As the Cookie Crumbles: Strategies for CSPs Marketing in the New Era

As the Cookie Crumbles: Strategies for CSPs Marketing in the New “Cookie-less” Era

Nicolas Carre

June 22, 2020

For 25 years, Communications Service Providers (CSPs) and Marketing organizations, in general, have relied heavily on browser-based cookies to track consumer behavior online. That data has come in handy for targeting customers with digital ads and customized Web experiences that play to the products and services those customers care about.

But the cookie is disappearing from the Marketing team’s toolbox. Apple got things underway last year with its Intelligent Tracking Prevention (ITP) software, and Firefox quickly followed with its Enhanced Tracking Protection (ETP). Both browser-makers now block third-party cookies by default as a push to empower users who may have privacy concerns.

Similarly, Google Chrome is phasing out traditional cookies over a two-year period and is enabling consumers to block cookies themselves. This activity, combined with new privacy regulations like the General Data Protection Regulation (GDPR) in Europe and the California Consumer Privacy Act (CCPA) in the U.S., has big implications for CSPs and other online marketers.

How will CSPs need to change their approach to Marketing to their subscribers? How can CSPs continue to monetize their services, upsell and market new services in this new ‘cookie-less’ environment? While AI/ML will undoubtedly help, what role might it take, what are key use cases for Marketing going forward, and what are the next steps you could take now?

What Are the Alternatives?

Instead of relying on browser-based cookies, which supply data about the online activity of individual devices such as PCs and smartphones and individual customers, CSPs have an opportunity to take a new, more holistic network-level approach to Marketing across subscribers. CSPs already gather a ton of subscriber behavior data automatically – data that is just waiting to be mined by marketers. Network-based data, in fact, supply a different view than what Marketing teams are accustomed to seeing with browser-based cookies, but it can provide much broader insight into interests and trends across their subscriber base for their marketing demand generation initiatives.

While the browser-based cookie approach provides granularity into the activity of a single user, network-based “cookies” can supply trend information at the network level. Once collected, an AI/ML-based Big Data analytics solution can classify and make sense of it. By doing so, Marketing efforts will become less about a home improvement store selling a drill to an individual customer and more about Marketing to groups or communities of subscribers that fall within a variety network trend classifications, which allow Marketing teams to measure the potential audience of consumers on the network to be addressed with a particular service category.

Monitoring Your Market Share

Let’s look at streaming video and online music services as examples. The traditional PC/smartphone cookie doesn’t show all the music or video activity of a given subscriber. Rather, it shows the aggregated view of download activity for a specific artist. While you won’t see that view by collecting network data alone, you’ll be able to see, instead, what apps customers are using and what websites they’re visiting across the mobile network, when and where. This can be quite useful if you want to measure the market share of a given service compared to that of an over-the-top (OTT) provider that may be using mobile network infrastructure to deliver video or music content to your customers.

CSPs that offer Video streaming services, for example, can use network-layer intelligence to see what other Video applications customers are accessing over their networks—such as Netflix, Hulu, Amazon Prime, and so forth and so on. You can monitor this usage to help ensure that you don’t lose Video customers to other OTT video service providers. You can also measure how the release of new content by your competitors affects your own video service consumers. That could be a critical step: the global OTT video services market has shown itself to be a formidable rival to traditional CSPs’ content services and is expected to double in size between 2019 and 2023 to $72.8 billion, according to PwC Entertainment & Media Outlook.

To determine what actions to take to protect your share, you’ll first need to learn how much of the Video content streamed by customers is your own and how much of it belongs to OTT competitors. You also need to understand how your subscribers actively using Video services are shared between solely using your video services, sharing their video usage with other competitive apps, or just using apps from your competitors. You need to address them separately.

Data about the specific apps and content consumed by your customers is available on your network but you need to make sense out of it considering the large scale of IP-based transactions circulating on your network. When you couple the data mined from the network Packet Gateway with the modern dashboards of today’s advanced analytics systems, it’s much easier to see at a glance who’s using what, when, where and how much on your network.

