "GA4 killed bounce rate because a high bounce rate was hurting everyone's SEO." Identify the two separate errors packed into this one sentence.

G. Ethics, privacy, and risk

  1. † A developer proposes capturing each form's phone number and email as parameters on the form_submit event "so we can match conversions to real people in GA4." State the policy problem, the privacy problem, and the correct place for that personal data.

  2. Explain, at a conceptual level and without giving legal advice, what GDPR and CCPA/CPRA each broadly require of a site that uses analytics, and how they differ in posture (consent vs. opt-out).

  3. † Describe what Consent Mode does when a visitor denies consent, and explain the phrase "a portion of what GA4 reports is modeled, not counted." Why does this make the "funnel of loss" important to disclose in a report?

  4. A vendor offers to "recover the traffic you're losing to cookie banners and ad blockers" using techniques that sidestep consent signals. Using this book's ethics, explain why that's the analytics equivalent of a black-hat SEO tactic, and what the honest alternative is.

H. Rivertown Strategy-File extension

  1. † Rivertown wants to add a fourth conversion beyond form fills, tel-clicks, and bookings. Propose one that fits a home-services business, say how you'd capture it as a key event (no PII), and state one thing it would and wouldn't prove.

  2. † Rivertown's booking runs through a third-party scheduling widget embedded in an iframe from a different domain. Explain why the booking key event might silently under-count or miss conversions, and what you'd flag to a developer before trusting the number.