Building a measurement plan: the capstone

Twenty-one parts have covered GA4 largely in the order features exist in the product. A real implementation runs in a different order — it starts with a measurement plan, written before any tag is installed, and every part of this series slots into that plan at a specific, defined point. This capstone writes one, real and complete, for a new GreenDesk product line: a self-serve, lower-priced tier launching alongside the existing sales-assisted product.

Step 1: define the real business questions first, not the events

Q1: Are self-serve signups converting to paid without ever talking to
    sales, at a genuinely sustainable rate?
Q2: Which channels bring in self-serve signups that actually activate
    (use the product), versus ones that sign up and never return?
Q3: How much self-serve revenue is being pulled forward from what
    would have been a sales-assisted deal, versus genuinely new
    revenue GreenDesk wasn't capturing before?

This is the real starting point a measurement plan gets wrong most often when skipped — starting from "what events should we track" instead of "what real questions do we need answered" produces tracking that's technically complete but doesn't actually answer anything decision-relevant.

Step 2: map each question to the specific events that answer it

Q1 needs: sign_up, self_serve_activation (custom event), purchase
Q2 needs: sign_up with channel data (part 6's Acquisition reports),
          self_serve_activation
Q3 needs: purchase, plus a custom parameter distinguishing
          self_serve vs. sales_assisted deal type

Every event and parameter here traces back directly to a real question from step 1 — nothing is added speculatively, avoiding the exact custom-dimension budget waste flagged in part 16's real 50-dimension ceiling.

Step 3: define event and parameter names before any code is written

Event: self_serve_activation
Parameters:
  plan_tier: "self_serve"
  days_since_signup: 3
  first_feature_used: "task_board"

Naming this deliberately, in the plan, before implementation begins avoids a real, common problem: a developer inventing an ad-hoc event name during implementation that doesn't match the naming conventions used everywhere else in this series (part 3's recommended-event-naming guidance), creating inconsistency that's expensive to untangle later.

Step 4: register custom dimensions and metrics in the plan, not as an afterthought

Custom dimension: Plan Tier (event-scoped, from plan_tier parameter)
Custom dimension: First Feature Used (event-scoped)
Custom metric: none needed — no new numeric value beyond value/currency
  already covered by the standard purchase event (part 9)

This is part 16's registration step, done deliberately as part of planning rather than discovered missing weeks after launch when a report shows no "Plan Tier" dimension available.

Step 5: define the real key events (conversions) explicitly

Key event: self_serve_activation
Key event: purchase (already marked key, part 8 — applies to both
  plan types, distinguished by the plan_tier parameter)

Following part 8's guidance directly — marking too many events as key events dilutes GA4's own optimization and reporting signal, so only the events genuinely representing real business value get flagged.

Step 6: plan the audiences and attribution view before launch, not after

Audience: "Signed up, never activated" (part 11) — for a real
  re-engagement campaign via Meta Ads retargeting
Attribution check planned: compare self-serve signups' cross-channel
  attribution (part 13) against the existing sales-assisted funnel's
  attribution, to see if self-serve draws from genuinely different
  channels or cannibalizes the same ones

Planning this in advance means the audience and the attribution comparison are ready to build the moment real data starts arriving, rather than being designed reactively after launch once someone asks for them.

Step 7: plan the consent and privacy handling explicitly

Confirm: new self_serve_activation event respects the same
  Consent Mode default-denied state (part 17) already governing every
  other event on the site — no new tag installation bypasses it

A measurement plan that doesn't explicitly confirm this is a real, easy gap — a new feature's tracking is sometimes wired up by whoever builds the feature, not the person who originally configured Consent Mode, creating a real risk of a new event firing without respecting the same consent state everything else honors.

Step 8: plan the verification step before launch, not after something looks wrong

Before real launch: run a full DebugView session (part 20) through
  the entire self-serve signup and activation flow, confirming every
  planned event fires with every planned parameter correctly populated

This closes the loop directly back to part 20 — verification isn't a step to reach for only once something already looks broken in reporting; it's a planned, standard step before any new tracking goes live for real users.

The complete measurement plan, assembled

1. Business questions (step 1)
2. Events mapped to questions (step 2)
3. Event and parameter names, defined upfront (step 3)
4. Custom dimensions/metrics to register (step 4)
5. Key events to mark (step 5)
6. Audiences and attribution views to build (step 6)
7. Consent handling confirmed (step 7)
8. Pre-launch DebugView verification (step 8)

This eight-step structure is the real, general template this series has been building toward the entire time — every part from 1 through 21 maps to a specific step here, and a genuinely well-planned GA4 implementation for any new product, campaign, or business runs through this same sequence before a single tag is installed, not after.

That's the complete Google Analytics series — from GA4's core event model through to a real, repeatable measurement plan. Combined with the SEO, Google Ads, Meta Ads, and Content Marketing series, Bright Leaf Coffee and GreenDesk now have a fully connected, end-to-end measurement story: traffic acquisition, ad platforms, content, and analytics, all tied back to the same real events and the same real businesses throughout.