AI for the Cable Industry
A practical guide for cable companies to use AI to improve customer interactions and increase profits. -- from Foundry.ai and Actifai
CONTENTS
- Abstract
- An AI Introduction
- Customer and Prospect Offer Optimization
- Data and Modeling
- AI’s other Cable Market Applications
- Conclusion
ABSTRACT
- Artificial Intelligence (AI) in the cable industry has primarily benefited the two or three largest companies who have the resources to develop these technologies entirely on their own;
- Recent advances in data availability, storage, and processing now make AI more easily accessible to the entire industry;
- The best AI initiatives are ones that focus on practical applications, not ones that are broad and all-encompassing, or ones that produce too little incremental profit to justify their resource expenditure;
- Customer and prospect offer optimization — including acquisition, retention, and upsells / cross-sells — is a targeted AI application for most cable companies with significant profit leverage;
- By aggregating demographic, behavioral, and competitive data and analyzing how each element impacts choice, AI-powered software can accurately predict the ‘right’ offer to present to a customer or prospect in order to maximize average revenue per user (ARPU) and customer lifetime value (LTV);
- Sales, marketing, and customer service teams are utilizing these tools and technologies to realize tremendous efficiencies and generate substantial returns for their companies.
AN AI INTRODUCTION
WE ARE NEARING THE END OF THE AI HYPE CYCLE
We at Foundry.ai have met with numerous CEOs and CFOs over the last three years, discussing how Artificial Intelligence software may improve the profitability of their businesses. The purpose of these discussions is to identify opportunities to apply ‘Practical AI’ — software powered by the application of data and math that is designed to statistically improve key business processes and increase profits.
In our experience, most of these initiatives will end up disappointing their sponsors. This perspective is aligned with the results of an extensive MIT / BCG report, in which more than two-thirds of corporate executives at companies with significant AI efforts reported that these efforts have not created material value. The problem is not that experienced senior executives are somehow misled by all this buzz. They know a hype cycle when they see one. They also understand the eternal verities of successfully introducing an important new digital technology:
This has been the playbook for the large companies that have most successfully introduced major new technologies from enterprise data warehouses (EDWs) to CRM systems to the Web, and it is a requirement for successfully introducing AI to create shareholder value.
The problem, rather, is that most executives do not have a sufficiently granular understanding of AI to allocate pilot-stage resources well. As with all technological advances, it is unnecessary for CXOs to understand the detailed inner workings of a piece of AI software, just as it is unnecessary for them to understand the detailed engineering of their mobile phones’ operating systems. But what they do need to know about AI is the answer to one question: What is it good for?
Most senior cable executives do not have a clear answer to this question. This is partly because AI has been (unjustly) mystified, and partly because AI in the cable industry has primarily benefited the top two or three companies who have the resources to develop these technologies entirely on their own. However, recent advances in data availability, storage, and processing are now making AI more easily accessible to the entire industry.
Thus, as companies begin undertaking targeted pilots to evaluate the impact of AI tools, it is important they do not address the wrong business problems, measure the wrong short-term metrics, and try to build platforms and roadmaps which — at this stage of AI’s maturity — will often do more harm than good.
MAKING AI MAKE MONEY
Choosing the right business application areas to address is the starting point for effective executive management of AI. Starting by taking a new technology and looking for places to apply it often results in tools that are not strategically important to the company’s mission, less profitable, and likely to be abandoned. In practice, we have found companies are much better off starting with the business problems, primarily because evaluating technical feasibility is a far more delegable task than judging where the profit and strategic opportunities sit in a business. The MIT / BCG report referenced above also indicates that AI initiatives under the direct purview of the CEO or other line executives are more than twice as likely to create value as those treated as a technology project reporting to the CIO.
Specifically, senior executives should identify a short list of core repeated decision processes with high-profit leverage that would be improved with better data utilization. We have rarely found them to be wrong about this. The work of the staff is then to estimate the value-at-stake for each process that is addressable with the AI technology of today — not the potential technology of five years from now. An AI pilot should not commence without a clear plan to generate measurable incremental cash flow within 12 months.
Every organization is unique, but there is one core business process that we believe holds very large AI profit opportunities for most cable companies: customer and prospect offer optimization.
CUSTOMER AND PROSPECT OFFER OPTIMIZATION
OVERVIEW
Efficient and effective customer acquisition, customer retention, and up-selling / cross-selling are core to driving growth and shareholder value for cable companies. At the core of these strategies is offer optimization, which we will define as presenting a customer or prospect with the ‘right’ offer in order to maximize the likelihood of a sale or save while optimizing for average revenue per user (ARPU) and customer lifetime value (LTV).
Best-in-class offer optimization processes allow sales teams to close more deals (outbound and inbound), marketing teams to generate more qualified leads (direct mail, online, etc.), and customer service teams to retain more customers and upsell / cross-sell products and services more successfully (phone calls, online chats, emails, and in-person). Despite its importance, performance varies widely across the industry, and offer optimization remains a major source of opportunity. There are a variety of reasons that this is true, including antiquated data availability, stale analytical and operational practices, and poor CSR adherence. Fortunately, two factors spurred by the introduction of modern and increasingly mature AI techniques are allowing organizations to close this performance gap.
