Comcast: How customer experience drives product development

Customer experience is one of those buzzwords that has come to mean anything, everything, and yet nothing at all. Although hype-mongers and tricksters have co-opted customer experience, in truth, the concept is profoundly important.


At its heart, customer experience means all touchpoints or interactions between an organization or brand and its customers. It’s a simple concept that is fraught with complexity.

Think about the touchpoints that a typical consumer or business buyer may have when interacting with a seller through the entire lifecycle or customer journey:

  • First, they research the product and company, seeking information directly from the brand but also from reviewers, friends, and other sources.
  • Having decided on a product or service, the consumer evaluates where to buy and may choose among product variants and configurations. Obviously, the nature of the purchase, whether consumer or enterprise, for example, dictates specifics of the actual transaction flow.
  • Following the purchase, the customer may need post-sales help with setup and initial use.
  • Eventually, that consumer may seek technical support or other forms of customer service.
  • At some point, the consumer may make another purchase, beginning the cycle again, or, hopefully, remain with the brand as a repeat buyer.

While the brand can control some of these steps directly — such as product features or technical support — other aspects of this broad customer journey are fully in the hands of consumers. No brand can control, for example, the discussions that consumers have among themselves outside of official company channels. Through its actions, the brand can influence, but not control, these conversations.

Customer experience is challenging because so many points of interaction must come together to create a positive impression in the customer’s mind. Product design, engineering, marketing, sales, support, and service all contribute to a buyer’s overall experience with a company or brand.

To dive into the complexities of customer experience, I asked the seven-time author of books on customer experience and digital transformation, Anurag Harsh, to share his thoughts. Harsh is the chief marketing officer at IPsoft [disclosure: a CXOTalk underwriter] that develops cognitive AI technology for customer service. He also co-founded the large publishing company, Ziff Davis, giving him a broad perspective on these issues:

Customer experience is not just a touchpoint or a series of interactions between a customer and a company, but a voyage. I call it a voyage because customer experience demands re-wiring a company’s systems, employees, and culture towards the sole benefit of the customer.

Companies must learn to view the world from the customer’s point of view and deploy resources around the customer’s needs, to build a company culture that screams “customer first.” This task is hard, especially in companies with large numbers of employees, because everyone must evolve together, speaking the same company language.

In customer support, for example, creating the right experience goes beyond conversations between a customer and support agent. You need a customer-first culture that performs well behind the scenes at the back-end. All this directly affects the quality of the agent’s response back to the customer.

Great experience happens when a customer-first culture meets the right back-end processes and technology, supported at every step by the organization’s leadership.

Because customer experience is profoundly important, I invited two of the world’s top practitioners to join me on episode 267 of the CXOTalk series of conversations with innovators.

Chris Satchell is executive vice president and chief product officer at Comcast, where is he responsible for the company’s product, design, and innovation teams. Comcast has annual revenue over $80 billion.

Brian Solis is one of the most well-known authors and analysts on digital transformation. He is a principal analyst at Altimeter Group, a Prophet company. Among his reports are the 2017 State of Digital Transformation and the Digital Change Agent’s Manifesto.

See also: Comcast: How AI, machine learning, DevOps, and a bit of hardware may make it a smart home platform

During this episode of CXOTalk, these experts discuss the subtle points of customer experience and present a framework for thinking about the problem. The conversation describes how Comcast uses customer experience as a reference point when developing new products and services.

Watch the entire conversation embedded at the top of this page and read a complete transcript at the CXOTalk site.

Here are edited highlights from the discussion:

What are the core issues around customer experience?

Brian Solis: If you look at the proper definition of customer experience, or employee experience for that matter, it is the sum of all engagements someone has with your organization throughout the entire journey and throughout the lifecycle. It’s not just about any one moment. It’s about how all those moments come together.

Chris Satchell: It’s about designing the entire journey. There’s a couple of things we focused on [when I worked] at Nike.

One was this idea of consumer brand business. Do what’s right for the customer first; then worry about the brand and the brand promise you make to the customer; then worry about the business. If you get the first two right, the third will come.

The second one was about thinking about the entire journey. Every touch point on that journey is an interaction with the customer. That can be positive or negative so that you could be building promoters all the way along or detractors all the way along.

You must think very broadly, and so you think way beyond when you’ve got a product installed, or you’ve got it in your home. You have to think, “How did I learn about it? How did I acquire it? How did I pay for it? How did it get to me? How did I install it?”

At Nike, we would think all the way to, “What is my interaction with an in-store athlete that was serving me?” because that is a great connection point with the company. Every point along a journey is your brand. You have to be authentic, and you have to serve the customer correctly there. That’s some of the things we brought here.

