People saw an ad, then bought the product advertised. Did the exposure cause the sale? It’s hard to tell. So do you need to invest in building out sophisticated attribution models in order to optimize your marketing? Not necessarily.
The goal for any marketing executive is to increase ROI, which you can achieve in two different ways: increasing the impact of your ads or decreasing the amount you’re spending on a campaign. Eliminating waste decreases that spend.
Controlling and measuring what doesn’t have an impact might be a faster route to ROI than pinpointing what did work with attribution models. The reason is pragmatic: It’s hard to tell what worked but relatively easy to tell what didn’t. This is especially true online because impressions that don’t contribute value are often a huge portion of the total.
Time and aggregation effects make impact data difficult to act on while a campaign is mid-flight, particularly when optimizing to offline spending. Intelligence often arrives late, or may not be actionable because results are aggregated differently than the buying decisions.
On the other hand, waste-oriented measures can generally feed back to programmatic buying optimizations in real time. If not, they can be used to adjust subsequent campaign parameters, such as targeting whitelists.
A Model For Understanding Advertising Waste
Advertising waste results from many factors. For example, consumers may never have an opportunity to see an ad because of its placement, or the ad may fail to load properly. Non-human traffic, targeting the wrong geographic regions, over-serving impressions, and poorly constructed audiences can create waste as well.
Marketers can estimate how many quality impressions they’ll serve on a given channel by understanding how likely these waste culprits are to occur. Armed with a realistic assumption of quality impression volume, marketers can calculate the effective cost of a media investment, as measured by an effective CPM, rather than a nominal CPM.
In the model we built with Ted McConnell, our partner and consultant, we preloaded pricing based on reasonable averages from a cross-section of national brand advertisers. We evaluated five categories: broadcast TV, ad networks/ecosystem video, network-guaranteed display, normal exchange-traded display, and Facebook in-feed video (see our model below).
Regardless of our inputs of cost and probability to deliver quality impressions, the beauty of the model is that any marketer can populate it with her own assumptions and historical data. This leads to greater truth about the quality of your impressions, which then leads to more accurate analysis when it comes to understanding impact and attribution of your investments.
When applying real-life assumptions around the deterrents of impression quality, it becomes clear that marketers can make a huge impact on their ROI by first managing their investments to reduce waste.
From the model and testing a range of scenarios, we pose five key mandates to reduce waste and improve ROI:
1. Advertisers should manage toward attention quality and outcomes by implementing any and all measures that reduce waste and create further insight.
2. Advertisers should consider managing frequency by limiting publisher or ad network diversity. The yield from having all decisions made at fewer points is likely to be much higher than the yield caused by supplier competition.
3. Advertisers should be certain that attribution models do not incorporate meaningless impressions. They are noise, not signal.
4. Advertisers should put their actual data into the provided model and use the outcome in their planning process.
5. Advertisers should not lose sight of factors beyond quality of audience, attention and reach. Characteristics like contextual fit still matter a lot.
To encourage feedback and discussion from the marketing community and drive discussion, here’s a download of our Effective CPM Calculator.
Max Kalehoff is CMO of SocialCode, a technology and insights company that manages digital advertising for consumer brands.