Industry Insights: View-Through Attribution
Posted on September 11, 2017
The topic of mobile ad attribution and measurement has surfaced at almost all of the top mobile marketing events this year and has been a divisive subject amongst advertisers and vendors alike, all of whom are calling out for more industry standards when it comes to measuring the value of impressions. Another driver for the need for more accurate attribution models has been the rapid growth of the mobile app market and the emergence of popular ad formats like video ads, which tend to be viewed rather than clicked on.
In the first in a series of interviews with advertisers, agencies and adtech account managers, we caught up with three adtech experts to gather their insights on the topic of view through attribution and to explore some best practices for attribution windows. Paul Oka is Remerge’s Sr. Account Manager based in our NYC office, David Raphael is a mobile marketing expert and CEO of Infinity Media and Matthew Hrushka is Rosetta Stone’s Mobile Marketing Manager.
Why is it so important to track view-through conversions?
David: As marketers, we need to know the impact of each dollar spent and it’s clear that post-click conversions only account for a portion of the impact. Some channels like display will never work purely on a last-click model. This is due to the context of the ad exposure and the lower click-through rate of display inventory relative to social and search inventory. It is a common mistake to dismiss these channels from the marketing mix simply because they can’t compete on a last-click basis. I’ve seen multiple organizations cut off display since it performed worse than other channels on last-click; in each case, they saw the efficiency of other channels degrade and quickly resumed display.
Paul: I would agree with David here. I like to think of an advertisers mobile marketing mix like a soccer team: you need to have different channels or vendors to play different positions in order to achieve your marketing goals.
Your channels that have the best last-click attribution figures are akin to a soccer team’s star striker - they are incredibly talented at scoring goals and driving “conversions”, but when we look at the most successful soccer teams or marketing mixes, they are not solely comprised of star forwards. You cannot win soccer games with 11 star strikers, just like you cannot drive impactful performance campaigns with hard last-click attribution channels only. You need other channels and vendors to move the “ball” (user) down the “field” (funnel) and get your star striker to score. Not attributing credit to your defense and mid-fielders can lead to a losing team.
Do you think VTA will become the industry standard given the current shift in ad formats?
David: The explosion of mobile video inventory is leading an increase in low cost-per-view but relatively high cost-per-click inventory, which will naturally increase reliance on post-view attribution. Perhaps more importantly, dominant platforms like Facebook are heavily advocating for post-view conversions and helping clients conduct formal lift tests to measure the true impact of post-view. Where Facebook goes, most advertisers follow. As the friction to experimentation decreases and more organizations are able to measure the value of post-view conversions, I am confident that last-click attribution models will continue down the path to extinction. After all, last-click attribution models are only dominant because they’re intuitive and easy to implement - you’d be hard pressed to find anyone arguing that last-click is an accurate representation of marketing reality.
Matt: I agree with David but unfortunately, the problem is currently not all networks have VTA integrations but as we as an industry get closer to a full multi-touch attribution model, they will be forced to adapt. Being able to weigh VTA allows marketers like myself to take into account impression value just like clicks will allow us to spend and forecast more efficiently. This advancement within the space will also allow a better understanding of the media you’re buying as viewability, quality indexes, etc. will play a larger role.
Paul: To add to what David mentioned about video inventory, I think with the adoption of un-skippable 15 second video ads, VTA becomes increasingly important to measure. Users are guaranteed to watch at least 15 seconds of the video ad before having the option to skip to view their content. In that moment, the users are not necessarily looking to make a purchase, but will recall the ad when they are going to make a purchase, remember it and come back organically. If last click only was considered, that video ad would not have any perceived value to marketers, when the reality is the opposite.
What would you recommend as a good attribution window for your clients?
Paul: This is tricky, as there really is no one-size fits all methodology. Some clients we work with are subscription based or have a “once in a lifetime” conversion event. Once a user subscribes, there is no additional conversion that they can drive, unless they churn and a re-activation campaign is needed. Other eCommerce clients can have users convert multiple times, whenever they choose so the attribution models need to vary depending on the conversion events, monetization and life-time value of on app’s users.
David: Paul is right. Unfortunately, there is no one-size fits all methodology. Attribution models are always going to be imperfect. At its best, attribution supports the marketing department with data about the incremental impact of each media dollar on conversion so that they can allocate their marketing budgets most efficiently. The biggest challenge with post-view attribution is to measure incrementality - that is, to discriminate between post-view conversions that would have occurred without an ad exposure vs. the post-view conversions that would not have occurred without an ad exposure.
I always recommend that my clients conduct a formal PSA test or run with a DSP that offers both VTA and CTA data insights on their platform to determine the split between view and click-through conversions. To add to that, and this may sound obvious, but you should only be attributing post-view credit if your ad was actually seen! The majority of impressions are not viewable impressions, which the IAB defines as having 50% of the ad loaded on screen for at least 1 second. I would recommend using a service like Moat or DoubleVerify to ensure you’re only attributing (and paying for!) viewable impressions.
What would you recommend as a good attribution window for advertisers focused on sign-ups or registrations?
Matt: At the start of the campaign, I would suggest smaller windows to ensure the network is building user/audience logic. By doing this the network will find common patterns between the conversions and optimize spend towards those conversions events. After that initial testing period is over, I would suggest opening up the windows slightly to understand what type of conversions are being logged outside of that window as “organic” to provide and show full value of the media spend.
What insights can I gain by tracking view-through conversions?
Matt: At its core, VTA is exactly that - insights based on the value provided by impressions. The network you’re buying media from doesn’t know more than what you are providing them, so building out your attribution SDK to pass meaningful data to your media provider will allow them to become more intelligent and provide you with more value in the long run.
Paul: VTA also allows advertisers to get directional data on creatives. If one creative set is driving more VTA conversions than another, it could be safe to assume that those ads have better recall to the users, they are remembering the ad and when the time is right to convert, they are coming back after receiving that ad theme.
Looking at last-click only will not allow advertisers to collect this data. If you cannot measure it, you cannot test it, if you cannot test it, you cannot understand what is working. Getting comfortable and stagnant in this industry is a death sentence. All advertisers do not need to include VTA conversions in their models, but at the very least, the data should be collected and measured.
Want to learn more about view-through attribution and how Remerge can help you uncover valuable mobile marketing insights? Contact your Remerge account manager or shoot an email over to firstname.lastname@example.org