The future for data analysis is learning how to create insightful conversations.

It’s true that this is partially possible, but I think it is worth understanding that there has been a real paradigm shift in how we engage with data and how we create the “insight” that business executives need. The opportunity today goes far beyond just predicting which product a customer may want to buy next. Think about the traditional data analysis environment. We usually had to find the data first as it was often scattered across different divisions within the business - different managers owning the data and different physical databases. Then we had to clean it up and try connecting it - creating keys or other links to connect various tables of data. Then we needed data scientists to query and engage with the data. They understood languages and tools like SQL, which allowed them to explore different questions and scenarios using the data they had. Within the customer service environment, there would only really be two types of data that could be queried - KPIs coming from the performance of the contact center team and the activity of the customers, their purchase history, for example. So the data analysis team could explore when a particular product is selling more and maybe target some special offers to customers that are likely to be interested in it, but we always needed people with the knowledge of how to extract these insights from the data - and that is after the data had to be cleaned up and prepared for analysis. Spotify gives a good example of how data analysis has changed. Imagine that I was an up-and-coming singer-songwriter from Stockholm. I had toured locally, and I released my first album and a couple of partnerships with other artists. How would I go about planning my first tour away from my home region? I can explore the Spotify for Artists service to see where people are listening to my music, which songs they like most, how often they are listening, and their demographic profile. This can help to plan the tour locations more easily, but once I announce some new tour dates, the Spotify app will also ensure that the people listening to my music in those locations are notified that a show is coming soon. British rock band Iron Maiden realized the value of thinking about data as a decision-making tool over a decade ago. In 2013, their manager informed the band that they were one of the biggest victims of pirate streams using BitTorrent in Brazil. This sounded like bad news - their music was being stolen, and the band was losing out. Instead, they started booking more and more live shows in Brazil. The data said that Brazil was the place where their music was stolen more than any other artist, so they decided to go and give the fans something that cannot be stolen - a real live concert experience and merchandise only available from the shows. Perhaps the key difference today - comparing the Iron Maiden experience to Spotify now - is that an artist can do this without understanding how to query a database or clean up the data. Anyone can now explore the data around their business activity just by asking questions. In many cases, this can literally be a conversation - look at how easy it is to have a typed or spoken conversation with ChatGPT today. The future of data analysis is who can ask the right questions. The data is out there and can be queried with ad hoc conversational requests. You can ask an AI assistant to highlight any unusual behavior or insight, but the real insight will come from smart people with a detailed understanding of your business and industry and the ability to engage creatively with the data.

The business magazine Forbes recently asked if AI will replace the work of the data analyst. The assumption that some managers have is that they should just be able to start asking “How do I improve my customer service?” or “How do I get customers to buy more products during each visit to our website?”

This is where a lot of management misunderstandings around AI still exist. The data is becoming easier to use, easier to access, and easier to query, but the questions still need to be based on specific problems that need to be solved or strategic issues that require critical thinking. Anyone who has learned how to code a computer will understand the problem. Whenever you tell a computer what you want it to do, you have to be very explicit because there is no room for assumption or interpretation. We are now entering a time when AI can help us to interact far more easily with data, but there is still a set of skills required to get the most out of our interactions with the data. The technical data analyst with coding and query language abilities has to evolve into someone with a creative insight into what the business wants to achieve and how the data may create insight into these aspirations.

Don’t ask “How can I improve the customer service process in my business?” do ask “What are the top three most common customer complaints or problems, and what are the triggers that cause them?” 

Then continue engaging in that conversation once the first response comes back…  We are in a new era of data analysis today, and the best analysts will be those who can create the best conversations - not the best database queries.