Soumita Roy Choudhury
Sales Director(Asia Pacific)
Nov 16, 2016
For long marketers considered mobile advertising as an afterthought to their digital strategy. With sub-10% budgets allocated towards mobile just a handful of years ago, there was little impetus to improve efficiency and ROI of spend. But with mobile becoming the first screen for most consumers, over half of digital spend is projected to be slotted towards mobile in 2018, particularly mobile programmatic.
As a result, top of mind for many advertisers, is to devise an efficient programmatic media strategy for their brands. And core to this is the harnessing of data to inform buying media from RTB sources. But this is easier said than done. Below are four critical aspects that every marketer should be aware of, to design programmatic buying strategies based on mobile data.
Yes, there are vast quantities of mobile data out there. Unfortunately, not all data is of relevance in the digital world. Take the example of Telcos. Intuitively it would appear that Telcos generate the lion’s share of mobile usage data– every mobile interaction must pass through one or more Telco pipes!! Unfortunately, most of this data is non usable in the programmatic context. Primarily reason for this paradox is that device IDs (mobile advertising IDs), the key for targeting digitally on mobile are, not available to Telco’s. Traditional Telco identifiers like IMEI, phone numbers, MS/ISDN etc. have little value in addressing mobile smart-devices.
Additionally, there is the problem of mobile web and its inherent anonymous nature. Safari mobile, the primary browser for iPhones, does not allow 3rd party cookies. This makes vast volume of web browser data anonymous and hence unusable for programmatic buying.
The mobile data ecosystem is complex with different players owning different slices of data. For instance, in addition to the aforementioned lack of device ID data, Telcos only have information about its own subscribers – if a Verizon cellular subscriber connected to an e-commerce site on Comcast broadband, Verizon wouldn’t be aware of the activity, and Comcast wouldn’t know this user on its network. App owners have knowledge about their users’ activities inside their walled gardens, but not what these users do outside. Uber, for instance, possesses valuable ride information about their users, but little visibility into their music consumption online. Marketers are often interested in holistic profiles – such 360 degree consumer models are hard to construct.
Mobile data is plentiful, but notoriously quality challenged. Nothing exemplifies this better than location data,
often touted as the most unique information derived from mobile usage. Location data available from mobile supply
is of highly spotty quality – location data from same device, a few seconds apart, could appear to be separated in
distance by hundreds of miles. Some of this inaccuracy may be attributed to fraudulent behavior. For instance, to increase
CPMs, many publishers pass spurious location data. Additionally, accuracy of location data varies by country of origin.
More developed economies have much superior location data compared to less evolved economies.
Another quality issue is the “blinding” or encryption of data,such as media usage data . Media usage is an important component for understanding behavior of smartphone users. For instance, someone with a cooking app installed is likely to be a culinary enthusiast. Similarly, a soccer app user, on her phone, is likely a soccer enthusiast. However, in cases when the source app media is blinded as is often the case, the information cannot be applied to inform media buys in any meaningful manner.
Mobile usage generates 10X more data compared to desktop usage, on a per interaction basis. This essentially implies
that to store and process mobile data, companies have to invest heavily in infrastructure and also explore ways to
store and process petabyte/Exabyte scale data assets. Hence, it is not surprising that companies are very cautious in
commissioning mobile data products.Traditional Ad tech companies have found this the hard way that investing in Big
Data technologies and building in house Data products are not only costly but deviates their focus from the core
business of running Ad campaigns. Moreover, skills sets required to build such systems are also scarce.
Mobile data, particularly in Asia, is in early stages of evolution and many startups are now designing innovative solutions addressing some of the above problems. At Mobilewalla our data scientists and engineers have worked over a year to create ingenious solutions to devise a mobile audience data product that not only provides unprecedented scale but also high accuracy and relevance. The Mobilewalla team is constantly innovating by conceiving new technologies and unorthodox business models to bring to markets solutions that will make data driven mobile marketing more effective. Stay Tuned!!