This level of analytics provides a holistic view of your customers which the smart cookie doesn’t provide at the same level of potential.

What Are Users Actually Experiencing?

In addition to telling you which applications your customers are using and from which content provider they’re streaming, network-based analytics tools can help you appreciate what customers are already experiencing. You can then fold that information into decisions, such as identifying the best candidates for receiving service promotions, the types of promotions that are resonating with the audience, and what each subscriber’s preferred delivery method is.

It would probably be a misstep, for example, to heavily promote additional video services to subscribers already frustrated by poor or unstable experiences with their Video services. Using the right analytics approach, however, you can see what applications and usage subscribers are running and apply additional quality of services (QoS) metrics to that user’s stream to understand latency, jitter, and throughput to put some quality of experience (QoE) perspective on each customer.

Identification of heavy video users with a good video score: The usage of a cumulated distribution function against the population of subscribers showing with a Video score of 4 or 5 over a time period, in combination with their individual volume of Video consumption allows the Marketing team to quickly identify the right target for a new Video product launch.

How does this work? Through APIs, your analytics engine ingests and correlates streaming transactions at the application layer (ISO Layer 7) for millions of customers per day. That Layer 7 data includes http header information, DPI/EDR signatures, and payload information. It is correlated with the context information (GTP-C) and can even get precise location data from the S1-mobile management entity (MME) LTE events. The engine will then combine these data feeds with customer demographic information such as average revenue per user (ARPU) range and data plan information.

Data is then refined by content classifiers, usage measurements, and QoS scoring to produce the customer analytic intelligence required to segment users by the intensity of usage, QoE score, and other criteria.

Top-tier video users showing an unstable and/or poor video QoE can become candidates for network improvement or alternative solutions to fix their problems. At the same time, the video users showing a decent QoE score and having video as primary data usage by volume share can be targeted with a Marketing campaign for new or additional video services without the risk of irritating them.

Build Custom Campaigns, Enter New Markets

There are other ways to classify users, too. Are there subscribers who consume large volumes of video and also consume large volumes of music, for example? And is there a significant overlap of these subscribers in specific geographic markets? If so, a campaign could be created to target heavy streaming video and music consumers in a particular market, sometimes using a specific device operating system.

It’s also possible for telecom providers to understand other behaviors that occur on their networks, such as online shopping activity, and use that information to determine new markets to enter, such as retail or banking, or promote other online shopping services. Your Marketing department can begin formulating offers that are based on actual customer interests—movies, music, sporting events, shopping—rather than offering traditional sales packages that are all about network speeds, feeds, and minutes.

Removing the Network-Speak

Classifying content in this way has traditionally been difficult. While CSP marketing team executives might have succeeded in classifying some content, typically the analysis describes customer behavior using obscure Internet terminology. For example, user behavior might be classified based on variables such as domain name service (DNS) and volume of HTTP and HTTPS web traffic usage instead of stating who’s using Amazon, CarFax, or Fandango, for example. Finding this information can help providers feed their campaign management systems by classifying subscribers’ content behavior in terms that are understandable and actionable.

While the browser-based cookie may no longer be an available tool in the Marketing team’s arsenal, CSPs do have more powerful network data they can use to deliver just the right campaigns to individual subscribers. Combining that data with intelligent content classification and QoS capabilities that are inherently part of IP networks and using the right analytics engine and dashboard helps CSPs reach a whole new level of marketing intelligence that will better serve both their subscribers and their top and bottom lines.

Nicolas Carre

Nicolas Carré is the Director of Product Management -- leading the Marketing Analytics at Guavus, a pioneer in AI-based big data analytics for CSPs. Nicolas has spent more than 12 years in the telecom industry, mostly working with CSPs on customer behavioral analysis. The learning experience from these marketing problems is Nicolas’ primary source of inspiration for designing Marketing Analytics products. Prior to Guavus, Nicolas gained valuable experience as a Data Architect and was involved in building several startups from the ground up – including Neuralitic, Inc., which focused on Mobile Subscriber Analytics and was acquired by Guavus in 2012.