Then from my time at Xbox, again it’s a lot about delivering the very best experience, not settling, and never being content with what you’re providing, no matter how good it is, because you have to think in this consumer world about how good you think your experience is. There is somebody out there merrily raising the bar on you. It won’t have to be in your sector. It doesn’t have to be in your industry.

Brian Solis: I wrote a book a couple of years ago called X: The Experience When Business Meets Design. It explained how to think differently about innovation by taking a step back and thinking about design for a new generation of customers and employees that are not in alignment with today’s corporate policies, processes, and even just how we think and how we think about productization.

The consumer doesn’t care about all of the politics and BS that happen within the organization. They just want the experience to be personalized. They want it to be great. They want it to be intuitive, maybe transparent in many ways. Innovation is as much about products as it is about policies, processes, and how we even work as well. I think the biggest thing is just shifting mindsets.

I look at today’s — I call them — Generation C. They’re not millennials. They’re not Centennials. They’re not Generation X. They’re just anybody who lives a digital lifestyle. What they all share is this heightened bar for expectations and these new behaviors. They’re impatient. For example, we talk about the uberization or the consumerization of technology. When someone uses Uber, that becomes their standard of engagement. When someone uses an Apple product, that’s their standard — or Google Search.

This new level of experience is blurring the line regardless of products or services that they want that same sort of intuition, that same sort of clarity and cleanliness throughout the entire journey. Yet, organizations are built on these 50-, 60-year-old structures that have all those things apart.

Chris Satchell: If I’m a customer, my reaction is, “Your org structure is not my problem.” We used to talk about this back at Xbox about trying to paper over the cracks in our org structure so that the consumer didn’t see. You’ll see so many companies when you track that product portfolio, it matches the org structure, and you’ve got to fight so hard to take that mentality out, and you’ve got to find leaders who will be selfless and say, “Okay. Yes, I have this release vehicle, but I’m going to take functionality from somewhere else. I’m going to give up something in my release vehicle because it doesn’t make sense to the customer.”

How do you empower teams to create great customer experiences?

One of the things that we’ve done that’s helped empower the teams — and it’s going to sound boring, but it’s so important — is we have this quarterly planning process. It’s how we take our annual goals for our portfolio and break it into quarters. Every quarter what we do is, the products managers, they get with all their stakeholders, wherever they are, including user research and what they want to do, and they write. They say, “For my area, here is a one-page spec of what I want to do,” and it’s something that can be achieved in a quarter for their end of the product or their product.

It says, “Here’s all the teams’ help I need.” What we do is we have this process where we stack rank them. Then we plan, and we just plan from top to bottom, making sure that any higher priority thing, you know, it fills resources in first.

There are 2,000 people in my organization and another 8,000 people we work with. What you don’t want to have happen is you start off on something and then find out one of the constituent teams can’t deal with the capacity constraints for that quarter, and you can’t deliver anything.

We solved that problem, but importantly, it gives all your partners somewhere to go. When they say mid-quarter, “We’d like to go in this direction,” or, “We want to change what’s happening,” you say, “Great. Talk to your product manager. If they like the idea, they can bring it to the next planning.”

It’s a way that we have managed to bring quarterly agility to annual planning. What we do is we only schedule 50 percent of our capacity that way. We call it “directed.” We do 50 percent of what we call “trusted capacity” where we just say to the scrum teams, “Hey, work your backlog. Put on your backlog what you know that you need, what the customer needs. That’s your capacity to manage. Go manage it.” We work very hard to carve off part of their capacity that they could just use to do the right thing. It’s taken us a year and a half of constant effort to get that to work, but it has helped us take the 36 teams that we’re feeding into video and make them more agile and coordinate across them. It’s agile writ large at a very big scale.

How do you measure ROI?

Here’s a controversial statement. I think it is pointless measuring ROI below the portfolio level for a given line of business. I sometimes have some very spirited discussions with our finance team around this. The reason is, we’ve got all these projects. They’re feeding into the overall experience the consumer gets. Then the consumer, especially in our business, has got a subscription they’re holding because of that.

When somebody comes to me and says, “Well, we need to know exactly what it costs,” I go, “Why do you need to know what it costs? You know what the portfolio costs.”

They’re like, “Well, so we can plan ROI.” I’m like, “How on earth do you know what the return is? There is no way to untangle these variables. That is impossible. It’s mathematically impossible. We don’t have that precision.”

