Which programmatic advertising KPIs fit each campaign objective?
Part of Programmatic media planning

The short answer
Programmatic advertising KPIs should follow the campaign objective. Awareness plans track unique reach, frequency and viewable delivery; consideration plans add survey lift or engaged visits; action plans track cost per conversion on a defined primary action. Each cost KPI needs quality guardrails beside it, such as invalid traffic and delivery on approved sites.
Key takeaways
- Pick the objective first; Display & Video 360 asks for a goal before it recommends a KPI.
- Reach is counted in people and frequency in exposures per person, so impression totals alone do not show either.
- Google Ads uses primary conversion actions for bidding and keeps secondary actions for observation only.
- A cost KPI is only comparable when audience, inventory and quality requirements are the same on both sides.
- One primary KPI per objective, a few diagnostics and fixed quality guardrails keep a plan readable.
How should a planner choose KPIs for a programmatic campaign?
Start with the business goal, then pick the measure that shows progress toward it. Google's Display & Video 360 campaign setup (opens in a new tab) follows that order: a campaign takes a goal (raise awareness, drive online action or visits, drive offline or in store sales, or drive app installs or engagements) and then shows recommended KPIs such as CPA, CPC, CPM, CPV, CTR and Viewable %. Google adds that "Goals work best when targets are realistic and closely reflect your objectives."
Definition
- Programmatic KPI
- A key performance indicator chosen before launch to judge whether a programmatic campaign met its objective, stated with its definition, data source and target.
Which KPIs fit a programmatic awareness campaign?
An awareness plan aims to put the message in front of the right people with enough repetition to be remembered. That calls for measures of people, not just impressions. Google Ads describes unique reach and frequency metrics (opens in a new tab) as measuring "the total number of people who were shown an ad and the frequency at which they were shown an ad," modeled across devices and, on connected TV, co-viewing. The page lists Unique users, average impression frequency per user over 7 and 30 days, and a frequency distribution from 1+ to 10+.
Exposure quality matters as much as count for awareness. A viewable impression under the MRC viewable impression guideline (opens in a new tab) requires 50 percent of the ad's pixels on screen for at least one continuous second for display. Where survey measurement is available, Google's Brand Lift (opens in a new tab) compares people who saw the ads with people who were eligible but did not, on metrics such as ad recall and awareness. Google notes that Brand Lift "isn't available for all Google Ads accounts."
| Objective | Primary KPI | Diagnostics | Documented source |
|---|---|---|---|
| Awareness | Unique reach in the target market at a set frequency | Frequency distribution, viewable rate, ad recall lift | Google Ads reach and frequency metrics; MRC viewability guideline; Brand Lift |
| Consideration | Survey lift on consideration, or qualified site visits | Engaged visit rate, video completion, CTR as a signal only | Google Ads Brand Lift; the advertiser's site analytics |
| Action | Cost per primary conversion action | Conversion rate, secondary actions observed, path to conversion | Google Ads primary and secondary conversion actions |
| Any objective | None: these are guardrails | Invalid traffic rate, share of delivery on approved domains | MRC invalid traffic standard; served placement reports |
Which KPIs fit consideration and action goals?
Consideration sits between knowing a brand and choosing it. Brand Lift can measure it directly with a consideration question, and Google says Brand Lift looks at goals like ad recall, brand awareness and consideration "rather than traditional metrics such as clicks, impressions, or views." Without a survey, the closest behavioral measure is a qualified visit: a session on the advertiser's site that reaches a product or comparison page. Click through rate can show whether creative is being noticed, but it is a diagnostic, not proof of interest.
Action goals need a precise definition of the action. Google Ads groups conversion actions into goals and separates two kinds in its help on primary and secondary conversion actions (opens in a new tab). Primary actions are reported in the Conversions column and used for bidding when their goal is used for bidding. Secondary actions "are for observation only," reported in All conversions but not used for bidding, unless they sit in a custom goal. That split is a useful model for any plan: one action the campaign is judged on, others watched for context.
Why is a lower CPM not proof of a better campaign?
A cost per thousand impressions describes price, not value. The figure falls whenever the plan accepts cheaper inventory, and cheaper inventory can be cheaper for reasons that work against the objective. Google's Display & Video 360 planning help (opens in a new tab) states the trade directly: "If you add more specific targeting, such as viewability or audience list targeting, the average cost will usually increase." A plan that removes those requirements will show a lower CPM and buy something different.
- Different inventory. A lower CPM may come from placements with lower viewable rates, so cost per viewable impression can rise while CPM falls.
- Invalid traffic. The MRC invalid traffic addendum (opens in a new tab) of June 2020 separates general invalid traffic, caught by lists and standard checks, from sophisticated invalid traffic that needs advanced analytics or human review to find. Cheap impressions that are later filtered are not cheap.
- Off list delivery. Spend that lands outside the approved site list may be priced lower and serve none of the brief's context.
