The Value of Fractional Attribution: Don't throw away the majority of your data

Our CTO Erik Thorson contributed his perspective on Fractional Attribution and the impact of using all available data.

The ideal of advertising is to know which ads drove a customer to make a purchase. Before programmatic advertising, the only way to accomplish this measurement, was to dedicate a phone number, URL, coupon code to an advertisement and see how many redemptions came through. This method is effective, but not massively scalable.

The rise of Programmatic buying and tracking brought this capability to a massive scale where there were millions events a day, each tailored to their user and each with their own set of attributes. Suddenly there was a feedback loop and inventory could be optimized in real time. The advertising ideal became reality.

While this innovation helped advertisers improve their tracking, it also opened doors for bad actors. Made for advertising sites started to arise and pull ad dollars out of the system, privacy concerns increasingly restricted the display ecosystem, and advertisers are now relying more and more on CTV as a trusted channel. The challenge now became: How do we apply the measurability of programmatic to the less identifiable world of TV?

Navigating Shared Devices

Televisions are shared devices and don’t offer the trackable 1:1 relationship the cookie provided. The most granular we can get with TV is household level mapping. If a person from an ad-exposed household goes to a web site and buys something after seeing the CTV ad, we can call map the conversion to a household level ad exposure. This conversion mapping is what enables performance CTV.

This measurability allows dollars spent to be mapped directly to dollars earned. The advertising ideal is realized on CTV!

The reality, however, is that mapping one exposure on CTV to a conversion does not fully show the story of how someone came to convert. In order to drive conversions, advertising campaigns take time, strategic placement, and consistency. A single exposure is not going to leave a lasting impression. The cumulative impact from a series of exposures is what drives someone to make a purchase, and the collective data from all exposures is more valuable and telling than any single one of those exposures on their own. Each of them should be given value when planning for or reporting results from a campaign.

Evolving from Legacy Tracking

Last touch, first touch or anything other than fractional attribution simply reduces the data gathered for a campaign to a small fraction of what it could be. The data available through a CTV buy includes time of day, channel, show, geo data, etc. These data points give buyers information on what inventory to buy to drive results and increase the probability of getting a conversion. The more data consumed, the more likely a buyer will be at driving results.

Lets say that it takes an average of 5 impressions before a user converts. If you’re using last touch conversion tracking, 80% of the data you could be using would be dropped from your optimization tactics. That’s huge! Losing 80% of the available data can mean the difference between driving scalable, repeatable results and missing the mark entirely.

Smart buyers need to leverage as much available data as possible to properly identify the drivers of success.

What does a fractional conversion report look like?

It’s not as clean as last touch. A fractional conversion model assigns a fraction of a conversion to each impression.

For example, take a series of ad impressions on the following publishers:

1. HGTV
2. HGTV
3. CNN
4. CBS Sports

If you are using fractional conversion tracking and all of these ads take place before the conversion (a site visit) happens, HGTV gets 50% of the conversion while the others get 25%.

This attribution gives you a proper view of the inventory that has the most relevance to your customer base. You should prioritize HGTV. In the case of last touch, all of the attribution goes to CBS Sports while giving no credit to the others when the majority of the conversion should really be credited to HGTV.

If you followed the guidance of Last Touch Attribution, you’d spend all of your money on CBS Sports. You’d be over-investing in the wrong publisher.

In advertising we deal in probabilities and throwing out 80% of the data hampers a buyer’s likelihood to drive performance. Last touch is an incomplete and ineffective way to scale results. We have the technology and the data to make educated decisions on how to buy ads and buyers should use tactics and platforms that enable them to drive success at scale over time.

Let's choose to be better advertisers by using all of the tools that we have at our disposal rather than simply taking the easy road at the expense of our clients.