About alanjporter

Writer of stuff - Books, comics, articles, and the occasional online ramble.

Machine Learning Isn’t Rocket Science

Take two astrophysicists, an Apollo engineer, a guy who designed parts of the International Space Station, a professor of robotics, and a random science fiction writer, and what do you have? It sounds like a dream sequence from the TV show, “The Big Bang Theory,” or the start of a science nerd joke. In fact, it was the make-up of a talk panel at a recent science fiction convention where I was one of the guests. The panel was ostensibly meant to be a retrospective look back at the days of Apollo, but like many such conversations, it soon turned to thinking about the future, which led to the subject of machine learning (ML) driven artificial intelligence (AI) and its current capabilities.

I expected an enthusiastic discourse, and so I was surprised when most of these actual rocket scientists seemed more ambivalent about the technology and its potential impacts.

A couple of observations caught my attention enough to tweet them out at the time:

“ML is great at recognizing patterns but not much else.”

“ML assumes tomorrow is going to be the same as today.”

Yet it seems these technologies are being received more enthusiastically elsewhere. Nearly every customer experience discussion and the majority of CX projects my team is engaged in these days includes some mention of machine learning and artificial intelligence (and often the two are used synonymously although they are different). Which got me thinking, how do the somewhat downbeat observations of a panel of space experts play into the world of customer data, and the ways we try to infer context from it?

‘ML Is Great at Recognizing Patterns but Not Much Else’

Machine learning is usually defined as “a set of algorithms and statistical models that computer systems use to perform a specific task without using explicit instructions.” It’s a subset of artificial intelligence that relies on patterns and inference to drive conclusions. In other words, as the scientists observed, it’s great at doing what it is meant to: Pattern recognition.

That means it can see what is happening in a data set, but not why it’s happening. That still (at the moment anyway) needs human interaction to derive context based on experience, knowledge, and a degree of intuition.

Machine learning can greatly reduce the workload and automate the process of recognizing patterns of behavior in large sets of customer data, but it is not a magic panacea for developing an understanding of why customers do what they do.

‘ML Assumes Tomorrow Is Going to Be the Same as Today’

The data we get from machine learning is a reflection of what happened the day the data was captured. For the purpose of pattern matching, there is an underlying assumption that the next set of data is going to be similar enough for the patterns and models it recognized to still be applicable.

Machine learning is not a predictive tool. It is a great way to analyze a lot of data and an efficient way to learn about repetitive behavior. But that’s it. The danger can be we take that baseline and believe that is how things will always be. Our customers acted that way yesterday, so they will act the same way tomorrow. If that was truly the case, to paraphrase Henry Ford’s observation, we’d still be riding horses. ML does not take into account the impact of disruptive social or technological influences. Overreliance on technologies like ML without understanding their role in developing a broader understanding of our customers can be just as much a blocker to delivering a good customer experience as any older system or technology.

We’ve Got a Long Way to Go With Machine Learning

When my wife and I get into my car on a Saturday morning, the ML system connected to my phone that analyzes my movements assumes we are heading for our favorite local diner. While that’s true around 80% of the time, on the odd weekend we head off in another direction, and the phone and GPS literally get lost for a while.

We have a long way to go (both figuratively and literally) with machine learning before it drives a true artificial intelligence-driven customer experience.

Customer Experience Isn’t About Fixing Discomfort, It’s About Preventing It

“Welcome back, Mr. Porter. Great to see you again.” 

Those were the first words I heard as I walked into my hotel after a day on-site at a client’s office. It’s always nice to be greeted in a hotel, especially one where you stay on a semi-regular basis. The greeting got me thinking, is there any better example of ‘baked in” customer experience than the hospitality industry? It is literally the key ingredient of the business (although I’ve stayed at my share of hotels where you’d think otherwise).

Technology Isn’t a Band Aid for Bad CX

How can we take the hospitality mindset and thread it through our digital transformation projects? I recently came across a sponsored post on my Twitter feed that declared, “Experience is everything,” and that we should “know what your customers are feeling so we can turn discomfort into delight.” It was of course a campaign to promote yet another experience management platform.

First, we need to realize that applying new technology isn’t the key to improving the customer experience. It should be an enabling tool, but if we don’t have the customer service culture in place, it’s a waste of time and resources.

But the Twitter campaign did get two things right: yes, experience is everything, in fact, it’s the key area of competitive advantage in today’s marketplace; and yes, we need to lead with empathy. But if you are starting with the assumption that your customers are already in some sort of discomfort when dealing with your company, then you have a systemic issue. Applying technology as some sort of band-aid will not improve the situation.

