Virtually every major telco has thrown its hat into the ad tech ring. They believe they are in a position to monetize their substantial data assets to construct rich consumer profiles for advertisers, in much the same way that Google and Facebook are monetizing their user data.
Just look at recent M&A activity, including Verizon’s purchase of AOL and Yahoo, Singtel’s acquisitions of Amobee and Turn, Telenor buying Tapad, and Altice buying Teads. Other global telcos have launched in-house ad tech subsidiaries, including AT&T in the U.S. and Axiata in Malaysia. Why this flurry of activity? Because telcos believe they have granular consumer data, at scale, that can establish a credible third front alongside search and social media ad platforms.
This hypothesis is mistaken.
Telco consumer data is far less comprehensive than they believe. However, most telcos are simply unaware of the large holes in their data. These gaps will be a major impediment to creating differentiated, powerful advertising stacks.
Telco data is not rich enough
Undoubtedly, telcos possess information about consumers that is both unique and revealing (billing information, call details, data usage). However, telco data assets suffer from huge gaps, preventing them from creating consumer models that are rich enough and complete enough to anchor advertising platforms capable of competing with those from search and social media vendors.
In developed markets, the bulk of subscribers are “post-paid,” meaning customers sign up for contracts and must declare information about themselves. Telcos only know information about the declared contract holder, not about others implicitly under the subscription. For example, a family in which the “contracted” subscriber is a 40-year-old male may have four devices under the contract—one used by a spouse, one for a teenager and another for the nanny of a younger child. But all devices inherit the subscriber’s declared profile, making the others effectively “ghosted”.
In fact, only one “named” subscriber exists for every three devices on post-paid subscriptions—meaning telcos only really have declared information on 33 percent of the consumer devices that subscribe to their network.
In emerging markets, the vast majority of subscribers are “pre-paid,” meaning there aren’t signed contracts. In this scenario, virtually no declared information exists. For security reasons, there are “know your consumer” initiatives in some countries, but the information required is minimal, and there is reason to doubt the veracity of the data.
Questioning location data
Consumer location data is generally regarded as one of the most actionable in advertising. Telcos frequently tout the wealth of location data they possess. Location-determination is performed through cell-tower triangulation, which provides accuracies that are not useful for making point-of-interest (POI) based-judgments. GPS data is more accurate, but telcos are often not in the logical “path” of GPS data flow.
Independent of accuracy, the timely availability of location information is also a big challenge. The data is stored in the file systems of network switches, in native formats, and retrieving it involves lengthy workflows between network engineers and data analysts. When building ad platforms, this problem is complicated by ad campaigns requiring “recent” data.
Access to consumer behavior
Another issue is understanding consumer behavior. Historically, telcos have tracked consumer behavior on mobile devices by observing web visitation patterns. The problem is app usage, invisible to telcos, frequently dwarfs mobile web usage of smartphone consumers. Plus, even when a mobile device is exchanging data over a network, about 60 percent of the time it is on a wi-fi connection on a carrier different from the telco, making it also hidden.
This means vast amounts of device usage is invisible to telcos, a serious impediment to establishing rich behavioral models.
Finally, telcos have a serious problem with digital “addressability” of their customers. Historically telcos have addressed consumers using telco-specific IDs such as phone and IMEI numbers, MS/ISDNs, etc. In the smart-device world, these traditional consumer IDs are not useful. The only way to deliver content into an app on an iPhone, for instance, is through Mobile Advertiser IDs (MAIDs) that are controlled by Apple. It has been impossible for telcos to map telco IDs to MAIDs, at scale. But this must be done in an effective ad platform.
Overcoming the challenge
These complex data issues are not insurmountable. A new class of data companies has emerged to address these problems and work with telcos to create specialized “vertical” DMPs. Employing big data, artificial intelligence and machine learning techniques, these companies are turning massive amounts of highly fragmented data into clean consumer models that enrich telcos’ native consumer data and help create the 360 degree consumer profiles needed to anchor powerful advertising stacks.
Axiata, a leading Malaysian telco, has partnered with Mobilewalla to create a powerful data-driven advertising platform that will launch in the next few months. According to George Chua, the director of data science and engineering at Axiata Digital Advertising, “we started off thinking of creating our service using general purpose DMP software and our own data. It quickly became clear to us that the big-data challenges with ingesting, storing and retrieving massive amounts of data made it infeasible to use off-the-shelf software. We also saw it was imperative to enrich the data we have on our own consumers, and we have developed a multi-layer data-driven advertising stack, in partnership with Mobilewalla, that is a blueprint for any telco wishing to bring such a system to market.”
The strategic convergence of big data with telcos is indicative of the enormous strides ad tech has taken on its path to maturity. There is no greater marker of sophistication than the ability to admit you don’t have all the answer. Couple that with a willingness to collaborate to develop the most savvy product. Every year may be “the year of mobile.” but with the help of big data, 2018 could shape up to be “the year of telcos.”
