Engineering Customer Experience: Customer Centric Feedback Loops

One of the best parts of Teradata culture is that we collaborate, and we collaborate a lot. It was no surprise when I heard from Cheryl Wiebe after she saw my CDP post. She recently wrote about CX from the Six Sigma perspective and asked if I'd share some of my ideas on rebuilding the customer experience through that same process improvement lens.

Six Sigma has been widely recognized for process documentation, measurement, improvement, etc. What happens if your processes are wrong, though? I don’t mean “can be improved”, I mean WRONG. A “wrong process” is one that focuses on the company instead of the customer. If businesses are not focusing on the customer first, a whole host of poor incentives can crop up. An example from my past life in contact center analytics is Average Handle Time (AHT) - it’s a great measure of how long your agents are taking on the average call. Remember agents: every second spent with one customer is a second you’re not spending with another one, solving their problem instead... callers aren't people, they're a problem waiting to be resolved.

What happens when you measure AHT?

Management flogs the staff to take less time on a call. Suddenly any call that may take longer than your prescribed 210 seconds is ripe for an “accidental” disconnect, leading to a call back; add in shoddy work that gets a person off the phone quickly, or the magic “transfer to a supervisor” to get them off your line. How does the customer feel in all of that? (Please comment below if this has never happened to you.)

What Should We Measure, Part I

As leaders became aware of the above issues (even as it still goes on…), they raised the question of “so what’s a better KPI?” Ideas were thrown out, among them: Net Promoter Score/System (NPS), First Call Resolution, Effort Scoring, Customer Acquisition; Problem Resolution. Each of these matters to someone, the question that should be asked is “Do they matter to the customer?”

One of the first “customer based” metrics is NPS. Some people love it, some hate it. I like it when it is done right. Ironically, most of the surveys I’ve taken have been written with the company in mind, not the consumer. NPS was supposed to be about the consumer, right?

Customer Experience Feedback Done Wrong

Who has fatefully pushed 1 after being asked by the IVR “we’d like to hear your opinion after the call, please press 1 to take our brief survey”? 10 minutes after you’re done fixing some problem you wish you could have fixed online, you get a call back from the robot: it asks you how likely you are to recommend Company X to your friends (ok so far), then it asks about if you went online, how many times did you call, did you try the chat function, how about our email, did you know you could manage your account online? What happened to “our brief survey”?

Think about this as the consumer: you had a problem that you had to / chose to call in to resolve. You went online, you chatted, you tried emailing, you looked at the app. You couldn’t fix it. And now some box is intruding into your life (you did press 1, don’t blame the box for everything…) to ask about all the channels you tried that couldn’t fix the problem.

When I led analytic departments, the last thing I wanted to do was remind a customer about all the ways we wouldn’t or couldn’t work with them, instead of reminding them how easy we made it for them to do business with us.

Customer Experience Feedback Done Right

Start with your consumer in mind. OK, I know… get some water, take a couple deep breaths, come back to that idea.

Make your consumer #1 in your mind. Find out everything you can about them before you talk to them: how often are they on the website, what do they look at on your web pages, how long do they stay there, do they follow you on social media, how often do they contact you and via which channels? Diagram out what those consumers are doing across all your contact channels and begin to understand the paths they are taking to conversion (new product, renewals, acquisition from prospect), service (contact center, apps, asking for help on social media), complaints (complaints), or churn (not just disconnect, but also fewer services, less engagement, less time on your platform).

You don’t have to ask a customer if they tried to solve a problem on the web anymore – you can see exactly what they typed in the search box, where they abandoned the webpage, what they talked about to a call center agent and if they were trying to resolve it online while the agent was talking to them.

Using path analysis with NPS and digging deeper into your customers’ experiences helps remove the recency bias that is a traditional knock on transaction based NPS. The customers contacting you typically fall into two groups: either happy or angry with your company. My experience is that the folks who expend the most effort (yet are still promoters) or the least effort to resolution (yet are detractors) tend to have the most valuable feedback about your operations.

When you ping them for feedback, you can ask more directed and essential questions that focus around perception of effort, you can ask “what would have made this easier for you?” and use natural language processing to get new ideas.

What Should We Measure, Part II

With an effective consumer data platform and the customer focused path processes in place, you can learn a lot without even talking to your customer. What things can you start measuring and processes can you start fixing?

• Completed on website or abandon prior to completion?
• Cross channel assist on conversion (“digitally assisted conversions in call center or retail store”)
• What topics online, in digital communications or traditional channels are of the most interest, how to emphasize the relevant features and benefits?

What are some features that lead to KPIs that focus on the customer while still helping you meet your business goals?

• Raise the number of people converting online versus higher cost channels (convenience)
• Setting appointments online to build commitment to convert in person (certainty around time required - you know you'll be seen at 1:30pm)
• Providing the conversion experience and problem resolution through the first channel used / channel preferred by the customer (ease, convenience, empowering your consumer)

Start with focusing your analytic efforts on understanding how customers are interacting within and across all of your current communication channels. Fix the obvious problems – buried webpages, missing functionality, incomplete information. There are lots of techniques to do this using both structured and unstructured data. The next step is to start cross-channel analytics: what did people say in a search box on your website versus what they said in a contact channel, why aren’t they able to accomplish their goals? Do you use industry jargon on your webpage, are your keywords things that only you would know? Do you deliver sales pitches (“upgrade your TV receiver”) when people are searching on error codes, are you trying to sell them a new product when they’re at their angriest with you?

Wrapping up, for now…

The thought leadership that companies used to make their manufacturing processes better can be turned toward making your customer experience processes better as well. It’s about focusing on the key elements instead of simply the end product. If you measure the number of widgets produced, you’ll maximize your throughput. If you measure the number of widgets produced without defect, you have to start looking deeper into quality assurance and standards of performance. It’s no different when you start looking at “number of customers helped.”

Take business focused perspective out of the conversation. I know, it’s tough, you have to make money, right? If you take the perspective of the consumer, first and foremost, you’ll design systems that make it easy for your customer to do business with you. As you build out reporting, ask yourself, “would a customer care about this?” There are obviously cost controls and such, and the contrapositive is the amazing reputational wins you get from stories like Zappos’ almost 11 hour call. If you measure things that customers don’t care about, you’ll soon start managing to and optimizing things that customers don’t care about. Do that well and customers will soon quit caring about you.

I’ll be on a panel discussion with Blake Johnson from Stanford and Cheryl Wiebe on the Sunday (14 Oct) of Teradata Analytics Universe 2018. We'll be chatting with our audience about applying scientific and statistical principles to customer experience. Sign Up or reach out for more information!