Every $ Accountable: The Dos and Don’ts of Programmatic KPI Setting
According to eMarketer, US programmatic display ad spend will increase 16% in 2022, and for the first time ever, video will comprise over half. By 2023, US advertisers spend on programmatic video will exceed that of linear TV.
But as media is increasingly bought through programmatic channels, it becomes harder to track and measure. With mass media, where set placements are bought, it’s much easier to reliably track effectiveness and efficiency. You can monitor whether an ad appeared, and how many times, to get an indication of reach and impact. And you measure the cost per reach point and optimize downwards over time.
In programmatic, you’re buying audiences rather than placements. With millions of potential websites, it is very hard to accurately track where an ad showed up, or the volume of impressions delivered across platforms. Reach is therefore estimated and is often under-reported. Volume-based efficiency measures such as cost-per-thousand (CPM) have been heavily relied upon, ignoring and often hindering effectiveness.
At the same time, the transparency challenges in programmatic mask the extent of this losing game – a laser focus on efficiency increases the risk of being exposed to undisclosed fees and poor-quality inventory.
Here are eight dos and don’ts designed to help advertisers to set suitable programmatic KPIs, ensuring investments are efficient and effective:
DO: Track and report real-time metrics.
The primary lever for programmatic is data, which means that there are numerous metrics that can be tracked in real-time. These may not provide a firm measure of the efficiency and effectiveness of a campaign, but they are indicative building blocks and are important for transparency.
At a basic level, make sure you have transparent investment data. Know who and what you are buying, and how it is performing. Ensure all placements are disclosed and reported on. Track key cost metrics over time (CPM, CPCV etc.) and analyze fluctuations to understand the cause as well as the impact on media quality. Monitor exposure metrics over time, such as unique reach, to give an indication of the impact on awareness. Cross-platform reach is difficult to measure given walled garden restrictions and data fragmentation, but unique reach is still perhaps the strongest indication of effectiveness within channels.
DO: Factor in quality.
Optimizing based purely on efficiency can do more harm than good. Quality metrics have to be incorporated. Know what quality tradeoffs you are prepared to make for the sake of efficiency before you invest by setting brand suitability guidelines. Set accountable thresholds for viewability (and attention), brand safety and ad fraud based on campaign goals. For instance, one might accept lower viewability or higher invalid traffic (IVT) for a performance campaign linked to outcomes than for a brand campaign where factors such as viewability have a much greater impact on objectives such as brand awareness. Go beyond viewability where possible to measure attention, leveraging the capabilities of both ad verification technologies and buying platforms that offer attention-based metrics. An ad needs to be seen and noticed to drive awareness or action.
Wherever possible, link quality, cost and exposure measures together. For example, track viewable cost-per-thousand (vCPM) rather than just CPM, or track cost-per-completed-view alongside completion rate. Use proxy measures to understand the ratio of cost to quality and volume, and track this over time.
DON’T: Rely only on campaign averages.
All investment should be held accountable, not just campaign averages. Make sure KPIs are reported by campaign down to the most granular level. Campaign averages are very useful to track what is happening over time but can mask the true performance. For instance, a campaign may meet the viewability threshold overall, however one element may have delivered a viewability level higher than the target, whilst other elements fell below.
DO: Link back to business outcomes.
Measure the impact on the bottom line. Performance campaigns should have a direct outcome measure versus investment (cost-per-sale /return-on-investment /Return-on-ad-spend) that supports real business objectives. Brand campaigns are harder to connect with outcomes but incrementally or lift (brand lift, conversion lift, ROAS lift, or exposed ROAS) should be measured where possible. Where the campaign is engagement focused, metrics such as click-through-rate (CTR), site visits, time on site, video completion rates, are useful and can give good indications of brand health. Remember to consider the media environment and consumer behavior in channel selection and when assigning KPIs. A user is less likely to click through from a display ad than from a search ad, for example. All metrics should ladder back to outcomes in an overall measurement framework.