And so, I think one of the problems is when people start measuring ROI. Measure it at an appropriate level. The level I think is appropriate is: Here’s what we invest in a business, and here’s what that business returns. If you start looking at features, and you start looking at product extensions and all these other things, and saying, “Well, we need an ROI,”

I think you’re missing the point in the modern world. You need to look at total investment, total return. That clears up a lot of the mess if you can convince people of that. I find that a lot of organizations love, would much prefer, to be precisely incorrect than right, because it gives them a sense of, “Well, we must be on it because we got all these detailed numbers.”

Well, the detailed numbers are fiction. We don’t know how the customer will receive it. How many of us see ROI projections that pan out?

Now, large-cap scale investment and capital investment, that’s a different matter. You can plan that. But, when it comes to consumer-based products, I just don’t think, other than the line of business, you can plan it. The first one is, if you can, don’t get caught in the game of ROI for small things. Talk about portfolio ROI.

Then what we measure depends. You’ve got your vanity stats because you want to know your population and what your monthly actives are and your unique users. But beyond that, you must measure, one, what you think is important. If you’re in a moment of truth, you need to measure success across a moment of truth. Maybe you need to measure net promoter score on one side then the other.

That means you must run experiments, take people through a new experience and measure what their net promoter score was at the end versus the net promoter score of people on the old path.

We have this idea of relationship net promoter score, so RNPS, which is the long-term [of] how you feel. Then TNPS, which is, through a transaction, how did you feel? Then other than that it’s, you’ve got to come back to the product teams. It’s like any good data science. KPI is no different. What question do you need to answer? You have to think about the questions you need to answer and then plan for the data to answer those questions.

From a development perspective, it’s great to put the infrastructure in to be able to say, “I want real-time stats. I want batch stats. I’ve got these different things that I want to get back from my application to make it very easy for developers to instrument.” Whereas, product comes in and says, “Could you find these things out for me?” They’re like, “Yeah, that’s easy. I can just go and add that.”

Beyond that, it depends [on] what you’re trying to answer for that question. If you’ve got a funnel problem with, “Hey, how do I track from when somebody downloads an application, how many people go through, set up an account, and they watch that first video and go to the second video?” That’s very different than saying, “I want to understand the heat map of how somebody moves through our user experience.” We’d say in England, “Horses for courses,” but it is about understanding the question; design your data feeds and your data analysis for the answer.

How does data help customer experience?

Chris Satchell: We have huge amounts of data on everything, whether it’s our products. You can only vaguely imagine how much data our network produces. We use it in many ways. We use it operationally to keep the service running, to give customers a great service. We also use it, as I said, to answer product questions, to understand where we should go next in our products.

We have a very strong machine learning, artificial intelligence, and deep learning set of core teams here. We’re using that data to recognize new insights in our products and also to

create new product experiences that you can only do with those intelligent methods.

The same with operations, feeding data in and looking for that sort of pattern matching recognition and next action recommendation that you can only do by using very deep networks to recognize all this data coming in.

We’re starting to use data as a way to change how we operate and as a core of how we build and the functionality our products deliver. I think that’s going to become common to many companies. Data will become part of the product.

We’re finding that the algorithms that are available are becoming a commodity. You can get great data algorithms everywhere.

The actual technology frameworks — whether it’s MXNet, whether it’s TensorFlow — analysis and modeling frameworks are becoming a commodity. The real thing you have as a company is your data. The models you build with that data, that is your secret sauce. That is your gold.

We’re very focused on using our data effectively. It’s a question of capacity. We have infinite amounts of great questions and things we can do. It’s just sequencing them through product development, through product insights, through network operations, and customer experience to get the most valuable things done first.

I think we always talk about big data. Now we’re talking about AI and machine learning, but all of that — let’s just remember, they’re just tools. Without great people thinking great ideas, without being able to develop it, without being able to take the insights or the data and have the actuation loop to affect things, there’s no point collecting it.

I used to joke that what would happen in the big data world, you’d have a board of directors that says, “We need data.” Dutifully, the company would go off and gather huge amounts of data. Then they would say, “Well, nothing is happening,” and so they go, “Ah, we need more data!” So you get even bigger data.

Then you realize a little bit later, you’ve got no insights from it, so you start building the insight engine. You have this, like, huge first bit, and then it narrows to insights. Then still nothing happens.

Everybody is scratching their heads, and then you realize actuation. There was no pathway to take the results we had and change the world based on that. You want it to look more like a pipe where your insights match your analysis match your ability to actuate.

There is also a cultural element where you need to check your ego and say, “Wow. I’m surprised. I had an insight. My insight was wrong, but I’ve got a new insight. Let’s go drive that.” If you can get those to line up, you can start making a change in the org.

