"Deep Packet Inspection can play a key role within cable operator networks; e.g., for traffic engineering and network security. But it presents significant shortcomings in holistically analyzing subscriber activity, which is key for both retention and growth." —Kate Mitchell, Edge Intelligence

Cable providers are at a critical juncture.

The number of consumers abandoning TV subscriptions for over-the-top offerings continues to grow. At the end of Q1, 3.4% of households cut the cord over the prior year, the highest rate ever, leaving about 83 million households paying for cable services in the U.S. This doesn’t include the increasing numbers among new households and younger demographics that have never subscribed to a pay TV service in the first place, a.k.a., “cord-nevers.”

Currently, approximately 13.5 million households (14% of all households) don’t pay for traditional forms of TV service. By 2021, eMarketer predicts, the number of cord-cutters will nearly equal the people who never had pay TV — a total of 81 million U.S. adults.

While this may all seem like doom and gloom for cable MSOs, it’s actually an opportunity to stop the cord-cutting trend and also win over cord-nevers through innovation. For cable operators to quickly turn the tide, it will require a stronger understanding of their subscriber base, or deeper than what’s possible with Deep Packet Inspection (DPI).

Seeing the Limits of the Old

DPI has been the default method over the past decade for examining and managing network traffic; it runs in line with production traffic or sends copies of packets to a network monitoring connection to inspect packets flowing through the network. Data is extracted from within each packet.

DPI can play a key role within cable operator networks; e.g., for traffic engineering and network security. But it presents significant shortcomings in holistically analyzing subscriber activity, which is key for both retention and growth.

What are those obstacles? First, it’s very challenging and costly to scale a DPI offering since it relies on inspecting at the packet level, on every port, at increasingly high network speeds. In addition, it can be difficult and immensely time-consuming and labor-intensive to gain customer insight from DPI systems since the hardware can be siloed and spread across many locations deep inside the network.

So how can cable MSOs obtain the subscriber insight they need to positively impact their business?

New Analytical Architectures

Big data analytics — the process of examining large and diverse data sets — can enable MSOs to discover hidden patterns, previously unknown correlations, customer preferences and other highly useful information to help them make more informed business decisions. And network data for cable operators is big, with hundreds of billions of records added daily, generated from millions of subscribers, and the need to retain trillions of records for analysis and compliance purposes. 

So the collection, real-time correlation, analysis and retention requirements placed on the analytical architecture are demanding — and many big data architectures are unable to keep pace. Analytics should provide the granular insight into and throughout the entire customer lifecycle that cable providers need to effectively support things such as usage-based billing, support-related inquires, proactive upgrades to bigger plans and anticipating those likely to churn. That knowledge can help inform activities directly geared to current and prospective subscribers.

For example, with the knowledge of subscriber behaviors garnered from big data analytics, cable providers can grow revenues through initiatives such as targeted promotions and customized product offerings. For those predicted to churn, better customer service and incentive offers may help in maintaining their business.

And for consumers who no longer subscribe to cable services but do still have data plans, providers can use big data analytics to determine their OTT viewing, web content and download data so they can figure out how best to monetize this use of their network. With this deep level of knowledge, cable providers can have accurate insight on data consumption to make sure usage-based billing and capped data tiers can capture revenue to offset what they’re losing from paid TV.

Going Deeper

While DPI still has an important role in supporting cable MSOs, it’s not cutting it in this time of cord-cutting. What’s needed is a way to understand subscribers on a deeper level than DPI can provide. By being able to better analyze the immense amount of data that’s available, cable providers can be well positioned to provide customers with personalized offers that resonate, incentives that motivate and service that delights — helping providers to retain and grow their business.

Kate Mitchell is CEO of Edge Intelligence, a distributed analytics platform.

Amidst the mind-boggling number of churn reduction strategies communications service providers (CSPs) have at their disposal, how are some leaders generating tens of millions of dollars successfully navigating the retention challenge? They are applying a test vs. control approach to answer questions such as: How should renewal pricing vary based on promotional price, tenure, PSUs, etc.? Which customers should we proactively call, and when? Can we profitably reduce the extent of save offers for some customers? Should we offer free streaming, a price discount or encourage a tier downgrade?

While provider executives know the importance of answering these questions — and most are invested in using data to curb churn — many still fall short in answering them accurately. Isolating how business actions affect churn over its natural rate is challenging. A few issues at play are:

• Home ownership: Apartment dwellers are naturally more likely to churn than homeowners as they have expiring leases that cause relocation.

• Tenure: Subscribers with less tenure are more likely to have promotions up for expiration.

• Seasonality: Customers who moved into a new apartment in the summer versus winter may behave differently.

Without accurately identifying each customer’s baseline churn, CSPs often incorrectly predict the impact of their retention efforts (accidently measuring differences in baseline churn for various customer segments instead of incremental churn caused by their actions). As a result, concessions are made to subscribers who would have renewed anyway; meanwhile, investments aren’t made in subscribers who could have been saved with the proper offer or outreach.

A growing number of CSPs are applying test versus control analytics to understand which retention initiatives are incrementally effective. The “test group” consists of subscribers who experience a given action, such as a price increase at the end of a promotional period. The control group is then constructed from subscribers who are similar to test customers across all other dimensions (e.g., age, tenure, homeownership), but did not experience that action (e.g., still on promotion). When highly similar subscribers are compared, any resulting performance differences can be directly attributed to the business action taken.

Knowing the overall impact of a retention program is important. Even more critical is understanding which resulting actions should be deployed to profitably retain each customer.

In an industry faced with an onslaught of new competitive threats, accurately understanding the impact of your business actions is critical. Before you incorrectly target more subscribers, consider using advanced test vs. control analytics.

Marek Polonski is a senior vice president at APT, an Arlington, Va., based provider of cloud-based cause-and-effect analytics software.