With Customer Experience, Actions Speak Louder Than Words 

Like the hospitality industry, thinking about the customer experience has to be woven into everything a company and its employees do, irrespective of their job-title or function, and it must come from the top. I once worked with a company where there was a lot of talk about “the customer” but very little seemed to be done to actually address known product issues. In fact, they had a reputation for ignoring their customers. The internal dialog seemed at odds with the external perception. Until one day someone said, “you do know what the CEO means when he uses the word customer? He means the shareholders and analysts, he means his personal ‘customers.’ He doesn’t mean the people who actually use our products.” That was like a lightbulb going off — it explained everything about that company’s culture and business model.

Compared this with another company I worked with. On the surface, it lacked the various functional roles you would expect would be necessary to deliver on the promise of customer experience. There was no customer support desk or call-center, no customer engagement manager, etc. In fact, no one had the word ‘customer’ in their job-title. Yet it has a top-ranked reputation for service with the people who actually buy and use their products. In fact, it also has a minimal marketing staff, because its customer’s word-of-mouth recommendations are such an effective marketing tool. The reason is that the idea that everyone in that company is part of the customer experience is central to the company’s culture, communicated from the CEO outwards.

That CEO understands you don’t have to be actually directly interacting with a customer to impact the experience. If the company has a customer-led culture it will be reflected outwards. Working in the accounts department sending out invoices? Think about how those invoices read to the customers. How easy is it for them to pay their bills? Designing a product? Think about how easy it will be to use. Putting together a website? Do you understand what your customers want to do when they engage with you online?

Customer engagement is a holistic experience. The customer’s don’t know, nor care, who works in what department. The response to any customer engagement, no matter where it originates, or where it’s picked up, should never be “Sorry that’s not my job.” It’s everyone’s job. If an individual can’t immediately help solve an issue, then build a culture where they take the details, find the person to solve the issue, and then follow up. (Never underestimate the power of following up.)

The Foundations of Customer Driven Cultures

I’ve outlined before what I believe are the three essential parts of planning any CX-related transformation project. They still stand as the core to baking in a leading customer-driven culture that will avoid the discomfort:

  • Know your customer.
  • Follow your customer.
  • Understand your customer.

But if you do find there are the occasional instances of discomfort (and to some extent they are inevitable), then rather than throwing technology at it I’d add a fourth and fifth edict:

  • Help your customer do what they need to do.
  • Follow up and build a relationship with your customer.

When Personas Go Wrong or, The Search for Fluffy

The woman on stage proudly told the conference audience how her team had spent three days to find just the right kitten for Emily.

Emily was a single working mother in her early 30s who lived with her 4-year-old daughter in a two-bedroom apartment. She was on a limited budget and often pressed for time. She also loved cats. Hence the search for the perfect kitten. The thing was, Emily didn’t exist. Emily was a persona dreamed up by the marketing team. The aim of the team was to create a series of recipes that used the company’s products — a series of recipes just for Emily. And they spent (wasted) three days looking for a photo of a kitten to accompany a made-up person.

I’ll be honest. I have a few problems with Emily — and others just like her.

Personas with too narrow a focus

By focusing on an individual as a persona you can narrow your focus too much and miss a large percentage of the customers and prospects who might benefit from your message. By creating messages “just for Emily,” the team was ignoring a wider need for anyone who wanted to create quick, nutritious meals on a limited budget. Personas should be focused on addressing customer needs, not on developing fictional characters.

It’s a marketing point of view

Often, as with Emily, personas are developed by the marketing team with little or no interaction with actual customers. Marketing teams are often organizationally isolated from everyday interaction with customers, which can lead to personas that reflect what the marketing team thinks customers are looking for, rather than what customers actually need and how they go about finding information. It is essential that your marketing team take into account real-life customer experiences and needs.

Customers are changing

I have seen many personas documented along the lines of “Emily goes to the website to do initial research, checks reviews on mobile, and uses the app to purchase.”

The customer experience evolves rapidly. I know my digital behavior patterns have changed over the last 12 months. You need to keep up with these changes. How often do you review personas to ensure that they keep up with new technologies and changes in how customers interact with your brand?

Still part of the “Sell and Forget” model

Historically, personas have focused on the buying behavior of a given set of potential customers. They are designed to drive people along the traditional sales funnel from awareness to lead to prospect to sale. But that only represents a small part of a customer’s overall interaction with a company.

How do personas fit with the continuous customer journey?

Once prospects become customers they shouldn’t be forgotten and neither should the relevant personas. How do your personas interact with your brand from delivery of the product through owning, operating, and getting support? Do you understand the full customer life-cycle of your personas and how their journey across every interaction with your company is connected and mapped?

