There is a lot of buzz around Artificial Intelligence and Deep Learning. Companies are looking for ways to leverage the power of analytics and machine learning in their products. In this blog, I am not going to focus on the benefits of AI, or elaborate on how AI is impacting our lives. Instead, I am going to explore the imminent question that every product owner is facing today — how do I incorporate AI into my product roadmap? Where do I start? How do I go about?
Here is a typical scenario. You are the product manager for an online SaaS product with a healthy growth curve. One day, the company executive walks to you and says — “Let’s incorporate Artificial Intelligence and Machine Learning in our product. Could you put together a plan for that?”. Although you have been thinking, exploring, and educating yourself on recent trends in data science, you have not seriously ventured into a concrete idea of how to bring AI to your product. Suddenly, now your research and exploration has a deadline and you have to come up with the plan or a plan for the plan! You are the owner of the product! it is your responsibility to put together a roadmap that incorporates AI while creating value and generating revenue.
But wait… is your product even ready for an AI strategy?
The answer to this question depends on what stage your product is. If you are in the ideation or development stage, then you need to brainstorm on how you can leverage data into your product. You are not looking for AI strategy but looking for an AI product. On the other hand, if your product has millions of customers, then you may have a lot of data that is being collected and recorded already. You need to dive deeper in the pool of data, leverage your domain knowledge and explore ways of using collective customer data to create value.
Let’s say you are looking at a product that has millions of customers. Where do you start now?
Here are some places to begin:
- Database schema — Most products store their data in a backend database. Get access to the database schema to get a holistic view on what data is stored. The cross relationship between the data items often times reveal interesting patterns that could be extracted, interpreted, and extrapolated to generate visual presentation to the user.
- Product Logs — The product logs typically captures information about timestamp, sequence of actions, queries that are run, results extracted, and performance metrics. Product logs would provide a deeper insights on user actions and their effect on data processing.
- Marketing Analytics — The marketing data, click analytics and user interaction would provide a load of information about the customers thought process when they are accessing your product.
- CRM analytics — Once the product is launched, the customers are the driving factor of your roadmap. As a good PM, you need to understand your customers persona. What best way then the CRM? It would provide rich data about customer’s profile, industry type, buying power, location, size, and branding. Analyzing your customers purchasing pattern would reveal a lot about the market trend.
- Chatbot and support tickets — The customer tickets and online chat data is a great place to retrieve information about user behavior, feature request, access patterns, and more.
The goal is to analyze the data from all these sources to formulate a useful and meaningful question. If you already have a question, then you need to find out how you can extract or predict the answer from the data.
Data Science was a “hype” a couple of years back. But now it is for real and if you don’t have a strategy around it, then you will be left behind. Customer’s data is the next valuable thing after revenue. It is up to you to find ways to convert the collective customer’s data into a value proposition that in turn could generate more revenue directly or indirectly!
In the next few blogs, I am going to brainstorm ideas on how to build a data strategy for your product. I would be more specifically looking at SaaS product but the the concepts are generic and would be applicable to any product. I would like to get your ideas as well — please feel free to share your thoughts in the comment section.