Brian Solis: [Chris just described] the biggest challenge I’ve seen with data. This is across the board. The challenges for any of this are human.

You’re working against a lengthy career of experiences that are behind every executive or decision-maker. They got to that role of where they are because they’ve made great decisions along the way. Those decisions have fortified their experiences and have validated their beliefs and their perspectives. When you’re trying to challenge convention, data only reinforces what you want to see or what you expect to see.

Being a data storyteller and having common language [means] getting data to tell the story of what is happening based on assumptions that will challenge convention. That’s the art.

Final thoughts on innovation?

Brian Solis: I often talk about the difference between iteration and innovation. Many companies think they’re innovative, but they’re actually being iterative, which I describe as doing the same thing, but better, whereas innovation is doing new things that create new value.

I look at the Comcast or the Xfinity remote as sort of this metaphor for the two. Buttons are iterative: backlit keys, dedicated buttons. Then the voice, the whole infrastructure for voice was innovative.

Chris Satchell: It’s a continuum, so I think small iteration is just micro innovation. You need innovation that’s small. You need innovation that’s medium where you’re expanding products. You need innovations doing completely new things, and you have people dedicated across that time continuum.

CXOTalk brings together the most world’s top business and government leaders for in-depth conversations on digital disruption, AI, innovation, and related topics. Be sure to watch our many episodes! Thumbnail image Creative Commons from Pixabay.

(Cross-posted @ ZDNet | Beyond IT Failure)

RPA often starts out like a teenage romance: a lot of enthusiastic fumbling around that ends quickly, frequently leading to disappointment

Yes, folks, that was one of the key takeaways one of the delegates pointed out at the FORA Summit in London last month, where a very mature conversation took place about the real future of operations in this lovely robotic age (download your full copy here).

This packed-out event was attended by 120 senior executives, the majority being senior buyside enterprise clients, joined by the CEOs of the leading automation solutions vendors, practice leaders across the leading service providers and global advisors. and the HfS analyst team.  This was a chance to get beyond that deluge of wooden marketing and sales hype that is murdering our sanity… and get to the real nub of the of the issues plaguing a confused – and fumbling – industry.

Ten Big Takeaways from the Discussions

1. RPA needs to move beyond the teenage romance stage. One delegate pointed out that RPA often started out like a teenage romance – a lot of fumbling around with enthusiasm that ends quickly, often leading to disappointment. Past events have focused on the importance of change management to the process, however, our recent study of 400 automation buyers shows that a lack of clarity around the business case is the major barrier to RPA adoption (change management rears its head after all the fun and games of implementing the software):



2. RPA hype is over and it’s nearing time to retire the term in favor of Digital Operations and the emerging Digital Workforce. Hype needs to move from replacement to enablement. The benefit of automation and AI are not reducing the workforce, but enabling machine to human and human to machine interaction. Helping enterprises and governments make better decisions with data. Building a more virtuous cycle with automation, decision making and data.

3. The Pace of Change Cannot Be Slowed – If You Aren’t Disrupting You Aren’t Surviving. Companies that view disruption as an opportunity and are not complacent are the most successful. Paranoia about the world ahead is your friend – driving staff to innovate and disrupt. Technology in this circumstance is a tool not a solution. Our customer panel said that there are ‘burning platforms’ already being created and businesses are going to have to come to a decision at some point soon to adopt. The supplier panel were agreeing that automation is surviving for big businesses and large enterprises have less than five years to sort this out.

4. The biggest challenge for Automation is the shift to scale. It’s not a technology problem, but an organizational change issue and how to achieve a broad set of outcomes at scale. Currently many implementations are sub-scale – tens or hundreds of bots instead of thousands they could potentially be.

5. Ultimately the world needs to shift its economic measure from effort to outcomes – where value is linked to achievement rather than the effort to achieve. The value of relationships need to be more interactive than ever, to make the shift towards outcome-based engagements, and away from effort-based.

6. The C-Suite is paranoid about the future and eager to make changes, while middle management is complacent and resistant to change. Culture is a major impediment to changing this dynamic. This requires a number of changes –change in the way companies operate, change in the skills that companies value, change in the incentives and the training that enterprises offer staff.

7. To adapt we need to constantly learn. This means better understanding of new technologies, better understanding of underlying processes and what can be improved. But ultimately it is about the best way to drive outcomes within the business.

8. We still need more use cases – especially as we look to Cognitive/ML/AI. As the hype shifts to higher forms of automation the need for use cases for all automation expands. There needs to be clearer understanding of where the value lies and where the process should begin. RPA is being passed over even when it offers 80% value for 20% cost and should be recognized as a valuable tool in an enterprise’s arsenal.