WHAT IS NEW
The first of these factors is the ability to create and integrate new forms of data, or the so-called ‘democratization of data’. Examples include online expressions of customer sentiment, audio feeds from sales and customer service calls, input from network monitoring devices, weather forecasts, customer demographics, real-time competitive pricing, local events, and behavioral trends — all of which can be rapidly processed, ingested into existing EDWs by leveraging lightweight data APIs, and tied to detailed customer and transaction data. Gone are the days when sales, marketing, and customer service teams can only use a handful of inputs to manage customer interactions.
The second is the capability to model and analyze data, including these new forms of data, in previously impossible ways. Techniques such as deep learning, gradient boosting machines, and word vectorization allow modern offer optimization processes to significantly outperform legacy ones — specifically, by considering the numerous, subtle elements that impact a customer’s choice and quantifying each’s impact on a variety of KPIs. With these methodologies, recommendations are not only more accurate but are also more granular, actionable, and include the capacity to improve over time.
DATA AND MODELING
Making progress in either of these dimensions may seem daunting upon first glance. However, in our experience, the technical challenge of building AI models that accurately predict sign-up propensity, acceptance probabilities, and the like is actually not the hardest part. Typically, the most difficult obstacles are data capture and data management.
DATA CAPTURE
Cable companies already capture huge volumes of customer, transaction, and device data. While each of these data elements is valuable for predictive analysis, organizations can drive significant improvements in modeling accuracy by augmenting this information with exogenous datasets (examples below). To do so, they can leverage a combination of publicly available data harvesting, public and private data APIs, and other search methods to efficiently gather, utilize, and update this data. A core input for any offer optimization model is competitive information — i.e., which competitors are offering what product / service at what price, and how does this compare to one’s own similar-but-not-identical offers. Competitive behavior can provide insight into how much customers are willing to spend and can illuminate opportunities to nudge up price without sacrificing competitiveness.
Additional sources to capture can include:
DATA MANAGEMENT
Traditionally, cable companies have stored this information in complex, convoluted EDWs. Converting all of this to usable inputs to generate offer optimization models demands the integration of expertise across multiple areas including a modern, usually cloud-based, data pipeline incorporating non-SQL data structures and federated query capabilities and sophisticated AI data engineering for data vectorization, compression, and feature extraction.
Simply put, chasing the dream of pointing a generic AI method at a pile of customer transaction data to “find patterns without bias” generally ends in disappointment. The benefits available from AI-powered offer optimization software require careful and expert programming, but are fully achievable.
IMPLEMENTATION
Implementation of these AI tools should seek to minimize change to existing processes. Essentially, the system needs to provide acquisition, retention, and up-sell / cross-sell offer recommendations that are simultaneously granular, comprehensible, and timely. And most importantly, if accepted they need to increase ARPU and LTV.
The acid test is whether the sales, marketing, and customer service teams follow the recommendations, which requires confidence that they are in fact correct. To that end, these AI offer recommendation tools allow reps to ‘look inside the black box’ to understand what factors are driving the recommendations. Additionally, these systems can incorporate both won / lost results and direct user feedback to self-learn and improve over time.
CONCLUSION
While much of today’s AI focus centers on flashy use cases and large transformational investments, most of the real successes for cable companies have come from using AI in highly practical ways to improve core business processes. We have seen executives who follow the guidelines in this paper successfully capture this opportunity, creating significant value within surprisingly short timelines. We hope these concepts help you in this important domain and more broadly as you consider how to drive focused, pragmatic, and profitable AI initiatives within your organization.
ABOUT ACTIFAI
Actifai helps leading cable providers optimize their customer interactions to maximize ARPU and lifetime value. Our technology was created in partnership with a major cable operator that asked an important question: “If our reps knew everything they could about our customers and prospects, what would they do differently?” We’ve applied artificial intelligence to optimize rep’s profit-impacting decisions as they grapple with “What offer is most appealing to a prospect?”, “What will retain this cancel-intentioned customer?”, and “Can I service this location?”
ABOUT THE AUTHORS
Ned Brody — Founder, Foundry.ai
Ned has led a variety of technology businesses over the last 20 years, having been Yahoo’s Head of the Americas, CEO of AOL Networks, Chief Revenue Officer of AOL, and CFO of LookSmart. Previously, he started and led Mercer Management Consulting’s Internet Practice and San Francisco Office. Ned holds a MBA and a B.S. from The Wharton School of the University of Pennsylvania with concentrations in decision theory and finance.
Venu M. Amar— Chief Commercial Officer, Actifai
Venu is the Chief Commercial Officer at Actifai. Prior to joining Actifai, Venu was at Nike, where he designed and led the integration of Zodiac, an AI startup the firm acquired in 2018. Previously, Venu was a Vice President at Zodiac and a Principal at Applied Predictive Technologies / Mastercard. Venu received a B.S. in Economics, summa cum laude, from The Wharton School of the University of Pennsylvania.
—
