Measurement
Part of Programmatic advertising measurement without vanity metrics
Decisions to settle before building a programmatic advertising reporting dashboard
A pre-build guide to a programmatic advertising reporting dashboard covering decisions, definitions, sources, attribution windows, first tiles and a pound example.
What to take away
- Most teams start with the chart and only later ask which decision it changes, so the dashboard is ignored after a fortnight.
- Build it from five questions the buying team already argues about, then map each to one metric, one source and one owner.
- Agree the reporting period and attribution window before anyone opens a visualisation tool.
- Record the source of every number, and name one person responsible for each tile.
Start from the decision, not the chart
A programmatic advertising reporting dashboard is a decision tool, not a wall display. Before choosing software, list the decisions your team makes each month: shifting budget between supply paths, pausing a creative, renegotiating a fee. Each decision needs one metric that would change it.
Decision-first dashboard test
Does this metric change a decision?
keep the metric
leave it out
If a metric cannot change a decision, leave it out. Teams that skip this step build dashboards full of impressions nobody opens twice.
Agree definitions before you build
Write a one-page dictionary covering spend, impressions, viewable impressions, conversions and cost per acquisition. State whether spend is gross or net of fees. State the currency: pounds sterling or platform currency.
One-page definitions dictionary
- Spendgross or net of fees
- Currencypounds sterling or platform currency
- Impressions definition
- Viewable impressions definition
- Conversions definition
- Cost per acquisition definition
Platform numbers rarely match. A demand-side platform may count a conversion on click, while an analytics tool counts a session on a later visit. The measurement and reporting pillar sets out how these definitions fit together across a client account.
Choose sources you can defend
Pull delivery data from the buying platform and outcome data from the client's own analytics. Google Analytics 4 documents how its event-based model records sessions and conversions, which helps when you explain why its totals differ from platform-reported figures.
Where lead generation is the objective, platform conversion tracking is part of the chain, and LinkedIn lead generation ads show how a form submission becomes a measurable lead event rather than a click.
Supply quality belongs in the same document. The IAB UK Gold Standard letter covers supply chain transparency, so you can note which parts of the chain the dashboard can and cannot evidence.
Map the attribution window
The attribution window decides which conversions count. A seven-day click window credits fewer conversions than a 30-day window that includes views. Neither is wrong, but mixing them across campaigns makes comparison meaningless.
Record the window on the dashboard itself, next to the conversion tile. If your client uses more than one model, the attribution methods guide explains how click, view and data-driven models differ before you commit to one in the build.
Build the tiles and name an owner
Which tiles belong on the first version?
Start with five: spend to date against plan, impressions and viewable rate, conversions and cost per acquisition, pacing by week, and one quality signal such as invalid traffic rate. Add tiles only when a decision needs them.
How often should the dashboard refresh?
Match refresh to the buying cycle. Daily data suits pacing checks during an active flight; weekly or monthly data suits fee and supply reviews. Timestamp every refresh so users know how old the number is.
Worked example with pound figures
An agency runs a lead generation campaign with a monthly media budget of £20,000, plus a £2,000 management fee. The dashboard shows £15,400 spent by day 21, which is 77 per cent of media budget with 30 per cent of the month left. That pacing signal prompts a review of bid caps.
The same month records 220 leads, giving a cost per lead of £70 on media spend, or £80 once the fee is included. Both figures are illustrative examples, not benchmarks. The client sees the media and all-in figures side by side, labelled.
Common questions
What is the minimum viable programmatic advertising reporting dashboard?
Five tiles covering spend, delivery, conversions, cost per acquisition and pacing, each with a named source and owner. Anything more should wait until a decision requires it.
Should the dashboard show platform numbers or analytics numbers?
Both, clearly labelled. Platform numbers explain buying performance; analytics numbers explain what happened on the client's site. Where they disagree, show the gap rather than hiding it.
How do we stop the dashboard going stale?
Assign one owner per tile and set a review date for the whole build. If nobody has opened a tile in a month, remove it or fix the source feeding it.
Does a dashboard replace written reporting?
No. The dashboard carries the numbers; a short written note explains what changed and what happens next. Clients remember the explanation, not the chart.
