[July 2020 Update]
Location targeting continues to be a very important signal for marketers to find their key audiences. Location data helps marketers find people who are near their business location, who have been at specific locations in the past, and who have a combination of behavioral preferences.
Particularly helpful right now is the deployment of look back targeting. In 2020 we aren’t going out as much, and our behaviors have changed. Still, location data allows us to see where people have been in the past. When we deploy look back windows we target people based on where they were recently.
Advertisers who want to reach LGBTQI to express solidarity with Pride can’t target 2020 Pride parades because they are not happening. However, we are able to identify and target the users who attended past Pride parades up to a year ago with messages of solidarity.
There are many uses for location targeting. Here are three:
- Restaurants targeting ads to people 2-5 miles around their address, and filtered to people who are seen eating out.
- Any spirits brand that wants to reach drinkers can target people who regularly went to bars in the last year.
- Target ads to people who recently walked into Coldstone Creamery for ads for an ice cream brand.
Earlier this year we wrote a comprehensive location marketing article for Street Fight Mag. In it we discuss ways to rebuild for the Covid economy using location advertising.
We also have more details about hyperlocal advertising in this video, where we go through a campaign setup with one of our hyperlocal data partners Factual.
Original 2017 Post
How do we know who the users are when we set up hyperlocal location targeting and audience based targeting?
The Short Answer
Location targeting data is tracked by your mobile device by latitude and longitude and sold to ad tech and data companies. Web behavior, purchase data and financial information that can be tracked digitally is tracked, anonymized, aggregated and passed around by an ecosystem of data providers and ad tech companies. Data is matched to publically available information like census data and other proprietarily created data sets using key identifiers like email address, cookie and device ID, which makes that data actionable to be matched with the an ad call.
This data is acquired by using latitude and longitude data from mobile devices. Each device has a unique ID, and the lat / long is being tracked. You can see some of this data on an iPhone by following the steps on this Business Insider article on the topic.
Apps are recording where we go, and that data is shared with a number of data companies including Factual, Verve and Ninth Decimal. So, then, we get to serve ads to people who are in the geo-radius we designate. For this campaign, we’ll look at 200-400 feet around an address depending on its size.
Cross Device and Household Extension
To extend the reach of this campaign we’ll capture the device IDs of users in our target area, and then serve them ads when we see them again. So in this case, if we see them at 5700 Wilshire we can then serve them ads when they are at home and on other devices. We use cross-device matching using probabilistic determination. Our partners look at the fingerprint of the devices (such as the specific apps on the device, the locations it goes to regularly and the wifi signals it regularly connects to) in order to match the devices that are owned by the same user. So, if we see a user connecting to the same wifi signal at home on their mobile device, and then we see some other laptops also connected to that device, we can say with over 97% certainty that the mobile device is owned by the same user.
Since these device IDs are being stored, we have the ability to go back in time and serve ads to people who in the past have been to these locations. So, we can serve ads to people who in the last three months have been to Wal-mart, bars, car dealers, retailers, or any specific commercial address.
If the targeting is broader; i.e. to the city or to multiple zip codes, we deploy behavioral targeting as well. In this case the ads are confined to the geographic area that we are targeting, and the ads are served to users who meet those behavioral targets. That same latitude and longitude data forges a view of the user’s interests, habits, and proclivities. Zip code, city, region and DMA targeting do the same. That data is then matched with other data providers (such as census data, banking data, credit card purchases, website behaviors) to create a richer view of that user.
When the user is calling up an ad a lot of information is passed through, including the location of the user down to lat and long (in most cases) and inclusive of the device ID, publisher name, publisher type, time of day, day of week and ad size. So then the targeting is matched to the info in the ad call.
That matching is done based on proprietary databases of user information, and strategic sharing of that information among partners. So, one company will have a lot of data about location (like Factual), and another will have a lot of data about purchases (say Mastercard), and another company will have a lot of data about owned identifiers (like device ID, email address, physical address and IP address – this would be SEMcasting and Acxiom, for example), and another company will have a lot of data about travel habits (like Alliant). The key here is the matching, so if one company has location targeting data and it wants to build out its travel segment, then it matches those key identifiers through Acxiom and with Alliant. All of this data is made non-identifiable.
Web behavior is straight forward. Your cookies track what you do online, and a data company sells that cookie data to other companies. Companies like Yahoo sell email address information to connect the email address to name and age. Mastercard tracks purchases and connects purchases to an email address to get further access to device IDs and cookies.
So, this is all to say that this infrastructure of data is readily available, aggregating hundreds of partners and the best in class advertising capabilities in the market today. When we operate a campaign we simply plug in the targeting parameters. We draw a radius around addresses, identify which of the over 93k data segments we want to target, upload the creative and launch the campaign. Then the hard work starts: optimize.
For example, if you’re a real estate company and you want to use location targeting to reach people who are movers you can use these data sets, and filter by age, gender, location and ownership type (own or rent). V12 has a lot of data tied to an interest in moving, so we pair that day with known renters.
- In Market > New Movers – V12
- Pre-Movers > 6 Month Pre-Movers
- Pre-Movers – Alliant
- Trigger > New Mover – The Data Alliance
- Lifestyles > Pre-Movers – Blue Kai
Each use case is different, but the infrastructure is there.