Carriers, devices and regulation - mobile targeting today can be a speed bump for marketers and publishers who drive to address or create scalable audiences. Placed founder and CEO David Shim believes understanding location data is the key to mobile targeting and Placed's analytics will help open it up as location becomes the new web page.
AdExchanger discussed the 16-month old startup with Shim recently. We began by discussing mobile targeting challenges today.
DAVID SHIM: In mobile, people are still looking at visitors, session counts, page views - when they should be looking at the important features that are unique to mobile. One of those features is location -- and not just latitude, longitude and what city they're in, but what businesses are nearby a person when they're interacting with mobile content.
Today, you can start to treat the device as a cookie, and the places that people go in the physical world are similar to web pages. We're essentially making web analytics for the real world.
AdExchanger: Can 'location' do in mobile what the cookie has done for PC-based display and video advertising?
There are a number of companies that AdExchanger has covered who enable location targeting: PlaceIQ, Placecast, and others. But you need to understand how an audience interacts with certain content, such as “3 percent of the time they're nearby McDonald's, 5 percent of the time they're nearby Wal-Mart, or 10 percent of the time they're nearby Best Buy.”
You need to have baselines before you start to jump into the mobile targeting piece. That's occurred in other channels such as online where comScore, for example, provides a baseline to say, "This is what the PC Internet looks like and here's the profile for these types of sites." …Arbitron for radio, Nielsen for TV… The market that we're going after is not necessarily the ad targeting. We want to inform the ad targeting.
But out of those thousand impressions on a mobile device, if you know that there's a McDonald's within two blocks, even if one impression drives that person to that McDonald's, that could be a very ROI-focused kind of buy.
We are very different from the traditional online targeting standpoint or online measuring standpoint. With cookies, they're opt-out. You go to a site, cookies will get dropped into your machine unless you have those blocked; and then, if you want, you can opt out of those cookies. It's a process that is working fairly well today.
With mobile content, the privacy focus is much stronger. When you download an application, you have to accept the set of permissions that say this application can collect location data. Or even when you go to an HTML5 site, and they use the geo-location API, a popup will say, "Do you agree to allow this website to get your location data?" The nice thing from a privacy standpoint is all the location data that we gather is all opt-in versus opt-out.
Could you share a use case for Placed?
In one scenario, mobile ad networks are interested in implementing place analytics with apps in their ad network. They want to create profiles of the places app users are nearby when they're engaging with content. A gaming app might have 10 percent of their audience nearby fast food restaurants.
They can start to look at all the apps they work with and see these profiles of nearby places where their content is being consumed, and start to package those apps up together. It's very similar to the old ad network rollups where they spot media across multiple sites, and they created a finance channel, a news channel, an entertainment channel, a sports channel.
Imaging being able to say, 50 percent of the time people are nearby a food and beverage restaurant when they're interacting with this channel of apps. Or this is the Whole Foods channel, so this audience has a higher skew to be nearby Whole Foods than any other grocery store. They can take these real places that people are nearby and use that as a selling proposition to make that inventory more valuable.
How is Placed relevant to a brand marketer's needs?
On average, 33 percent of mobile content is consumed when someone is in a state of movement -- walking, driving, or in transit... riding a bus or train. Using that data, we've had one marketer say, "I've always wanted to enable voice controls for this application, and I haven't been able to make the case because my team doesn't believe that people are using this app on the move." Now they are able to look at that hard data and make a business decision on the product that adds in voice controls based on this dataset.
Another marketing use case: if I'm a RedLaser or a Decide.com or an Amazon, and I've got a barcode scanning app, I can say 12 percent of scans happen at a Best Buy. Ten percent of scans happen at a Wal-Mart. Five percent, at a Target. It starts to give me this visibility of who is my competition offline. Where are my customers going in the physical world?
When you create a mobile app that's very brand focused…let's say you're testing a Coke app or even Adidas or Nike, you don't have any hard metrics other than visitors and maybe shares and session time. If I've got a Coke app, and part of it is figuring out where can I buy a Coca Cola, I can see when people are out looking for a Coke in the real world.
We can start to give them this data point that before just wasn't available. The closest you were able to get before was to go on on Webtrends, Google Analytics, Omniture, and see the big circle that represents Los Angeles.
Is there an application here for the mobile web?
We will be launching something with mobile web. We're beta testing right now with a couple of different scenarios, and the plan is to not only support iOS and Android, but also HTML5 or mobile web.
What's the difference between mobile web and mobile apps from your perspective, in terms of the challenge?
With mobile web, we still see a lot of activity where there isn't much movement. Someone might be using their laptop or their tablet, and they're just kind of sitting around. There's not as much movement versus when someone has an Android app or an iOS app on their iPhone. They're moving around, so you see more variance in terms of where they physically are in the real world.
What's the revenue model?
It's completely free. It's very similar to what web analytics was 10 years ago. The minute Google said, "We're going to make it completely free," you saw a huge uptick in people that started to look at analytics.
We think we're in the very early stages of location analytics where it's going to be important to have a solution out there in market that people can consume without barriers to entry.
How will you pay the bills?
In the longer term, our goal is to enable premium features in the analytics solution. This is focused around benchmarking and adding that as a premium feature.
People will start to be interested in understanding that it's great that McDonald's is nearby 15 percent of the time when someone interacts with my app, but what does that mean compared to the U.S. population? If it's only two percent, that's really interesting. I can take that to a media planner and say, look, we've got a lot of inventory available, and I can tell you that our audience is eight times more likely to be nearby a McDonald's than the U.S. population. You should advertise on my site, or you should advertise on my app.
We don't plan on getting in the targeting business. We think there needs to be an independent player on the location side.
What milestones do you aim to achieve in the next 18 to 24 months?
We want to increase the adoption of place analytics usage. We want to see an uptake with not only app developers but mobile content owners; marketers asking, what do your place analytics numbers look like?
How big is the potential pool of app developers?
You've got 900,000-plus apps out there. Figure 10 apps per developer, which is pretty high, and you're still looking at almost 100,000 unique app developers in market.
What's your headcount?
We're at 12. We're looking to add three to four more in the very near term. Our team is very data-focused, very science-focused. We've got four Ph.D.'s that are focused on mining data and making our model stronger. We have four engineers that are focusing on improving upon the pipeline and and supporting this large set of data.
How does what you're doing – and location targeting in general – relate to real-time bidding and programmatic media buying?
With a mobile phone and a mobile Internet connection, even if you have the best mobile service, sometimes you'll get spotty service. And when you're dealing with responses that are required in 20 milliseconds, it becomes very difficult to do real time buying.
It's starting to go there, and it will continue to go there, but there are still some technology hurdles.
People are going to… look at it not on an impression basis, but on a session basis. Rather than do real time bidding against an individual impression, you can imagine that at some point someone might say, okay, we couldn't make a decision within that 20 milliseconds just based on the Internet connection speed of the device and other factors. We can make a decision on that impression, but we know that user's going to come back again because they're consuming this mobile app. You can make a decision on that second impression that comes up based on the data that you collected on the first. There are ways to make it work.
By John Ebbert