Breaking Bad Location
Paul Thompson, MD at Blis, argues that advertisers need audience scale as well as accurate location data for their campaigns to be successful.
A lot of the “new kids on the block” in the mobile advertising world have recently been talking about the issue of location data “inaccuracy” – the term used to describe inaccurate, fraudulent and incomplete location data which is often supplied by publishers to increase the value of an impression. And quite rightly too. However, verifying the quality of data should be standard practice in any business dealing with huge volumes of data and not some massive revelation or secret sauce.
These new kids are also missing a crucial point. Removing data also removes audience scale which is required to make the proposition work for advertisers. So the question is: How do you remove bad data while maintaining a scalable audience?
We agree that if location advertising is done badly it wastes advertisers’ money because their ads are not sent to the right location. The industry is trying to address this issue and while location-targeting verification does filter out bad location data, it can remove as much as 85 per cent of the original audience. Using data mining, data analysis and data extraction techniques come as standard across the location advertising industry and Blis has been doing this for years. The companies operating in this space are using different technologies to tackle this issue, but we are all in agreement there is a lot of bad data.
Negative brand association
Even good location data needs filtering to ensure it is “brand safe” for an advertiser. Location data provided by certain dating apps, for example, might have accurate lat/long GPS coordinates, but it may not be an appropriate ad placement for an advertiser and could lead to negative brand association. If your location provider has built the right proprietary tools to understand where the lat/long data is coming from, it can remove the wrong impressions for your brand. Again, this kind of filtering should be standard across the industry. Less than 5 per cent of the data coming into the Blis platform from publishers can be verified as ‘good data’ and matched to our customer’s campaigns.
But with all this filtering removing location data, how do you build audience scale for an advertiser? Lat/long does provide good proximity, but it doesn’t provide real scale. If a location-specific ad is to generate successful engagement, a brand needs a bigger audience. This is where believe we can bring something else to the table.
We have the ability to build audience scale in location-targeting by understanding the relationship between IP addresses and specific buildings where there are all kinds of devices connected – mobiles, tablets and laptops. Our platform constantly matches IP addresses to locations globally at a rate of 3m IP addresses to building locations per day. We refresh this data every 24 hours to ensure its accuracy and context. This is on top of the 330,000 global points of interest built into our database. It ensures we can always build considerable audience in any location by the addition of more connected devices and their screens.
Another way to build audience scale is through data partnerships with the leading public wi-fi ISPs, O2 and Sky, covering more than 40,000 hotspots in UK high street retailers and other outlets. By overlaying this second-party location data to campaigns we can again add significant audience scale for an advertisers that would not have been there by only relying on lat/long data. More screens, more accurate location data, bigger audience.
There is bad location data everywhere and it’s available to everyone. Filtering it out solves this problem but without adding other content, demographic and third-party data, advertisers will always be faced with reducing scale. This is the issue we are seeking to address.
Paul Thompson is MD at Blis