The Utility Customer Experience Investment: Unlocking the Full Potential of Personalization
- Sep 22, 2018 12:45 am GMT
- 673 views
Look around and it’s not hard to see: no matter which industry, we’re in the midst of a revolution—in customer personalization. Thanks to innovations in customer analytics pioneered by Amazon, Netflix, Pandora and others, not only do companies have the ability to offer greater levels of personalization and sophistication in the consumer experience—but customers are now coming to expect this.
In a handy summary of how the customer experience landscape is changing, with increasing connectivity across a multiplicity of devices, MarTech Today’s Andy Betts has gathered a compelling list of trends:
- Eighty-one percent of consumers want brands to understand them better and know when and when not to approach them (source: Accenture).
- By 2018 (this year) more than 50% of companies will redirect investments towards customer experience innovations (source: Gartner).
- Sixty percent of marketers struggle to personalize content in real time, yet 77% believe real-time personalization is crucial (source: Adobe).
- Ninety-four percent of marketers are focusing on their data and analytics capabilities, personalization technologies and customer profile data management capabilities to deliver personalized customer experiences (source: Forrester / Janrain).
Indeed, today’s marketers, customer experience managers, and journey architects have the right idea about where to concentrate their personalization technology investments. Modern customer personalization for any organization requires leveraging customer data, predictive analytics, and consistent delivery of recommendations across any channel the customer may reasonably expect (including mobile, call center, IVR, portals, email, SMS text, direct mail, even smart speakers). But while utilities are starting to think about the importance of this level of personalization, many have not yet begun to develop the initiatives required, to provide comprehensive recommendations across communication channels, or to manage and prioritize those recommendations on an ongoing basis.
What’s more, getting to sophisticated and accurate analytics-based personalization is challenging enough for a consumer products e-commerce company, where consideration must be given to past purchase history, browsing behavior, marketing campaign interactions and available demographics. But for gas and electric utilities, delivering the kind of personalized experience that today’s customers have come to expect is exponentially more difficult.
In addition to all of the above, utilities have to factor in analysis of customers’ energy usage (including average kWh load, peak periods, derived cooling/heating degree days, etc.), rate plans, income eligibility, program history, premise details (heating/cooking fuels, age, square footage, AC type, owner residency, single/multi-family, etc.), creditworthiness, the weather, meter type, geography, neighbor comparisons, outage history, and much, much more! Companies such as Amazon and Netflix, which are heralded as the gold standard for personalization, have the advantage of being able to focus their energy primarily on websites and apps with the singular goal of driving revenue. Utilities don’t have this luxury. They are mandated to serve all customers equally and must deliver on multiple objectives, from driving revenue and reducing costs to serving low-income communities, achieving energy efficiency goals, improving customer satisfaction, meeting regulatory requirements, and improving operational efficiency. The combination of these highly disparate challenges can make embarking on the journey of personalization seem daunting.
The good news is, even given these extra challenges, utilities can harness the abundant customer data available to take advantage of modern machine learning and analytics tools that can deliver the personalization their customers expect. The catch is—utilities unfortunately can’t rely on the exact same personalization and analytics strategies that retail-focused companies use (and this even applies to utilities that are trying to increase their own e-commerce revenue).
It All Starts with the Customer Journey
When talking about personalization, what we’re really describing is the ability to understand the customer’s world at a critical moment in their journey, and then putting the most useful and attractive offer in front of them. If that offer is grounded in a deep understanding of the customer’s needs right there, then the likelihood of that personalized offer being adopted increases dramatically.
Retail companies typically make personalized offers for a specific product or service they would like to sell and is likely to be adopted. But for an electric or gas utility, true personalization must include a wider selection of messages that support the utility’s diverse goals. These range from energy products to new rates, EE programs, savings tips, and safety messaging, as well as expanded e-commerce offerings (advanced thermostats, power strips, light bulbs, etc.); not to mention, for some utilities, an ever-widening array of value-added services, from installation to appliance repair, to tree service, to energy storage, to solar, to vehicle financing! In other words, modern utilities have a much wider pool of Actions to recommend.
