Home Alone (Parts I – IV)
I find it hard to imagine that anyone out there hasn’t seen at least one of the movies from the Home Alone franchise. Just in case, the basic plot revolves around a child being left alone to fend for themselves and fights off criminals lurking in the neighborhood. Despite his high profile reputation within the family, when things begin to get hectic, he is forgotten.

Analytics, as a part of a CX program, feels very much the same. Everyone knows it is important, it screams for attention, yet is often forgotten in the rush of implementation – and the first time someone notices, the options are limited.

Make a Plan for Analytics

When preparing to roll out a new CX program or refreshing an existing program, an analytics plan should come ahead of a roll-out. I’ve seen this get confused with “we’ll let every department add a question” or “what will my dashboards look like?”. This type of thinking tends to result in a trend of declining response rates and limiting the program’s potential. With this approach, by the time you reach the phase where you can start analyzing data, you realize three things:

  1. You have some great looking dashboards – while this may seem like a success, it isn’t what will drive the success of the program within the management of the organization. If the numbers seem relevant, the dashboard’s consumers will be satisfied with having a way to track these numbers, they will now be measured on with some fancy charts or widgets. However, this will only inform them of those CX measures and will ultimately lose their usefulness. Dashboards – no matter how fancy and detailed – is not the same as analytics.
  2. You’ll have “peacock measures” – Getting every department to add their input will result in a very long survey and will ultimately detract from the customer experience through the very tool you are trying to use to improve it. As a result, a CX manager will work with departments to find the department’s one or two critical questions. To avoid being the department with the bad score, they will try to game the question, so it makes the department look favorably on the dashboard being developed. No one wants to be the department with the only bad score. It will never be as obvious as “Which days of the week do we provide great service? Those that end in ‘-day’ or ‘other’?”, but it is targeted to solicit a specific type of response.
  3. You can’t tie results to business measures – When the outcomes are predetermined, your analytics will show no true relationship to business performance. Sure, there will be some correlation between the information displayed on your dashboard and success, but the predictive nature of the measurement will not be there. It will lead to many data mining dead ends and limit what you can do analytically with the data.  

While your plan will not be perfect, it will help make an efficient survey experience and provide the opportunity to connect to the business questions. I’m often caught saying that analytics don’t come from analyzing the data you collect during the survey rather it comes from connecting that data to your business. It comes from understanding causation instead of correlation. It comes from analyzing churn risk, then comparing it to company churn. If we’re brave enough to admit it, we may find that the plan didn’t work for some areas, but knowing the plan gives us a path to adjusting our plan forward.  

Don’t Let Analytics Be Your Kevin

At the beginning of the first Home Alone movie, the daughter is asked to count children. Without the experience in handling the “McCallister Family Organization”, the daughter mistakenly counts a neighborhood kid instead of Kevin. That creates a problem in their plan immediately before they’ve even left for the airport. As you begin your plan, it is easy to ask your vendor or CX partner to develop an analytics plan for you, however, you’ll end up getting the same outputs as asking a teenager to develop it. The plan will be optimized to fit into their framework [avoiding too much work] and can leave you scrambling after the CX journey has begun into finding ways to fix the situation. You may end up with some very good looking dashboards though.  

Instead of building the analytics plan, a good CX partner will work with you to develop the plan based on business needs and within the framework of their approach – that is the compromise you should be looking for. The plan will allow for questions that tie to business measures such as customer churn, per-customer revenue and profitability. All your questions in the CX survey should revolve around those goals with the limited space and time you have. The CX journey will be a great one when you have a plan, contingencies, and a framework to guide you. If you leave too much to others, you may find yourself on an airplane having left a very vocal child – that I’ll call analytics – home alone.