Why it is a strategy document
It is tempting to treat measurement as a technical task to be done after the campaign is built. The order is wrong in two ways. First, tracking set up after launch cannot recover the baseline or the early results, so the campaign cannot be judged. Second, and more important, the questions a measurement plan answers are strategic: what the business counts as success, what evidence will be accepted, how credit is shared between channels. Those are decisions for the people who set the objectives, made before the people who build the tracking start. The objectives and KPIs guide produces the inputs; this plan turns each KPI into something that can actually be captured.
What the plan contains
The objective-to-event map
For every objective, the KPI, the event or events that make up the KPI, where each event happens and what qualifies it.
| Objective | KPI | Event(s) | Where it happens | Qualification |
|---|---|---|---|---|
| More qualified leads | Sales-accepted leads per month | Form submitted; lead marked accepted | Website form; CRM | Accepted by sales within the period |
| Higher first-purchase conversion | Purchase conversion rate (CVRShare of visits or clicks that convert.) | Purchase completed | Checkout | Excludes refunds within the return window |
| Reach the launch segment | Unique reach in segment | Impression | Ad platforms | Audience defined per platform |
The map exposes vague objectives immediately — an objective with no event cannot be measured — and it exposes events that cannot be captured with the current set-up, which is the tracking work that must be done before launch.
The source-of-truth table
Every number will be available from more than one tool, and the tools will disagree. The plan names the reference for each kind of number and says why the others differ.
| Number | Reference source | Also reported by | Expected differences |
|---|---|---|---|
| Spend, impressions, clicks | Ad platform | — | None; the platform is the only source |
| Sessions, on-site behaviour | Analytics (GA4Google's event-based analytics platform. or similar) | — | Clicks exceed sessions (bots, blocked tracking, abandoned loads) |
| Platform conversions | Ad platform | Analytics | Platform counts view-through and modelled conversions within its own window; analytics counts on-site events with its own attribution |
| Revenue, orders, customers | Order system or CRM | Analytics, platforms | Analytics misses refunds and offline; platforms double-count across channels |
The point is not to crown a winner but to stop the recurring meeting in which two reports are compared and neither is believed. When a report cites a number, the plan says which source it came from.
Naming conventions
Campaigns, ad sets, ads, content and links need names that a report can group by without a human translating. A conventionAn agreed pattern for naming campaigns, ads, events and tracking parameters so that reports can group them reliably. fixes the fields and their order — for example market, channel, objective, segment, campaign, creative, date — and the UTMURL tags identifying traffic source. parameters that carry them into analytics. The convention is boring and the single highest-return decision in the plan: without it, every analysis starts with a cleanup.
Attribution settings
Which model and which windows the reference reports use for crediting conversions to channels, and the acknowledgement that platform attribution and analytics attribution will differ. At the strategy level the plan decides the stance: for example, last-click in analytics for operational optimisation, platform-reported figures for in-platform bidding, and periodic incrementality tests for budget decisions on creation channels. The attribution page covers the models; the plan records the choice.
Reporting cadence and audience
Who sees what, how often, in what form. A weekly operational view for the people running channels (diagnostic metrics, by channel); a monthly management view (KPIs against targets, by objective); a quarterly strategy review (programme-level numbers, mix decisions). Each report draws on the same plan, so the numbers agree across levels.
Ownership and verification
Who owns the plan, who owns each tracking implementation, and the check that must pass before any campaign spends: a test event visible in each reference source, named correctly, attributed correctly. The campaign brief references this check rather than redefining it.
A measurement plan preventing a familiar argument
A retailer's paid-social agency reports a strong return for the quarter; the finance team's order data shows revenue flat. The measurement plan already states that platform-reported conversions include view-through within a seven-day window and are the reference for in-platform optimisation only, that the order system is the reference for revenue, and that creation channels are judged on reach and on incrementality tests rather than attributed revenue. The meeting is short: the platform figure is what it was always going to be, the revenue figure is the one the business reports, and the open question — whether the social spend is incremental — has a scheduled test rather than an argument.
A minimal plan for a small business
One page for a local service business
Objective: more booked consultations from the service area. KPI: consultations booked, reference source the booking system. Events: booking form submitted (website, tracked through the tag manager to analytics and to the ad platforms), phone call from the site's tracked number (call tracking), booking confirmed (booking system). Qualification: a booking counts when confirmed, not when the form is sent. Diagnostics: sessions by source, form starts, form completions, calls, cost per booking by channel. Source of truth: booking system for bookings; analytics for site behaviour; ad platforms for spend and clicks only. Naming: every campaign named market, channel, service, month; every link tagged with matching UTM values. Attribution: last-click in analytics for channel comparison; platform figures used only to optimise inside each platform; disagreements expected and explained in the monthly note. Reporting: a weekly check of bookings and cost by channel by the owner; a monthly summary with the three KPIs. Verification: a test booking from each channel before any spend, visible in all three systems with the right name. The whole plan fits on one page, and it answers every question the owner would otherwise ask the agency.
A business with no accounts yet can write this plan as one step of the tutorial on setting up the marketing basics for a new business, once ownership, analytics, consent and naming are in place.
Pre-launch verification checklist
Reporting layers
| Audience | Cadence | Contents | Source |
|---|---|---|---|
| Channel owners | Weekly | Diagnostic metrics by channel and campaign; pacing; tracking health | Platforms and analytics, per the source-of-truth table |
| Marketing lead and management | Monthly | KPIs against targets by objective; allocation changes; notable diagnostics | Reference sources only |
| Leadership and finance | Quarterly | Programme-level outcomes: customers, revenue, blended cost, brand trend; mix decisions | Order system or CRM; brand tracking; incrementality results |
Building it
- Start from the KPI tree in the objectives guide; list the KPIs and their diagnostic metrics.
- For each, write the event, the place and the qualification.
- Inventory the tools that exist (analytics, tag manager — GTMTag management system. or similar — ad platforms, CRM, order system, email platform) and assign each number its reference source.
- Write or adopt the naming convention and the UTM scheme.
- Record attribution settings and the stance on platform-versus-analytics figures.
- Define the reports: audience, cadence, contents, source for every number.
- Hand the plan to whoever builds the tracking; the analytics cluster covers that implementation, including server-side signals such as the Conversions APIServer-side event sharing with Meta. where browser tracking is incomplete.
- Run the verification check before the first campaign spends; repeat it after every site or tool change.
When tools change
Platforms change attribution defaults, consent rules reduce what browsers report, and sites get rebuilt. Each of these breaks a measurement plan silently: numbers keep arriving and mean something different. The plan therefore records the date and version of each tool's settings, and the verification check is re-run after any platform notice, consent change or site release. A visible break in a trend with no change in activity is treated as a measurement question first and a performance question second; most sudden drops in reported conversions turn out to be a tag that stopped firing or a window that changed.
Common mistakes
- Measurement after launch. No baseline, no early data, a campaign that cannot be judged.
- No source of truthThe system whose figure is treated as authoritative for a given number, such as billing for revenue or the CRM for pipeline, with other tools reconciled to it.. Every report is contested and the loudest number wins.
- No naming convention. Analysis becomes archaeology.
- Platform figures as revenue. View-through and modelled conversionsConversions a platform estimates statistically for activity it could not observe directly, such as unconsented or cross-device journeys. presented as sales.
- One attribution model for every decision. Last-click for budget decisions starves demand creation.
- Reports with no owner. Dashboards that nobody checks and nobody trusts.
The core metrics guide explains the numbers the plan captures; the analytics hub covers the tools and the tracking plan that implements this one.
