By the end of this year, mobile will likely account for over half of all spending on programmatic advertising, according to eMarketer, and programmatic buying already accounts for a quarter of all spending on mobile — nearly 50% more than on desktop.
In many ways, mobile has always been a programmatic-first platform, says Lauren Moores, vice president of analytics at Dstillery. Street Fight recently caught up with Moores to talk about location data fraud, the role of location in programmatic, potential impact of Apple Pay and emerging attribution models.
There’s been some tussling over whether the quality of the location data flowing through ad networks presents a meaningful problem for the industry. Can you talk a little about whether you believe data quality was, and is now, a serious problem? Or is it more of an overstated speed bump of sorts. Would you call it fraud?
I do think it’s a problem, but I don’t think everyone understands it. When we first started looking at exchange data, the location information was very few and far between. You always had the user’s IP address and the time of day, but the amount of inventory with latitude/longitude data was typically about 10-15 percent [of total inventory.] As mobile expanded and more people entered the space, we saw a huge increase in the amount of location-enabled inventory from 20 percent to over 80 percent of impressions.
In terms of weeding out the bad data, we have something we call manhole cover, referring to one location with thousands of devices associated to it in a short period of time. If you have thousands of devices coming from a single location, we throw it out because we know it’s incorrect. In other cases, there’s situations where the bulk of the inventory from a given publisher is being driven by 10 percent of the devices. We know that’s just garbage.
Programmatic buying is set to account for the lion’s share of digital media spending on desktop in the next few years. Where do we stand on mobile, and what impact will that have on the value of publishers’ location data?
What’s interesting is that in many ways mobile almost is programmatic-first. Marketers who work with mobile are using it through programmatic means primarily because it came along after the reticence for programmatic on the desktop side. If you aren’t using programmatic in any way, you are missing out on the data science and behavior implicit in what programmatic can bring to the table.
It’s becoming increasingly common for folks in mobile advertising to argue that clicks are a terrible proxy for performance. Do you think the dependence on click is a data problem, in that most of the commerce activity might happen elsewhere and often offline, or an industry problem — simply the inertia of past behavior.
It’s more the latter. The use of last-click attribution is an industry problem that has lasted 18 years and is in part caused by both the lack or the inaccessibility of offline datasets. And it’s a bigger issue in mobile than desktop because of the screen size issues. We have done research that shows that the best apps that have high CTR are those that are flashlight related or kid related.
Ultimately, the value of advertising needs to be measured through sales, and sales can be measured on the desktop side more easily than mobile web and app, as the latter is more fragmented. Since we still have 9 out of 10 purchases happening in-store, we need to tie online behavior with offline spending and use it as a good measurement for brands that have brick and mortar locations. Up until the last couple of years, this has been difficult, more qualitative and inconsistent. We still have a far way to go to a consistent, global solution for access to measuring mobile conversions on apps and web.
A lot of folks have pointed to Apple Pay and other mobile payments efforts as a potential way to help connect online media and offline consumer spending. When, if ever, do you think a mobile ad startup will offer a click-to-pay ad? How could it impact the future.
Apple Pay and other click-to-pay models will help the marketer close the loop between digital media and offline spending for the iOS side of the equation. And with Google also entering the market, marketers will have two different sources of this attribution. For the rest of us, we will need to work directly with the marketer to get access to data to build better models and provide better measurement.
In many ways, the current in-store “tap and pay” approach actually breaks the data-tracking path. In general, the more solutions a merchant or brand supports, the more fragmented the data access becomes. If I have choices as to which I use to pay, then the marketer has to ensure that they are receiving all those different data types.
How do beacons play in the mobile advertising landscape. Let’s say, for instance, a meaningful number of stores install beacons. What would be the play for ad tech?
That’s the key: a meaningful number of stores to actually integrate it into their system. That could be exclusive to marketers depending on the platform available for the beacon. But it really plays more to loyalty and engagement and your customers now, less so on prospect at the moment.
If you think about it, if you’re coming into the store, you’re there for a reason. The accuracy of exactly knowing where you are in the store is going to differ on the technology. I can be standing in front of the Cheerios box and it can tell that? No, but you are going to know I am in that particle lane. The whole IoT technology will help with beacons but it’s more about the watch you are wearing, or the in-store kiosk, something that’s different that engages with us.
Liz Taurasi is a Street Fight contributor.