Why is your data useless?

• 01 Intro
• 02 Customer Segmentation and Customer behavior
• 03 Business Optimization
• 04 Final thoughts

01 Intro

A very common article title that I read in newspapers or any traditional media is “Data is the new oil”, but when I read the articles they are constantly very disconnected from reality. These articles are filled with big words such as “Big Data”, “Machine Learning” and “Artificial Intelligence” but very seldom explain why data is important for a business and where and when data gets its value. Here I would like to share a few of my observations from my experience working in the Finance and Auto industries of why you (I am talking to you business owners) should care about data. You see, in contrast with money, data has no value at all…

Let me explain… What I mean is that Data is a business asset that gets its value from its use. Many companies are very proud of their CRM, databases and expensive servers but they lack specialized people and procedures to make use of their data beyond being a record of transactions. The thing we need to understand is that 99% of a company’s databases are actually just a byproduct of their softwares used in everyday work. Their main purpose is not to be explored but to be a historical record and facilitate people’s work, which is great! But when seen with the proper eyes and techniques, data then it works like a lantern in the night. Beyond exploiting current data, I also encourage you to become data oriented and make data generation part of your design process for new products and services. These are some ways to exploit your data and generate value from it.

02 Customer Segmentation and Customer behavior

Since companies and product diversity has increased immensely, customers have become very educated and judgmental when purchasing a product or a service. Therefore, customer understanding is a major task for any company. Customer segmentation, is the search for clusters or groups clients that are similar. Similar in what way? The starting point should be your demographics. Is there any similarity among your customers’ age? Or city of origin? How about sex distribution? Beyond demographics, can you group clients based on their purchases? How much did they spend? In which order did they acquire your products? Or maybe from how they learned about your business? The name of the game is finding out ways to group clients. This means that if a client is in a cluster, he is similar(in some way) to the people within the cluster, but very different from people from other clusters. Once you found your clusters and understand what are the fundamental characteristics of the cluster, then you can classify your clients earlier, and offer them the right products in the right manner. You can choose to focus your ad campaigns to convert the groups you are lacking or strengthen the groups that are more loyal, this is now a business decision but with the clientes clustered you now know the impact of your decision.

Useful algorithms: XGboost and K-means.

After understanding “Who is our client?” We should continue our research into analyzing customer behavior patterns. Which is understanding how customers interact with our products and services. This is a central step for the design of an experience as well as to improve your services. A textbook example is the one of a supermarket and analyzing which items are bought together. We can expect milk, bread and butter to be purchased together commonly, but you may find other interesting patterns such as diapers and beer being bought together (For real, look it up). This intel then may be exploited with discounts or rearranging the stores. Another example example for analyzing customer behavior is for an optimal design of the digital experience. This can be done by doing a careful observation of how users navigate your website. For instance how many clicks users do on average to get from the home page to where they want to go, then fix your website so they can get there faster, which impacts your sales funnel directly. You can design and optimize your experiences from observing your clients through the magnifying glass of data

Useful algorithms: Apriori, K-means, XGboost.

03 Business Optimization

If we learned something from Henry Ford, it is how a small change can improve your assembly line immensely. Reducing 1 minute to make a car doesn’t change much, when you build 1000 cars a day, that adds up very quickly! Making your company data oriented can help you optimize all kinds of processes regardless of which industry you are in. Logistics, stocks, distribution, energy consumption, employees bathroom time…(Forget the last one). Anything can be optimized when you have the data available. With the data in hand, and the eyes of a Data Scientist, you can create models to predict, anticipate, and optimize any pipeline and you’ll see the results adding up at the end of the month.

04 Final thoughts

We live in a wonderful time where technology can change our businesses in ways we couldn’t imagine before. A proof for this is how everyday a new startup is created from just an idea plus some lines of code and then make any well established industry tremble. We’ve been using data in businesses to improve sales and profits for a long time, but now the way we interact with it has changed and it has to change even more! We need to think of data as an asset for the company. For every product or service we design, we have to ask the question: What data can it generate for us? If no useful data is generated from your new product or service, then it’s not finished and should fix that before launching. Change your mindset and become data oriented.