So for a utility to determine the optimum mix of data, analytics and delivery, they first have to identify the particular customer journeys they would like to impact, which will then lead to the pool of Actions they want to recommend. For example, a program focusing on low and moderate income (or income-qualified) customers would have Actions offering discounted energy rates, fuel assistance, home weatherization or bill payment plans. By first establishing the customer journeys and the related action pool, a utility then opens the door to figuring out the required collection of data and energy analytics necessary to make their personalization story sing.
Putting Energy Analytics into Action
The power of using state-of-the-art, yet proven, analytics is that all applicable utility customers can be evaluated to determine which action(s) within the pool are most relevant to them at a given time.
And the good news is, there are more options out there than ever for businesses seeking to better use analytics to create more personalized experiences. For utilities, however, there’s a danger of over-reliance on analytical approaches that are primarily—or only—driven by e-commerce or marketing campaign data. Although extremely useful for consumer products companies, data points that primarily rely on Web views, MyAccount logins, shopping cart abandons, click-through rates, etc. aren’t sufficient for powering the analytics required to generate insights on a full portfolio of utility Actions (which, in addition to e-commerce products like smart thermostats, power strips and light bulbs, can also include safety messages, new billing plans, EE program recommendations, and much more). Each of these Actions requires a specific mix of data and personalization analytics to figure out the very best action (and the second best, and the third, etc.) to recommend to any customer based on their unique situation and energy usage at a particular moment.
So what kind of analytics are required for the specialty needs of utilities? Generally, there are five key categories of analytics that will be required for utility personalization: (1) core energy analytics, (2) forecasting, (3) peer analysis, (4) propensity, and (5) ranking. Each of these, of course, is a rich area of data science, and applicable across lots of industries. Where things get specific for utilities is the unique mix of energy-related algorithms, regressions, tools, models, and data sources, behind them.
Because “analytics” is a broad term it can be useful in the world of energy personalization to cluster the various combinations of analytical methods required for each personalized Next Best Action (NBA) recommendation into conceptual groupings that one might think of as “gears” (since, when put together, they form the engine that powers the utility personalization experience). Some of these gears require utility-specific data; for example, Average/Peak Load requires AMI data for a specific customer. Similarly, others gears, such as Weather Normalization or Cooling/Heating Degree Days, combine utility data with third-party data sources, such as local weather.
In this way, each desired action is driven by a mix of gears that uncover the required energy and customer insights to power utility customer personalization.
Solving the Data Problem
Believe it or not, when it comes to utility analytics, more data isn't always better—it has to be the right data. For example, I discussed above why, though highly relevant for retail sites, utility personalization can't be reliant solely on marketing interaction or web traffic data. Two more reasons for this are what’s referred to as the Data Sparsity and the Cold Start problems. Data Sparsity refers to the fact that most e-commerce-optimized analytics assume quite large sets of marketing interaction type data to power recommendations. For large retailers, this is the case. But most utility websites and/or marketplaces have much, much lower volumes of web visitors, and typical retail recommender algorithms simply aren’t optimized for these low interaction volumes to make accurate NBA recommendations.
Similarly, the Cold Start problem refers to the fact that most retail-focused personalization analytics require a minimum monitoring period before the system has gathered enough data about customer interactions to start making NBA recommendations (and even then, sufficiently high volume of website traffic are necessary, as noted above). For example, a common best practice for e-commerce analytics is to let the software observe website behavior for at least 30 days before launching—with the potential to substantially delay and complicate the process of launching new utility customer NBAs.
There’s some irony here, of course, because utilities are overflowing with data! The issue is that horizontal solutions haven’t optimized their analytics or data object models for the unique needs of utilities. Again, adapting these can lead to unnecessary implementation and operational expenses, and limits the ability to quickly scale up products and services.
Instead, utilities are better advised first to identify the internal data needed to enable the gears required, then to precision-source the necessary third-party data from the many available vendors, such as Acxiom, CoreLogic, Experian and more. When done right, such a system can be automated and scaled up as the number of Actions, analytics—and needed data types—increases over time.