- Wrong denominator. Impressions do not equal people, so a lower CPM says nothing about cost per person reached.
The fix is to compare cost only on matching terms: the same market, audience definition, inventory rules and quality floors. The CPM guide covers the arithmetic and the difference between CPM, vCPM and eCPM.
How should quality KPIs sit beside cost KPIs?
Treat quality measures as guardrails with fixed floors, and cost measures as targets inside them. A cost target can be traded for scale; a guardrail is not traded. When a guardrail is breached, the campaign is reviewed before anyone looks at whether cost improved. Write each guardrail with its source, because two measurement providers can report different figures for the same delivery.
| Guardrail | What it protects | Where the figure comes from |
|---|---|---|
| Share of delivery on approved domains | That spend reached the planned context | Served placement report reconciled with the approved list |
| Invalid traffic rate, general and sophisticated | That impressions came from real visits | The measurement provider's filtration report |
| Viewable rate and measurable rate | That ads had a chance to be seen, and how much delivery was measured at all | The viewability measurement report |
| Suitability incidents | That ads avoided the client's excluded content | Post campaign placement review |
| Reach in the target market | That delivery reached people in the brief's country or region | The platform's reach report, with its geography setting |
Delivery on approved domains needs a reconciliation step, covered in the guide to served domain reconciliation. For reach and frequency planning arithmetic, see reach and frequency.
How many KPIs should one plan track?
No standard sets a number, so the useful rule is structural rather than numeric: one primary KPI per objective, a short set of diagnostics that explain it, and the quality guardrails the plan applies across campaigns. Google's split between primary actions used for bidding and secondary actions for observation is the same idea applied to conversions. A plan that names eight equal KPIs gives the team no way to decide when two of them move in opposite directions.
KPI section of a programmatic plan
- One objective is named, and the primary KPI follows from it.
- The primary KPI has a written definition, a data source and a target.
- Each diagnostic says which question about the primary KPI it answers.
- Quality guardrails are listed with floors and their measurement source.
- Cost KPIs are stated on matching terms: market, audience, inventory rules.
- The conversion or survey window is fixed before launch.
- The plan says what happens when a guardrail is breached and who decides.
How does PlacementGraph vet the site list before launch?
What does a KPI plan look like for one advertiser?
Fictional example
A replacement window company in the Dakotas
An unnamed replacement window and door company serving Fargo, Sioux Falls and Bismarck plans a fall display campaign on home improvement, regional news and local lifestyle sites, followed by a second phase aimed at quote requests. The client's first draft lists CPM, CTR, impressions, clicks, quote requests and cost per quote as six equal KPIs.
The agency splits the campaign into two objectives. The awareness phase takes unique reach in the three metro areas at a set weekly frequency as its primary KPI, with viewable rate and frequency distribution as diagnostics. The action phase takes cost per completed quote request as its primary KPI, defined as a submitted form, with form starts and calls tracked as secondary actions for observation only.
Both phases share the same guardrails: invalid traffic rate, share of delivery on the approved site list and suitability incidents, each with its source. CPM stays in the report as a cost diagnostic, compared only with earlier campaigns that used the same inventory rules.
Frequently asked questions
Is click through rate a good KPI for display awareness?
Treat it as a diagnostic, not a goal. A click shows that someone acted on the ad, which awareness campaigns do not set out to cause, and clicks can include invalid activity. Reach at a set frequency, viewable delivery and, where available, survey based ad recall describe awareness more directly.
Should KPI targets change during a campaign?
Keep the definitions fixed and change targets only for a stated reason, such as a budget change, and record when and why. Google also warns that Smart Bidding needs a 7 to 14 day learning phase and that changing budgets, targets or conversion goals during it resets the learning window.
Can one KPI cover websites and YouTube in the same plan?
The objective can be shared, but the measures differ by surface. Google Ads reports reach and frequency for Display, Demand Gen, Video and App campaigns, while view based measures apply only to video. Set a primary KPI for the shared objective, then state which diagnostics apply to each surface.
Who should own the KPI definitions in an agency and client relationship?
The client owns the objective and the definition of a conversion, because both come from its business. The agency owns how each KPI is measured and reported and flags when a definition cannot be measured as written. Both sign the KPI section of the plan before launch.
Sources
- Create a campaign (opens in a new tab)Google, Display & Video 360 Help
- Create a plan for a campaign (opens in a new tab)Google, Display & Video 360 Help
- Measuring reach and frequency (opens in a new tab)Google Ads Help
- About Brand Lift (opens in a new tab)Google Ads Help
- About primary and secondary conversion actions (opens in a new tab)Google Ads Help
- Invalid Traffic Detection and Filtration Standards Addendum, update (opens in a new tab)Media Rating Council, June 2020
- MRC Viewable Ad Impression Measurement Guidelines, version 2.0 (opens in a new tab)Media Rating Council, August 18, 2015
Researched and written by PlacementGraph Research. Read how we research, source and update these pages. Corrections: diana@placementgraph.com.