Measuring Programmatic Performance: KPIs and Metrics
Programmatic campaigns generate a large volume of data, which is both an advantage and a challenge. Having access to hundreds of metrics doesn’t automatically translate into understanding campaign performance. A clear framework for which KPIs matter, and why, makes it much easier to optimize campaigns and communicate results.
The measurement hierarchy
It helps to think about programmatic metrics in three layers:
- Delivery metrics: did the ads actually reach people in a meaningful way?
- Engagement metrics: did people respond to the ads?
- Outcome metrics: did the campaign drive the business results it was meant to?
Each layer depends on the one below it. Impressive click-through rates don’t matter if the underlying impressions were never viewed. And strong engagement metrics are only valuable if they connect to actual business outcomes.
Delivery metrics
Impressions
The basic unit of programmatic measurement. One impression equals one ad being served. Raw impression volume is a starting point, but on its own says little about quality.
Viewability
An impression is considered viewable when at least 50% of the ad’s pixels are in view for at least one second (two seconds for video), according to the MRC standard. Viewability rates vary considerably by placement, device, and ad format. A campaign with high impression volume but low viewability is reaching far fewer people than the numbers suggest.
Invalid traffic (IVT)
Not all impressions come from real humans. Invalid traffic includes bot traffic, accidental clicks, and other non-human activity. Reputable DSPs filter for IVT, but monitoring IVT rates, especially on new supply sources, is important for maintaining measurement accuracy.
Win rate
The percentage of auctions entered that result in a won impression. A very low win rate can indicate that bids are too conservative, that floor prices are higher than expected, or that competition for specific inventory is intense.
Reach and frequency
Reach measures the number of unique users exposed to a campaign; frequency measures how many times each user saw an ad on average. Managing frequency is especially important for brand campaigns. Too few exposures may underdeliver the message, while too many can cause fatigue and waste budget.
Engagement metrics
Click-through rate (CTR)
CTR measures the percentage of impressions that resulted in a click. For direct response campaigns, CTR is a useful signal of creative relevance. However, CTR varies significantly by format and placement. Display CTRs are typically much lower than search, which is expected and not a sign of poor performance.
Video completion rate (VCR)
For video campaigns, completion rate (the percentage of users who watched the full video) is a better measure of engagement than impressions alone. Quartile metrics, tracking completions at 25%, 50%, 75%, and 100%, give a more detailed view of where viewers drop off.
Engagement rate
For rich media or interactive formats, engagement rate captures interactions beyond clicks: expansions, video plays, swipes, and other format-specific actions.
Outcome metrics
Cost per acquisition (CPA)
CPA measures the cost to drive a defined conversion event: a purchase, a form fill, a store visit, an app install. It is one of the most direct measures of campaign efficiency for performance campaigns.
Return on ad spend (ROAS)
ROAS divides revenue attributed to a campaign by the amount spent on it. A ROAS of 4x means the campaign generated four times its cost in revenue. Calculating ROAS accurately requires reliable attribution, which is increasingly complex in a multi-touch world.
Cost per visit / store visit attribution
For retail advertisers, driving foot traffic is often the primary goal. Store visit measurement, which estimates physical visits attributed to ad exposure, has become an important metric for local campaigns. Methodologies vary by provider, and numbers should be interpreted with appropriate context.
Brand lift
For awareness campaigns, brand lift studies measure changes in metrics like ad recall, brand awareness, or purchase intent among exposed versus unexposed audiences. These studies require a controlled test design and sufficient sample sizes, but they provide evidence of impact that click-based metrics cannot capture.
Attribution and the challenge of multi-touch measurement
Determining which ads drove which outcomes is one of the hardest problems in digital measurement. Last-click attribution, giving 100% of credit to the final ad a user clicked before converting, is simple but often misleading, undervaluing upper-funnel touchpoints.
Multi-touch attribution models distribute credit across multiple interactions. Data-driven attribution uses statistical modeling to estimate each touchpoint’s contribution. None of these models is perfect, and they can produce different conclusions from the same data.
Advertisers should use attribution as a directional guide rather than a precise accounting. Consistency in methodology over time matters as much as the specific model chosen.
How Arago helps
Arago surfaces the metrics that matter most for local retail campaigns, from viewability and frequency to store visit attribution, in a unified dashboard. By connecting delivery data to real business outcomes, Arago helps local retailers understand what’s driving results and where to adjust spend across channels.