Get that right and the satisfied customer persona can be your best advocate to generate even more business.

Was the kitten really necessary?

When you are developing needs-driven personas to help you understand customer behavior, your process needs to be systematic, efficient, and based on data. Building an emotional backstory for a character is all well and good if you are working on your latest novel, but it can be a time-consuming misdirection in developing effective customer-driven personas. How many customer interviews could that marketing team have done during the time it took to find the perfect photo of Fluffy?

Not Another @#$&! Survey

We all do it. Come on, admit it. I do it a lot.

You’re at the store and while giving you your receipt (which is probably three times as long as it needs to be), the cashier grabs a pen, circles or highlights a QR code or website address, forces a smile (if you’re lucky), and asks you to take a survey to “Let us know how we’re doing.”

Do you take those surveys? Probably not. I suspect that most people do what I do: toss the receipt in the nearest trash can.

Considering that I earn my living in the customer experience industry and like to think of myself as a customer advocate, it seems a little disingenuous of me to ignore those attempts to capture my voice as a customer.

Survey fatigue

The problem is that those sorts of surveys actually contribute to poor customer experience. Why should I provide a retailer with feedback and information that generates no tangible value for me?


I suspect that most customer surveys just add to a stockpile of data that no one looks at. This is just data collection for the sake of data collection, an exercise undertaken so someone can check the box when asked if the company has a program for capturing customer feedback. And when every retailer does it, the impact is the same as it would be if no one did. The surveys become meaningless. We have reached a point of survey fatigue.

Stop asking, start listening

When the average response rate to customer surveys tops off at around 10 percent, isn’t it time to stop doing them? Or at least stop doing them the way we are? If we really want to develop effective strategies for capturing the voices of our customers, it’s time to stop asking questions and start listening instead.

That doesn’t mean that surveys can’t be a useful tool. If used correctly, they can be a great way to start a conversation with your customers.
When engaged in a consulting gig, I often use surveys as a way to develop an understanding of how people feel and what works (or doesn’t) with the processes and technology the organization is using. These surveys get response rates of 60 to 90 percent and provide a lot of useful insights. Rather than blanketing a large group of people with generic questions, I target the surveys to discrete groups, with questions that relate to their day-to-day activities and that demonstrate an understanding of what the respondents are trying to accomplish and the challenges they face.


Surveys and voice-of-the-customer strategies should not just be about answering the question “How are we doing?” They should ask, “How can we improve things for you?”

Know the customer, help the customer

Every time you reach out to customers, you should demonstrate that you have listened well enough to know their needs and that you can help them. As a minimum, to demonstrate that you know the customer, tailor the conversation around the following topics:

  • What products they use.
  • What interactions they’ve had with your organization.
  • What’s important to them.

Then you need to demonstrate that, if they provide you with feedback and share information, you can add value and help them in the following ways:

  • Making their lives easier.
  • Reassuring them and/or directing to them more information.
  • Teaching them things that might be helpful.
  • Rewarding them.

Gathering useful information and opinions from your customers requires you to do more than simply gather data. The purpose of the exercise should be to develop an understanding of their needs and challenges. By responding in ways that add value, you demonstrate that you understand your customers, which will help you capture their true voice.

Are Your Customers Shouting Into The Void?

Many years ago I ran the support organization for a small software company. We had a whiteboard on the wall opposite the area of the office where my team sat. Everyone walking into the break room could see it. It showed the number of customer calls or emails we had each week, how many support tickets were still open, and how many we had resolved.

Above it sat another sign that said, “We are not a black hole.”

Don’t ignore customers

While the figures on the board were what we reported to the CEO each week, it was that simple informal sign that became our mantra. We didn’t want our customers to feel as though their requests for service were disappearing into a black hole. Let’s face it: no one likes being ignored, but more often than not ignoring people is the standard operating procedure of many support organizations. Even if it isn’t intentional, that’s often the way it appears to the customer.

It used to be easy to monitor and listen to your customers. They either called, emailed, or even wrote actual letters (remember those?) when they had problems. There was really no excuse for being a black hole and not responding to them. Today, providing support is much a more complex undertaking. There are an overwhelming number of channels that customers can use to communicate with you, and while you may be able to monitor most of them, it is almost impossible to capture them all — especially when customers come up with new, unofficial channels to make themselves heard, like the gentleman who was so unhappy with a company that he painted his complaints on the side of a van.

In my experience, companies respond to the voice of the customer in one of four ways:

  1. Ignore it.
  2. Capture it and do nothing.
  3. Acknowledge there is a problem but take no action.
  4. Acknowledge the problem and provide a solution.