9. The purpose of digital is to bring humans and technology together. One of the panellists made the comment that digital was not about specific technology or about a transformation. “Digital” is about the bringing together of humans and technology. It is the interface, closing the gap between the two.

10. Change management remains a vital component of automation strategies. The difficulty in delivering at scale is exacerbated by poor change management and planning. It’s clear from our event in Chicago and in London that enterprise customers and service providers need to spend more time on planning to get automation to work effectively. One senior buyer representative said “change is not like flipping a light switch… more like a dimmer where it comes to full light over time and every new leaders is a new start.” So there needs to be a clear outcome and commitment – one of the main topics of conversation during the event was around the need for better change management to ensure that nothing is left behind in the race to transform. With important advice that “change management is about educating people slowly toward what the world will look like tomorrow”.

The Bottom line: Here are the anti-fumbling themes taking the conversation to New York this coming March…

To conclude the London summit and take the narrative onto our biggest and baddest FORA summit yet, the following four themes will steer the next phase of this industry mandate:

The Technology – a means, not an end. Data is the currency of transformation

Like with many new technologies, analysts, consultants and industry practitioners become obsessed with definitions and the demarcation between automation variants: in this case RDA, RPA, AI, Machine Learning, Cognitive, and all their permutations and combinations. Whilst this might be important for market sizing and positioning – many of the conversations in London reinforced the point that technology is a means, not an end – deemphasizing this definitional obsession. All these pieces of tech are tools, not solutions themselves. Without a coherent, end-to-end business transformation strategy, “dabbling” with automation technologies frequently does more harm than good, at best yielding only meagre results.  Given the amount of potential disruption to legacy work practices businesses are facing, a deeper transformation strategy is required which will take automation at scale – “you need to go big,” as one participant put it, to get real benefit from automation.  But first, organizations must map out the path to understand where they’re going.  This brings with it another crucial part of the transformation recipe – data. Understanding the centrality of data to the digital enterprise – how to acquire, structure, interpret and act on it – is essential

The Value – shifting the metrics from effort to outcomes

Much of the discussion during the event focused on the outsourcing services industry, in part because that’s where the prevailing labor-arbitrage business model is under existential threat, and in part because that’s where automation technology is already being deployed at scale.  During his keynote Phil Fersht observed that “Transactional outsourcing’s death throws began in 2012” – dating its demise to the rapid emergence of RPA.  However, there is a new, business model within reach.  Providers have meaningful experience with automation technologies and valuable know-how, while buyers desperately need expert help with design and implementation.  What’s needed is a new value proposition – one that separates effort and time from cost and revenue, and shares risks and gains.  “Clients will have to contribute value to their suppliers,” as one participant put it, and providers will have to become more innovative and willing to expose their balance sheets – in short, being less transactional and more consultative.

The Talent – taking the robot out of the human and putting insights back into the process

As has been discussed at the FORA and HfS Summits in the past – and as noted by Professors Leslie Willcocks in London, automation is not about replacing humans, automation “takes the robot out of the human.” Taking the mundane and process-centric tasks to free up employees to engage in more meaningful activities.  Artificial Intelligence, on the other hand, augments and extends the human mind, empowering humans to make more consequential decisions.  Together, they fundamentally change human behavior and workplace management paradigms.  In the digital future, all employees will need skills in data analysis and interpretation, and middle managers in particular will need to be able to connect the work they supervise with the outcomes the business requires.  Both must be granted what one participant called “permission to change” the way they have traditionally operated, and business must invest to equip them with new skills to succeed.

The Change Imperative – the way operations support the business itself needs to be redesigned

There is a growing awareness that we are at a step-change – a discontinuity – in business history.  The challenge presented by digital and automation technologies can only be met successfully with a commitment to transformational change; incremental, tactical approaches will only yield limited results and risk failure.  As never before, senior executives in every industry face existential decisions about the future of their enterprises, and will need to “make themselves uncomfortable,” as one participant put it – to re-imagine their businesses based on the centrality of data and digital relationships (see Technology above).  They will need to shed the constraints of the “as is” and articulate the journey to the “to be.”  Success will be measured not on beating last quarter’s results but on the ability to see and grasp the scale of change required and create a viable and compelling digital vision for what one participant called the “journey to improvement.”

HfS Subscribers can click here to download their complimentary copy of the London FOR A Mandate “Arise the Digital Middle Manager!”

(Cross-posted @ Horses for Sources)