What skills do you and your customer experience team need to be successful? (5-minute read)
While pure luck was a major factor, my use of Microsoft Project version 1.0 on my personal Mac certainly had a major positive influence on my promotion to the European HQ of Digital Equipment Corporation. I printed out multi-page GANTT and dependency charts for a major project and used them to present to my future manager. It made me look like I knew what I was talking about. That was what I felt too: seriously competent.
Ten years in Geneva and many projects later, I had a new manager. Don Gordon invited me into his office in Munich to review a major program I was managing. After about ten minutes he said, “Maurice, you know almost nothing about program management.” I did my best to stay polite while disagreeing strongly. He said, “OK, just answer this question: What is a deliverable?” I have to say that I struggled to answer. And I quickly understood that it was a critical and fundamental concept in program management. A deliverable is the thing you can observe so that you can tell that a piece of work has been completed. For example, if you have completed some customer research, the deliverable will probably be a report or a presentation.
Program management skills are critical
The single most important skill set needed for customer experience improvement is program management. You and the people in your team have to be able to get things done in a professional and predictable manner. The easy way to learn the basics is to attend formal training such as that given by the Project Management Institute (PMI) or PRINCE2 training. Courses are available in many languages. Learning the basics from a colleague who has already attended formal training is an acceptable minimum.
Some program management concepts are quite important when you are presenting a new initiative for approval. The single most important thing you have to avoid is called ‘scope drift’. This usually happens when the program completion criteria are not well-defined. Imagine a situation where your improvement initiative is the implementation of a Chatbot as an option for customer support. When is the implementation complete? Is it when you have completed a proof of concept? How about a pilot covering three countries? Or have you completed your part of the work when it has been implemented in 87 countries and 20 languages? I strongly suggest you define this type of thing before starting.
Success criteria are also important and should be agreed before starting. Let’s suppose your Chatbot project is complete when you have implemented a pilot in three countries. Has the pilot been successful? How do you know? Is success defined by a number of cases closed, a customer satisfaction metric, a cost metric, or what? To avoid scope drift, you have to agree these points before starting.
I will try to avoid using the term Data Scientist. (Oops, I just used it.) Correct interpretation of structured and unstructured customer data is the second of the two core customer experience skills. Within that broad area, the most important knowledge areas are statistics and at least some basics of Natural Language Processing.
I feel strongly that statistics do not belong in any executive-level presentation. By this I mean you should avoid presenting things like confidence intervals. If you happen to get someone who loves statistics in your audience, everyone else will go to sleep while you answer his or her questions. You should simply invite anyone who wants the technical details to see you after the presentation. However, you still need to be confident that the numerical part of your work has been done competently. Any basic statistics course should do, and I don’t have any particular recommendations, though I suggest that you need to get beyond the Normal distribution. (Net Promoter Scores, for example, are not Normally distributed.)
Unstructured data too
If you use the Net Promoter System, and indeed most other customer feedback systems, you will get a lot of unstructured data too. This means the text answers to the open questions you ask, such as “What could we do better?” Software to analyze lots of text in lots of language is in constant evolution. At the time I write this, most software that is available on the market works extraordinarily badly.
The most common defect is the requirement for lots of human intervention, mainly to ‘tag’ answers, assigning them to an existing list of categories. As soon as more than one person is doing the work, the results are inconsistent at best, incomprehensible at worst. The second-most common defect is that a lot of the software returns single-word categories. For example, the most common word found in a survey about a restaurant might be ‘food’. But what about the food? Was it good or bad? Who knows? Great Natural Language Processing software groups unstructured data into themes; multi-word concepts, each of which groups answers that may use different words that have similar meanings.
The final skill you need is the ability to present compellingly. While I suggest a formal course on such skills, I don’t believe it is enough. I believe you have to be a storyteller. That will be the subject of another blog.
While the above is new content, I make related points in some our other books. All books are available in paperback and Kindle formats from Amazon stores worldwide.