All advertising has a goal, which is to make the consumer aware of a product or service and encourage them to buy. In this spirit, more marketers are focusing on attribution measurement and trying to learn how every dollar they spend on media influences the consumer outcome and how different elements contribute along the consumer path to purchase.
Without an accurate understanding of how each piece influences the consumer’s decisions, marketers are essentially throwing part of their media budget into the trash. If you can’t understand which channels drove behaviors, you can’t make informed decisions on how to allocate budget to optimize those channels.
This is one of the biggest challenges facing marketers, and solving it requires a two-step process. Step one is understanding the consumer purchase path across devices and time, accounting for all the media touch points, whether they were paid, organic, earned, or owned. The second step is taking that path and figuring out how to distribute credit for the end result, whether that was a lead, sale, profit, or another metric.
That may sound easy on the surface, and there are countless vendors that have promised that marketing attribution can be solved right now. That’s sadly not the case.
Marketing attribution isn’t easy because the consumer path can get quite lengthy, especially with multiple consumer IDs and devices used to interact with media. And what if a consumer conducts all of their research online but makes the purchase offline? The reality is that online media drives offline sales, and vice versa. There is a similar divide in desktop and mobile, where mobile ads may drive desktop purchases. If mobile ads can never be tied to a purchase, they look much less effective than they actually are. Factor in ad blocking, viewability, and other digital media issues, and it gets harder and harder to paint a clear picture.
Tying online campaigns to sales
At Datalicious, we follow these steps to clean, anonymize, and match personally identifiable information (PII) to cookies and IP addresses, effectively matching sales and campaign activity across channels:
- Match sales to people and households using CRM or POS data.
- Remove PII and anonymize data.
- Match people to digital identities using deterministic user (cookie) ID syncing.
- Match any leftover, thus far unmatched, people using probabilistic IP address matching.
- Build the purchase path across channels and platforms including sales as the final touch point.
- Develop and apply attribution modeling to determine which touch points were the most powerful in determining the final sale.
- Aggregate data at channel/strategy/campaign/creative/etc. level for reporting.
How can this method be applied to real campaigns? Let’s look at an example from the automotive industry. The process starts with sale data, shared by the dealer or gathered from third-party registration sources. This data is then anonymized and used to match an auto sale to a household. Deterministic identity matching then helps tie that purchase to online activity. Once that approach runs its course, we move on to a probabilistic approach to make a match. At that point, we have a sense of the various different people in a household and what kind of research they’ve done. When this research activity is stitched together with advertiser log files, it’s possible to see which touch point was the most important in driving the final sale.
Understanding three billion destinations on the consumer journey
Attribution can also be applied to historic data to provide marketers with insights into future campaigns. The telco company, Telstra, wanted to understand how the approximately three billion user touch points it processed per financial year were driving revenue. There were more than two million different paths to purchase in fiscal year 2014, and 2.5 million in 2015, spanning 29 products and attributing sales to 7 different digital media channels.
To provide a full-path analysis, we reprocessed more than 24 months of conversions. A weighted attribution model accurately captured purchase path data and provided an accurate view of digital marketing’s performance in the media mix. This led to 25% growth in digital investment in 2014, which led to a 65% growth in sales orders. The following year, gross revenue grew by 87%, thanks to the media optimization insights.
While marketing attribution is an incredibly technical process, and one that will continue to change as the data market evolves, it can make a big difference. No single partner has all of the raw data needed for 100% accuracy. What marketers need right now, and going forward, is a partner that will continue to evolve its attribution measurement alongside all of the movement in the media landscape. At the end of the day, the best attribution results will come from the marketer and solution provider working together, rather than the marketer simply accepting the results at face value.