Get real with Blis: The key steps marketers should take for data accuracy and assurance

Share on linkedin
Share on twitter
Share on facebook

On day 2 of our Get real with Blis educational week, we’re looking at data accuracy and how marketers can ensure the data they’re receiving and using in their marketing campaigns is accurate and reliable.

When you’re aiming to reach and engage consumers, how can you be sure you’re getting high-quality data from your media partners and vendors? How do you know it isn’t fraudulent, inaccurate, or collected illegally?

With 64% of the population of India owning a mobile phone, 24% of which are smartphone owners there’s a lot of location data to be found in India – which means there’s also a lot of insight available into consumer behaviour. Reaching consumers while they’re on-the-go is important for marketers in India but gaining real-world intelligence into what consumers are actually doing, via mobile location data is vital in gaining a true understanding of consumer behaviour.  

Insights into consumer behaviour can be shared from the data collected from mobile phones, especially those that marketers want to engage with. That’s because, logically, where a person goes throughout the day says a lot about who they are. Whether a person goes to the same office every day, the gym three times a week, or the same café every morning, we get an important glimpse into how they live and what matters most to them.

The burning question, however, is how do you know that data is accurate? Marketers need to be cautious about the data they use – not only because it must be permission-based, but also accurate. When it comes to location data, a few metres can make all the difference in the world.

How your data is represented in a visual format is a key indicator in identifying whether the data is any good. Location data is generated by unpredictable humans, so if it is accurate or reliable data, it should appear irregular if you plot it on a map. There may be large clusters of activity around densely populated urban areas, but no true patterns should be seen. With bad data, on the other hand, you are likely to notice patterns, such as straight lines and repetitions that look more like they were generated by a machine.

At Blis, we’ve become experts at identifying bad data. In fact, it’s easy for us to spot if one single app is passing data that is being promoted as good lat/long data but actually isn’t. Only about 20% of the data that Blis receives makes it past our stringent quality standards and into campaigns. 

Below are the key reasons why the majority of data fails our quality assurance test and what marketers should be looking out for when using location data from third party suppliers.

 If the data isn’t precise: Centroid data may be accurate, but it’s not precise – and precision matters when it comes to location campaigns and gaining real-world intelligence on what your audience is actually doing. Precision is typically measured by the number of decimal places in the latitude and longitude (aka “lat/long”) provided in the bid request. The number of decimal places is a direct correlation to the level of precision eg. a lat/long that’s been rounded up to one decimal can represent a country or region accurately; two could be a county or district. Three can identify to the street level, but five can precisely define the location of a single tree, and eight can identify a single paperclip on your desk. We consider anything less than 5 decimal points too imprecise for campaign use.

If the uniques aren’t quite unique enough: If we see an unusually large volume of data originating from a single lat/long, we know something isn’t right. This is because when we talk about “uniques” in reference to location data, we’re not thinking of visitors or device IDs, we’re actually referring to unique lat/longs. At Blis, we require a minimum of five decimal places for location data to pass our QA test. From “single tree” level to spaces that are literally two square millimetres, we would rarely see more than one or two devices in a location at a time, so if we see more, that’s a red flag.

If the mobile data was generated by centroids: Campaigns requiring less precise location data might use centroids, centroid data doesn’t meet Blis’ standards for precision and accuracy. “Centroid” is really a catchall term for any central location that’s been used to identify where a user is at that moment, but it generally refers to latitude and longitude data that originates from a central IP in a metropolitan area. While it may be accurate, it’s notprecise: A centroid may be in Mumbai, but the users could be in Kajupada or corso Mankur. If you’re looking for shoppers of a particular store on the Linking Road, centroid data won’t help.

These quality assurance measures prevent nearly 80% of bad data from reaching our platform, and there are additional filters we apply to ensure our clients receive only the highest quality, cleanest data available. Smart Pin, our proprietary data cleansing technology, was purpose-built to validate the accuracy and confidence of all the data we receive.

Marketers should be aware of the data they are receiving and using in their campaigns and feel assured that if they notice any of the above trends, they need to question the accuracy of that data and whether it is right for their campaign, or to be used at all.

We’ve now covered what real-world intelligence is and how to ensure you are using the most accurate data in your marketing campaigns. Up next, the importance of data transparency. 

Share this insight
Share on facebook
Share on twitter
Share on linkedin
Scroll to Top

Thanks for subscribing.

You will soon receive our
latest news & insights