Deliver, Optimize, Learn ...
Precision energy analytics and unified data are critical, but they don't do the utility any good if those recommendations don't get in front of the customer wherever they are at the moment of need. Real, end-to-end personalization, then, means the ability to integrate consistent messaging across any and all channels the customer is present in—from call center to chat to IVR to web to mobile to text to social media to digital assistants.
Of course, each of these channels has its own integration formats, standards and challenges. Being able to support integration across any channel requires putting the work in upfront to ensure that data is unified and properly structured, and is able to leverage flexible APIs (application programming interfaces) for integration. Integrating personalized customer NBAs into the plethora of channels available to utility customers can seem daunting. It is important to develop a channel strategy to capture the greatest business value through planning, implementation, training, metrics development and reporting. By staggering delivery channel deployments, lessons learned can be incorporated into each successive implementation.
However, even if personalized NBAs are successfully deployed into every possible customer delivery channel, there’s one more mission-critical step before personalization can be sustainably achieved—ongoing optimization. This is because the various customer data inputs that feed machine-learning based insights and NBA recommendations are always changing. The customer has a child, is laid off from work, buys a more energy-efficient refrigerator or an electric car, and suddenly their energy use profile changes: recommendations should change accordingly.
In addition, every time a customer interacts with their utility or adopts one of the products, rates or programs recommended, fresh data is generated, which can be fed back into the analytics engine to improve future NBAs delivered to that customer. By combining continuous energy use analytics, customer interactions, and what is already known about available products and customer profile history, this ongoing feedback optimization can even proactively indicate the applicability of a product the customer did not previously qualify for. The result is a level of dynamic engagement that delivers a game-changing customer experience.
Getting Started in the Real World
OK, so personalization sounds like a great idea. But how and where can a utility realistically get started when facing a seemingly overwhelming number of potential customer journeys, offers, programs, products, messages and channels? What’s more, any such new initiative could require cross-departmental buy-in, changes to well-established processes, and the need to integrate into aging IT systems—all of which can seem like a big enough challenge to make the most hardened utility leader run for the hills.
Now, this is the part of the article where we're supposed to describe the formula for easy personalization for any utility. But we're not going to do that, because I don't know the particular environment of every utility reading this. Instead, what I will say is that as impactful to customers as personalization initiatives are—from a new IVR to a re-architected Start/Stop process—they can’t achieve their full potential until data-driven, prioritized Actions are being successfully routed to every relevant customer. What’s more, without data-driven personalization, utilities will be unable to fully gain the organizational benefits of these recommendations, including OPEX savings, c-sat improvement, additional revenue, and increased program enrollment.
Fortunately, there is a manageable way to get started. Like many projects, even enterprise-level personalization is often best tackled by starting small. One best-practice is to utilize a 1/10/2 approach. This means selecting one key customer journey, ten related customer Actions to serve up, and two relevant channels, each of which should be strongly aligned with both critical customer needs and the utility’s current business objectives (such as reducing repeat calls, increasing customer satisfaction, and so on). By using this approach, utilities can get started on the path to customer personalization with a limited, manageable scope that allows them to prove the value and then expand with confidence over time.
For years, utility customers were content with stable rates and limited exposure to outages, but now they are expecting more across all utility touch points. Following the lead of personalization pioneers like Amazon and Netflix, personalization techniques such as the ones we’ve described today have proven effective in increasing value and customer satisfaction across industries ranging from retail to media to telecom to banking. Utilities can leverage the success of these personalization front-runners to achieve material results for their organization and, more importantly, for their customers. By starting with single impactful customer journey, utilities can apply proven technology to create a unified view of customers across data sources, leverage data science to match the right Actions to each relevant customer, and deliver these best-fit Actions to customers as appropriate.
This is modern personalization. With this foundation utilities will be equipped not only to better scale their growth to support an increasing number of customer journeys and Actions over time, but ultimately, to deliver a better customer experience for everyone.