Solving the Dirty Data Problem in Location-Based Advertising | Street Fight

Solving the Dirty Data Problem in Location-Based Advertising

Solving the Dirty Data Problem in Location-Based Advertising

targetMike Schneider is a guest contributor. If you would like to contribute a post to Street Fight, contact us here.

Location and advertising go hand-in-hand. One great benefit of location-based advertising is the ability to set a targeted location preference in order to reach the consumers most valuable to a brand. A second benefit is reaching those consumers at a time and place when ads delivered by brands are most relevant. Accurate location-based data means better targeting for brands, more relevant content for consumers, and optimized eCPMs for publishers.

It should all work perfectly. Right? So why do brands see clicks from Asia when setting a targeted preference for New York? This is a longstanding problem that frustrates both brands and digital agencies alike. And it’s solvable.

Our work with leading ad tech companies has shown that 80% of the location data appended to ad inventories is inaccurate. The inaccuracies come from antiquated IP positioning providers who trace ad requests through the Internet to find the hub access point and assign the hub’s latitude and longitude to the IP addresses. Most IP providers use this network topology method to generate random latitude and longitude points within broad geographic regions. This essentially devalues location solutions, which are most effective when they are delivering precise locations. While country or state-level data may be trustworthy, low-resolution data prevents ad tech companies from assigning accurate context to their audience – such as demographic information or determining if a user is inside of a venue or walking down the street or at home.

Without accurate data, advertisers can inadvertently serve the wrong content to the wrong people in the wrong places. Even worse, if advertisers build audience personas on inaccurate location data, the efficacy of a campaign could be close to nil.

Making the assumption that “any location is good location” is costly for advertisers and their agencies. Worse, it’s a missed opportunity to deliver highly-targeted — and desired — content to waiting audiences. Estimating location to a broad block level area or zip code depletes both profitability and performance. For example, advertisers trying to target “New Parents” could waste their inventory on broad location points instead of focusing on the segmented, and accurately targeted, audience they want to reach, i.e., frequent visits to Babies R Us.

One relevant case study that highlights the significant impact of precise location targeting is our work with the data management platform, MEDIATA. The company needed accurate location to build user intelligence and to deliver relevant advertising, but like all server-side ad tech platforms, the only consumer location data source MEDIATA could access was IP address. Using Skyhook’s Hyperlocal IP, which positions over one billion IP addresses worldwide, often within 100 meters, MEDIATA could offer more contextually relevant mobile ads to their customers. With access to Wi-Fi, GPS, cell tower and device accelerometer data, MEDIATA could get a much better picture of the landscape, and deliver content that had time — and location — relevance.

What was the result? A 20 percent increase in the effectiveness of MEDIATA’s campaigns. More accurate user data translates to successful and impactful advertising.

Exact location also yields an increase in the effectiveness of programmatic bidding algorithms and the value of ad inventories by reaching consumers where and when they are in decision-making mode. Tagging inventory with enriched location data, like user behavior along with exact location, adds a brand new layer of meaning to ad units–and a time and place relevancy to consumers.

mike schneiderMike Schneider is VP of marketing at Skyhook.

 

 

Find out more about the importance of accurate location data at Street Fight’s Local Data Summit in Denver on March 5th. Hear from and network with top executives from Bing, Esri, xAd, and more. Register now and save $400!

2 thoughts on “Solving the Dirty Data Problem in Location-Based Advertising

  1. Mike, you bring up some great points. I’ve been watching this trend over time and was thinking about writing an article as well, yet from a slightly different angle. I might suggest there are two sides to the “dirty data” equation. We need both sides to be more accurate.

    One is what you highlighted, which focuses on the data associated with targeting a user/customer. The targeting of a users must improve over time and become more accurate. Hopefully this will occur as the overall scale of the audience improves and marketers can be more surgical in their targeting. As you know, there are a number of different constraints holding this back, some technical others audience based.

    But the other half of the equation has to do with the data associated with the business, hypothetically the data associated with the business trying to target customers. At the end of the day the targeting you mentioned is intended to drive foot traffic through the door of the business. Yet I often see bad data representing the location of the stores themselves. Unfortunately too many businesses have their locations based on the physical street address. Often this can get you close, but in my opinion it’s not nearly accurate enough. Especially if a store is located in a plaza, shopping center or a mall. In an ideal world we should know the actual “center of store” lat/long, all the entrances to enter it and how to navigate there (i.e. access road, best parking lot entrance, etc). Then use this data as the basis for the targeting customer…on the other half of the equation you mentioned.

    Targeting is only as good as its’ ability to identify the value of a user…based on the accuracy of where you want them to go.

    I like your solution and look forward to seeing how it evolves. ~ From Brett Hallinan at PositionTech

  2. It mustiness menstruum course into a elucidate compact. paid written essays Your essay was a toil of bang, they should get and maintenance as they plica over the pages, tied whether it is good a small bit. You may moreover be sounding for an say-so that will loan credibleness to your report

Leave a Reply

Your email address will not be published. Required fields are marked *

Name *

2 thoughts on “Solving the Dirty Data Problem in Location-Based Advertising

  1. Mike, you bring up some great points. I’ve been watching this trend over time and was thinking about writing an article as well, yet from a slightly different angle. I might suggest there are two sides to the “dirty data” equation. We need both sides to be more accurate.

    One is what you highlighted, which focuses on the data associated with targeting a user/customer. The targeting of a users must improve over time and become more accurate. Hopefully this will occur as the overall scale of the audience improves and marketers can be more surgical in their targeting. As you know, there are a number of different constraints holding this back, some technical others audience based.

    But the other half of the equation has to do with the data associated with the business, hypothetically the data associated with the business trying to target customers. At the end of the day the targeting you mentioned is intended to drive foot traffic through the door of the business. Yet I often see bad data representing the location of the stores themselves. Unfortunately too many businesses have their locations based on the physical street address. Often this can get you close, but in my opinion it’s not nearly accurate enough. Especially if a store is located in a plaza, shopping center or a mall. In an ideal world we should know the actual “center of store” lat/long, all the entrances to enter it and how to navigate there (i.e. access road, best parking lot entrance, etc). Then use this data as the basis for the targeting customer…on the other half of the equation you mentioned.

    Targeting is only as good as its’ ability to identify the value of a user…based on the accuracy of where you want them to go.

    I like your solution and look forward to seeing how it evolves. ~ From Brett Hallinan at PositionTech

  2. It mustiness menstruum course into a elucidate compact. paid written essays Your essay was a toil of bang, they should get and maintenance as they plica over the pages, tied whether it is good a small bit. You may moreover be sounding for an say-so that will loan credibleness to your report

Leave a Reply

Your email address will not be published. Required fields are marked *

Name *