Conversions will be harder to measure post-cookie demise, with view-through measurement impossible. Testing and leveraging existing data to establish proxies is now critical, as is exploring intra-platform and consumer identity solutions to track these important outcomes. Now is the time to establish bespoke advertiser benchmarks and model future measurement, utilizing combinations of incrementality measurement, in-platform attribution, data clean rooms and media mix modeling (MMM).
DO: Focus on the long term.
Shift from short-term measures like ROI / ROAS to more long-term customer lifetime-value (LTV). The long-term health of customer relationships will be increasingly important as we lose the power of third-party cookies. Whilst operationally challenging, advertisers need to leverage their own first-party data, create internal data lakes and stitch data together to have a better understanding of LTV, enabling more long-term, value-based decision-making.
DON’T: Set buy level price commitments.
The majority of programmatic buys are not negotiated upfront; prices are determined during the auction for each impression. Countless variables influence impression price (audience, geo, device, format, browser, location, KPI etc.) making year-on-year comparisons unsuitable and inaccurate. So don’t set traditional price commitments at an individual buy level, it limits flexibility and has a detrimental impact on media quality. It can also result in agencies making decisions based on commitments rather than what is strategically right for the brand.
Instead, ensure programmatic media is held accountable with thresholds based on MMM. Maximum performance-linked cost thresholds (CPA, CPCV etc.) can be set at a campaign or objective level based on the value of the touchpoints. This enables buy level flexibility whilst maintaining safeguards. But it is important to recognize the limitations (of thresholds and MMM) and have a process to enable exceptions for the right strategic opportunities.
DO: Focus on the consumer.
Increased data regulation has come in part due to relentless retargeting of consumers and a lack of consideration for consumer experience and privacy. Advertisers (as well as the wider industry) must do things better and put the consumer first.
Advertisers need to better understand the quality and accuracy of data used for targeting, as inaccurate targeting can harm consumer relationships and erode brand image. Look to set data quality standards. This is especially important as advertisers expand and access first-party data for targeting, ready for a future without third-party cookies. Model on-target reach and effective CPM of different data sets to understand data value – it is worth investing in quality data that is compliant and reaches desired audiences with less waste through direct partnerships with publishers or with supply side platforms.
Frequency also needs careful consideration and management. With the upcoming loss of cross-platform frequency capping (due to cookie depreciation), advertisers will need to be careful not to wastefully over-expose consumers to their ads.
DO: Analyze, test and evolve.
Analytical capabilities are key. Hold programmatic investment accountable with an ongoing governance and analysis program. This requires deep platform knowledge, as analysis needs to go beyond just the core KPIs. Advertisers should ensure the vast amount of data available in programmatic platforms is being used to reduce waste and maximize value.
Ensure that KPIs are suitable to the individual programmatic buying platform. Every platform and algorithm are different (and constantly evolving) so developing knowledge of platform best practices and optimization tactics is important in setting the right KPIs. This may require building internal capabilities or leveraging external support.
Furthermore, the validity of KPIs should be continuously assessed. This is more important now than ever. Advertisers need a rigorous testing roadmap for the next 12 months to establish their measurement model and the right programmatic KPIs for a cookie-less tomorrow.
In summary, programmatic KPIs need to be bespoke to an advertiser’s business, with a clear accountability framework linked to the business’s bottom line. Programmatic continues to evolve at break-neck speed and so must measurement. With the imminent demise of third-party cookies, now more than ever it is crucial for advertisers to assess the suitability of their programmatic KPIs and evolve measurement practices.
Advertisers must be analytical and ensure robust ongoing governance. It starts with an audit. Assess whether KPIs are fit for purpose now and in the fast-approaching cookieless world. Then take a test-and-learn approach to establish and evolve programmatic measurement on an ongoing basis.
For further perspectives and a lively discussion around KPI setting from experts in the industry, check out our panel discussion: Every $ Maximized: Holding Programmatic Investment Accountable.