Unfortunately, the response to any given request too often depends on when and where within the organization it was received and handled. This leads to an uneven experience. Companies do better when they treat customer input as a single unified data set.

Empathy first, followed by action

So how do you go about using that data set to deliver what the customer needs? The first rule of thumb goes back to not being a black hole. Acknowledge that the customer has a problem. Do that, and you’ll be way ahead of most companies. However, while empathy is all well and good, customers prefer action to empathy.

How do you enable your teams to take actions that help customers?

Give your customer-facing teams access to intelligent content.[fn: For more on Intelligent Content I recommend Intelligent Content: A Primer by Ann Rockley, Charles Cooper, and Scott Abel. (XML Press) 978-1937434465. ] Content is an expression of everything a company does, and it needs to be valued as an asset across a company. To solve customer problems and provide positive actionable feedback, you need to be able to tap into that pool of content in efficient ways that allow the right pieces of knowledge to be pulled together to provide personalized responses. That content can come from knowledge bases, technical documentation, support articles, operating schedules, customer profiles, or machine-learning chatbots. Match that content with current marketing campaigns and offers, and you can pull together positive customer experiences that help solve problems, further engage the customer, and continue to build brand engagement.

This comes down to taking a holistic, strategic view, this time in regard to your content. Look not only at what your content was created for but also at where else you can use it to answer customers’ questions. That is not a quick or easy task, but it is one that increases efficiency, leverages your content assets, and allows you to respond to the voice of the customer in the best possible way.

Let the Customer Experience Drive Your Technology Design

It all started with a new Christmas decoration my wife had set her heart on. She’d been back and forth wondering if she should buy it. Last weekend, we found ourselves back in the store, where she eventually decided it would be ideal for a space on the hallway table.

And so we found ourselves at the register, ready to pay for the new decoration. As is the case at most stores these days, the transaction was processed through a payment terminal card reader. You know, the sort where you can either insert or slide your card, or sometimes even wave your phone over it to pay for goods. It was no different from the hundreds of other terminals that we have encountered when shopping: In addition to the places where you could physically insert your card, it had a nice screen, a keypad and an electronic touch pen. 

All went well until my wife got to a screen asking her to confirm the amount. The green OK icon on the screen was an exact image of the green OK button on the keypad, so she pressed the button on the keypad. Several times. Nothing happened. I tried it too. Still nothing happened. At which point the sales associate smiled and said, “Oh, everyone does that. You need to press the OK on the touchscreen.”

An Experience Problem

There was no text on the screen to indicate that it was active, so why in that case was the OK icon made to look exactly like the corresponding button on the keypad? Or why wasn’t the system designed to accept input from either the screen or the keypad?

If “everyone” tries the keypad first, then you have an experience problem.

Two steps later, after my wife had inserted her debit card, the screen flashed up a question asking what sort of card it was. Why didn’t it recognize the card from the data swipe? The list of options it gave for the card type were clearly being driven from a backend financial system and the terms it used had no relevance to the average shopper. The sales associate had a cheat-sheet that allowed her to identify the correct response needed for the average bank-issued debit card. 

If you are presenting everyday consumers with professional jargon that an employee has to translate, then you have an experience problem.

Focus on the Needs of the User

As these sort of encounters tend to do, our shopping trip got me thinking about customer experience technology. 

It used to be that most technology interfaces with a customer interaction point were system-specific: One type of action being served and captured by one piece of technology. However, as the customer experience technology stack has evolved, we now have single interaction points that are either driven by, or feed, multiple backend systems. We also have, as is the case with the payment terminal, interfaces being developed that combine different ways of entering responses or acknowledging actions. 

When to use which action, and which technology to use, can be a difficult trade-off, but more often than not we find those decisions are being driven by the limitations of the technology rather than the needs of the user. By the time my wife had finished paying for her decoration, she had used the screen, the keypad and the pen. Why did that transaction need three separate types of action, when it all could have been done on a touchscreen?

Experience Should Inform Design

I’ve long been a believer that system limitations shouldn’t drive the experience. Rather, the experience should inform the system’s design. 

As the technology stack continues to evolve, we shouldn’t be throwing new technologies at individual problems, we should be taking a holistic outside-in view of how those new solutions will impact the way our customers do business with us. I realize that may not always be practical, but it’s a philosophy we should strive for when developing and implementing customer experience technology.

This is the season when we get to wish for a better world, and I’d like us to move toward one where the customer experience is a seamless, frictionless affair driven by great innovative technology.

By the way, my wife was right: The new decoration looks perfect on the hallway table. Happy holidays!

