Attribution. It’s one of the most important topics of 2018; many advertisers are realizing how critical attribution is when it comes to evaluating multi-channel performance. It’s a bird’s-eye view of how online advertising—how each touch point—impacts customers and drives revenue for businesses.
Why Last-Click Won’t Cut It Anymore
The most common model used in the past was “last-click,” which gives credit to the last interaction a user had with the advertiser—for instance, the user searched an item of clothing, clicked an organic result, and then bought that item. It’s simple, and last-click data is usually free. But, you get what you pay for, which is to say an incomplete picture.
Let’s say the user took these steps:
- Searched for running shoes on Google, clicked a shopping ad to the advertiser’s site, then got distracted and abandoned cart
- Scrolled through Facebook and was served a retargeting ad but didn’t click
- Browsed a favorite blog and was served a remarketing ad but again, didn’t click
- Searched for the brand, clicked an organic search result, and finally purchased
In this situation, who gets the credit? Arguably, if steps 2 and 3 didn’t happen, the user might still have purchased running shoes, but they could’ve gone to a competitor. And there are plenty of scenarios that look a whole lot like that one—complicated and impossible to reconcile by giving one interaction 100% of the credit.
So you’re convinced you need more nuance in your attribution, right? Let’s talk about multi-touch (or multi-channel) attribution.
Your Multi-Touch Attribution Options
To set the attribution stage, here’s a quick breakdown of the most common models:
The first three are what we call single-source models; the latter four are multi-touch models. Regardless of which one you choose to go with, multi-touch models give a larger picture of each user’s path and funnel leading up to a final sale or lead.
Multi-Touch Works: A Quick Case Study
One advertiser in the finance industry made the switch from single source to multi-touch in Q3 2017. This particular advertiser’s main KPI is ROAS, or return on ad spend, and they’d found tremendous success with what we traditionally consider higher-intent channels, like search. On search, they were consistently hitting their ROAS targets throughout 2017 and able to grow their investments to drive revenue growth for the business. However, they were struggling to hit their goals in channels like display.
The advertiser knew there was value in these channels, but the data just wasn’t there. ROAS for their display campaigns on Google was 50-70% lower than their search campaigns, and they had begun cutting budgets to reinvest in better-performing channels.
By Q3 2017, this advertiser’s display monthly budget had dwindled from $60k to only $20k, with most of the spend going toward retargeting small segments of audiences that worked well. They tried prospecting campaigns, targeting similar audiences to their customers, but ROAS was so low that they quickly paused. Despite the small, segmented remarketing audiences performing well, ROAS wasn’t showing growth, and managing such a low spend on those campaigns wasn’t worth the investment or time. By the end of July, the advertiser decided to pause their display campaigns.
While their display campaigns were paused, the advertiser decided to roll out a multi-touch attribution model, specifically Time Decay, giving credit to all touch points within a 7-day lookback window, including view-throughs, which they used Sizmek tagging to track. This was in opposition to their first-click model, where they previously only gave credit to the first click within a 24-hour window. After being paused for a month, they re-launched their display campaigns with a test budget of $30k, evaluating performance with their new multi-touch model.
Performance was strong out of the gate, with ROAS up nearly 200%, exceeding their expectations. They were able to finally see the impact of these more upper-funnel channels, open up their targeting to broader remarketing audiences, and even launch prospecting campaigns on display. By October 2017, the advertiser was able increase their monthly display budget 350% from what it had been pre-multi-touch, nearly $90k, while hitting their ROAS goals, which now matched search.
Through this experience, we learned that single-source attribution models fail to tell us how middle interactions contribute to the user journey. Even though these touch points might not be how users first find out about a brand, or even serve as the final interaction before a sale or lead, they help ensure that the brand or product is top of mind and sway the user towards completing the final step in their path. Though it can be a huge investment, it’s worth it to figure out the true value of channels like display or Facebook. Who knows: you could end up tripling your investment and revenue on a channel previously deemed ineffective.
Editor’s Note: Marin TruePath
Marin TruePath is a cross-channel measurement solution that provides you with actionable insights across the customer journey. With view-level conversion data that accounts for each device, marketers can finally make more informed budget allocation decisions with a complete picture of the customer journey.
For more information, reach out to your Marin Customer Success Manager or contact us.