Customer Familiarity can be defined in one of two ways: how much a customer knows about a company’s processes or how well a company knows their customers. Although it brings certain issues of loyalty, commitment, and retention with it, how well the customer knows how to do business with the organization only matters when a company wants to change a process – and even then, the companies that worry about changing their processes know how to deal with it (for the most part, they probably did it already).
A company knowing their customers is a far more interesting – and complex topic. As the next step in my series on how to leverage analytics in an organization, I want to explore a little bit about how and what we learn about our customers.
At the beginning we had customers that simply bought our products or services. And we did not care much about whom they were, we simply produced our products in the best possible way and then we sold them in the market. At the very best, and not in all cases, we retained simple identifying information on customers – name, address, phone numbers, product they bought, and not much more.
We never used that information – OK, sometimes we used it for service and warranty purposes – and the value of knowing that Joe Smith bought a SNW-10 widget was not of great help for much other than tracking inventory and figuring which one sold better than the others. Reporting, if any, was usually just a summary of products sold, or some other identifying information such as 20 people from Reno, NV bought the product last month.
Someone came up with the idea of building profiles for customers. The idea was that if a 29-year-old, white male, Cadillac driver, golf player, who lived in Southern California liked SNW-20, then all other people like him were potential buyers for the same product. The theory goes that buyers often buy what their peers buy – whether they know their peers or what they bought, or not.
We set out to capture as much profiling information as possible, we called it demographical data or demographics, and we used it to segment customers in many different ways. We then used those segments to create marketing and sales programs, to create service solutions – and to report. Reporting was done by a process called cross-tabbing, or selecting certain demographical data and cross-referencing it against something else (e.g. 29-year-old from Saint Louis, who rents an apartment).
CRM systems were born and the idea was that profile data we had could be mixed with transactional information, collected by CRM systems, and we could build what was then called the 360-degree view of the customer, or a total picture of them. Using transactional data we could both learn what customers wanted and predict what they needed – in theory.
The era of CRM saw the birth of customer analytics, as we try to discern from all the information we collected what was wheat and what was chafe (i.e. what was worth using for profiling and predictions, and what was – well, just stored). By using that analysis we learned that transactional and demographical data could not provide a complete picture – it lacked data on the intentions of the customer.
Enter Surveys and Enterprise Feedback Management. Organizations began to use surveys, first focused on customer satisfaction only, then more on intentions, needs, and wants. These feedback events, consolidated into an EFM system that integrated with the demographical and transactional data collected in CRM systems, focused on Behavioral (what they did) and attitudinal (what they wanted to do) data from customers.
Reporting began to change. Since we had analytics we began to create insights instead of reports. Analyzing the data and finding patterns and trends in them gave us the ability not only to predict better what customers may want and need (inferred from the insights), but also allow organizations to understand better customers behaviors and model experiences behind them.
Later, when the use of open-edit or text-box questions was added to the surveys, we saw the data collected was very valuable. Now customers could express, freely, how they felt and what they wanted even if they were not asked directly in the survey. The amount of data collected, in unstructured form, became a treasure of sorts, where companies could read and learn more about their customers than they ever could before. This data is still today critical to all efforts across the organization to learn more about the customer.
The last era of customer familiarity came with the social evolution we are currently experiencing. Sentimental data, or sentiment analysis, began to crop up. One thing we did learn from doing surveys is that customers tend to answer them in the way they think the organization wants them to be answered. I wrote about a method you can use to ensure a 90% customer satisfaction, for example, that works in real life.
Analyzing the sentiments in addition to the behavioral and attitudinal data yields a true view into the customer’s mind. Knowing their state of mind when they decide what to do, or what they will do, is an incredible insight that organizations can use to improve their processes and their business. It does not replace any of the other data, it complements it.
There are two things that happened in this evolution, and they are very different:
- The quantity and complexity of the data increases as organizations leverage analytics (see my previous post on how to do that well) to find more valuable insights and they use those to build better experiences, better products, better services and – well, better businesses.
- The detailed personalization of each customer diminishes in favor of a community of customers with similar likes and dislikes, needs and wants, and similar profiles. Having a community makes it easier to assign specific attributes to it and leverage it for analytics .
So, now we are left with loads of useful, but unstructured, complex information, better profiling of less personalized customers, and some insights – what is the next step? More analytics.
Parsing the edit-boxes and comments collected, creating structured data models from them, analyze them for behavioral, attitudinal, and sentimental data structuring the unstructured, and use those new data models and insights to improve products and services.
What do you think? What did I miss? Do you see this as the most interesting turn of events yet? Let me know your thoughts…
This post is the second of a series of sponsored research posts I am doing with Attensity on the use of Analytics in CRM.