Is Your Voice of the Customer Program All Talk and No Action?

voice-of-the-customer-feedback-survey

A recent conversation reminiscing about those far-off days when we could go out to eat and attend all sorts of types of entertainment reminded me of a couple of customer experiences in the fast-food world that left an impression – not always good.

A Hot and Cold Customer Experience

On the way to a concert one evening, my wife and I stopped to grab a quick meal at one of our regular Tex-Mex fast-food restaurants. After patiently waiting in line, we got to the front and the server asked, “Can you wait a few minutes while I fill these online orders?”

I was not impressed with them giving priority to online customers over those who took the time and effort to be in the store, so I tweeted my displeasure. By the time we were sitting in our seats at the concert an hour later I had received an apologetic response from the company’s social media team and a promise to follow up.

The follow up was an email on Monday with a generic “please rate your experience” survey attached. The good feeling created by the social media team’s quick response was undermined by the apparent disconnect with the customer service process.

Bland Food, Seasoned Service

Contrast that to another occasion when I ordered delivery from another local fast-food chain. They emailed me when the order left the restaurant with a link to an interactive app where I could actually track the driver’s progress towards our house.

The food arrived ahead of schedule, hot and well presented with a print out on each container with the details of each individual order. But it wasn’t as tasty as we felt it should be based on our in-store experience — none of the food had enough of the sauces that give the chain its distinct flavor.

The following day I had a survey call from them too. Not a generic email, not even a robocall automated survey, but an actual person who listened. When I mentioned the lack of flavor she said she’d pass it on. An hour later I had a call from the manager of the local restaurant asking for details of why we weren’t 100 percent satisfied with the taste of our meals.

Guess which chain will get our service next time we want some fast food?

Suffering From Survey Fatigue

There’s a well-worn saying that we have two ears and one mouth so we should listen twice as much as we speak. Too many companies use surveys to try and measure the customer experience, and in doing so say they are listening to, and capturing, the voice of the customer. But the truth is that these don’t really work.

We are all suffering from survey fatigue. Every single store on a recent shopping trip asked us to fill in a survey by going to a URL the sales associate helpfully circled on the receipt. It sometimes feels as if you can’t undertake any retail transaction these days without being surveyed.

And how many of these do you fill in, or respond to? I’ve seen and heard industry statistics that suggest that up to 90 percent of customer experience surveys are ignored. Why?

Some of the most common reasons for the low response rates include:

  • Too many surveys from too many sources
  • Sending surveys after too much time has elapsed since the customer’s interaction
  • Expecting the customer to initiate the action (i.e. “Please go to this website and let us know what you thought”)
  • Customers don’t see any changes due to the feedback they give, so they stop giving it.

The last point speaks to the heart of the problem: companies are collecting data, but they are not listening to what the customer is saying.

A survey on voice-of-the-customer programs revealed that:

  • 75 percent of companies are only collecting or analyzing data without deriving many actionable insights
  • 46 percent are only collecting data without analyzing or doing anything relevant with it
  • 23 percent collaborate around this data with other groups
  • Only 2 percent transform their business using collected data and insights derived from it.

The ‘What’ and ‘Why’ of Customer Experience

How can you change that? Firstly by acknowledging there are two distinct types of questions you can ask in order to measure the customer experience: “What?” and “Why?

“What” questions may be the best way to capture data. They pose questions such as: What is the level of customer satisfaction? What is the likelihood to recommend? By collecting answers to these and similar questions, surveys can provide answers related to why satisfaction ratings are high or low but are often without context.

“Why” questions provide context and sentiment: Why is this customer calling? Why is that customer pleased or upset? Why, exactly, does this customer want to return, cancel or upgrade?  Interaction data includes customers’ open cases, phone calls, helpdesk tickets, sales orders or any other customer interaction information that gets recorded and tracked.

In order to derive the most meaningful insights from collected data, companies need not only to understand the “what” and the “why” of customer interactions, but must also be able to correlate the two. Customer experience managers need to take a holistic approach and consider both the feedback and interaction data as one unified data set.

By taking a holistic viewpoint of measuring the customer experience, it becomes easier to identify and plan actionable changes. A customer experience manager can start from a single customer mention of canceling service in a call recording and roll up from there to view survey scores, Net Promoter Score responses, and other related feedback.

The magnitude of the problem can then be assessed by rolling up to satisfaction levels related to certain features of the product or service in question. If the issue is compelling enough to merit a response beyond that to the individual customer, it will be easy to define an action plan accordingly.

The Start of a Relationship

When the customers know they are listened to, and that their feedback is bringing measurable results and changes, they are more likely to continue to respond and develop an ongoing relationship.

After all, listening is the key to any successful relationship.