Data Product Revolution
In the last 15 years, there is a huge increase in Data Product development; e.g. Netflix Recommendation Engine, Spotify Playlist Curation, Google Maps Route Prediction, Autonomous Cars, etc. We can add hundreds or even thousands of great examples to this list. But why is there a revolution in this direction? Or let me ask the first question, what is a Data Product?
What is a Data Product?
In his book, “Data Jujitsu: The art of turning data into product, 2012”, DJ Patil defined a data product as “A product that facilitates an end goal through the use of data”.
Nowadays almost all mobile applications and websites use data but not necessarily all of them are considered data products. The main differentiation here is, the end goal of the data product is to use data to generate value. For example, Gmail is not a data product but the smart compose feature is a data product.
Data Product Examples
Today we are surrounded by Data Products but we are not even aware of them. We can’t differentiate them from the core products. Especially for non-technical people, it is almost impossible to detect. Here are some examples;
- Spotify Wrapped: Spotify is a leading music streaming platform. In 2016 they announced the collected data between January and December as a marketing campaign. The campaign has been recurring every year and is a personalized Data Product for individual users.
- Netflix Recommendation Engine: Netflix is a leading movie, tv show, and series streaming platform all over the world. Their major business model is to retain as many customers as they can over subscription models. In order to retain more customers, they need to improve the engagement of their users. The core Data Product that serves this requirement is their personalized recommendation engine.
- Gmail smart compose: Gmail allows users to compose emails in a plain text format. In 2018 Google introduced smart compose capability. This Data Product uses various machine learning techniques to generate the most appropriate sentences and words for the user in real-time to minimize e-mail creation efforts.
- Google Maps route predictions: Google Maps help you to move from X to Y location with various methods; e.g. public transportation, scooter, etc. For each process, the end user receives alternative routes with the predicted ETAs. In 2020 Google shared a glimpse of its algorithms and how it built this Data Product.
Hooked Model
In his book, “Hooked: How to Build Habit-Forming Products”, Nir Eyal describes the hooked model in four steps; Trigger, Action, Variable Reward, and Investment. The core of this model is to create a habit-building experience for the consumers so that they don’t need any nudge to use the product. Basically, users use the products because it is their habit to use it.
Why do we love Data Products even though we don’t know them?
Take a step back and think about the product you love and try to split the Core Product and the Data Product. I will take Spotify as an example.
I love listening to music. From the Walkman era till today, it has been a habit in my daily life. Whenever I have free time I listen to music; e.g. while cooking, doing laundry, and reading books. If you are someone like me, not a social media user, or not following any popular trends, it is becoming a little bit annoying to find out new artists and bands. Collecting vinyl and getting stuck at amazing 90s is also not a sustainable solution.
This is why I am a massive Spotify user. I am not using Spotify because of its fancy UI components or its huge archive. Similar components can be found in alternative products or even better. But I use Spotify because of its labeling, categorization, search engine, recommendation engine, and playlist curation capabilities. With these capabilities, I can be introduced to artists and bands that I have never even heard about or chance to find out with the limited time I have in my daily life.
So the Data Products make the Core Products smarter, easier to use, and more personalized and the users are keen to retain and make them a part of their daily routine. Simply converting normal products to habit-building products.
Before and After ChatGPT
Nowadays there is a huge hype about ChatGPT. Everyone is speaking about how amazing it is, and how fascinating the outputs it produces are, it is going to replace X, Y, and Z jobs in the next years.
We have to admit that the product they have built is amazing. I believe the main success of the product is not coming from the technical complexity or solving a life-saving problem in our lives. It is coming from its simplicity, accessibility, and easy-to-use features.
Conclusion
In the last 15 years, many companies invested in data products to make their core products smarter and personalized. With ChatGPT we are entering into an era in which the core product is the data product. I believe we are going to see many examples in this direction. Nevertheless, if it is a core product or making the core product better, the Data Product Revolution is inevitable.
Thank you for reading
In this article, I set up a common understanding of Data Product terminology. In the next series, I will deep dive into different kinds of data products, and how to set up cross-functional teams to build the products and nurture this culture in your organizations.
If you are interested in Data Science, Machine Learning, Product Management, and building Data Products, you can start following me.
