Delivering retargeted ads to mobile users is a challenging practice. Marketers must decide whether the same consumer browsed through multiple mobile sites and apps, then evaluate the extent to which they can serve relevant ads without appearing creepy.
Retargeting firm MyThings believes it has found a solution through a combination of probabilistic technology and various identifiers. The 9-year-old company’s strategy appears to be working; its client roster includes ShopDirect, Orange, Toys R Us and Adidas.
AdExchanger spoke with CEO Benny Arbel and Jake Engwerda, VP of sales for the Americas and APAC.
What problems are MyThings trying to solve?
BENNY ARBEL: In the past two years, we’ve seen programmatic advertising and especially retargeting become uniform, template-based and black-box. This can work for smaller advertisers, but not for large savvy marketers. They want more insights, more data and a campaign that can actually meet their unique business and branding goals.
Customization is about giving marketers a choice of business models and a way to segment their audience according to their business goals – such as acquisition, repeat purchases, increased AOV [average order value], loyalty, cross-selling or reactivation.
Our focus on transparency is about providing access, via an advanced dashboard, to data on multiple levels including campaign, segments, media domain, creative, mobile and channels.
What differentiates your mobile retargeting solution?
BA: We’re seeing more and more companies launching mobile retargeting solutions. But if you take a closer look, you’ll find that these often lack the complete mobile stack. This could be from a variety of reasons: Either it’s limited to the mobile Web, or it’s not synced within a device, or it doesn’t run on mobile-compatible infrastructure that works with multiple identifiers like cookies, the AdTruth ID, IDFA and Android ID.
An incomplete mobile stack will also lack sufficient integrations with cookieless iOS and in-app inventory sources, or it can be based on first-party cookie dropping – which isn’t scalable and does not work on RTB because it requires a passback. We developed our mobile retargeting offering with a complete stack in mind.
What are you targeting against?
JAKE ENGWERDA: We’re using onsite activity. We do some upper-funnel things, but the core of our engine is built on what people are doing onsite. We started serving everything on HTML5 and we work with AdTruth to know whether this device is the same device you saw the bid request coming from.
Has your relationship with AdTruth changed since Experian bought them?
JE: No, we signed our contract with AdTruth before that happened and we feel that we should be fine for a certain amount of time. We’re not seeing any impact from the acquisition so far.
Can you give me an example of a mobile retargeting campaign?
JE: A typical use case example would be that someone is browsing through a site on a mobile device, maybe to get to a specific product. AdTruth is embedded in our tags, and every time a bid request comes through, we get the unique identifier from AdTruth and we choose to bid on that particular impression.
The ad we serve will be on HTML5 and it will be dynamically tailored to the products that the person looked at, and we’ll also show some related products using our recommendation engine.
How do you avoid cases where a consumer is served the same ad for something he or she already purchased?
BA: We have controls like frequency capping to make sure consumers are not seeing the same ad over and over again and analytics on our dashboard that help clients understand which ads are most relevant to a consumer.
Which success metrics are you using?
JE: We often use CPMs and sometimes CPC, but the click is not necessarily king. We know that not everyone clicks, so what we like to use is a CPM or a “Cost Plus” campaign, which gives a little more flexibility and transparency to margins. Cost Plus refers to the media you actually pay for on the exchange, plus a prenegotiated percentage.
At this point, we don’t do CPAs on mobile because mobile doesn’t have the same infrastructure for cookies as the desktop. With mobile, it’s mostly AdTruth’s numbers and there’s no third party to compare it to, so if we were going to be paid on a CPA, there’s a lot of potential for confusion.
Something that I think is interesting going forward is AdTruth being used by ad servers, which to me is an interesting way for the industry to move. If we’re using AdTruth to do our retargeting and ad severs are using AdTruth as a way of reporting back how accurate our retargeting was, no one is truly judging the third party. I don’t know how that’s going to play out. To me it seems like a possible conflict.
Where do you get your data? Is it mainly first-party data or are you also using third-party data sources?
JE: We will pull in third-party data when a client requests it, but there’s so much to be gotten from first-party data, and when you starting using data that many other people are bidding on, your cost of bidding is going to go up, because everyone is looking for a wealthy 18- to 35-year-old woman.
What we know is unique to the individual and more valuable than those data aggregators, so when we go to bid, quite often no one else has identified this person for what they are, which allows us to get in cheaper and the client gets a better return on ad spend.
What’s on your road map?
BA: Advertisers are looking for complete cross-device solutions and we currently support cross-device retargeting by using log-in data. To identify users across devices who did not log in, we’re developing an algorithmic identification solution that will be able to statistically predict that the same user is active on several devices.
