Data Analytics experiments with Tableau
Wouldn’t it be better if we validate our ideas and assumptions before we start building a product? It would save the company so much time and money to build a minimum lovable product (MLP) or minimum valuable probuct (MVP) [whatever is your mojo!] if you were sure that the customer wants it. With many analytics tools, you can have a scrappy approach to validate your ideas/assumptions with your customer. Many a times, the customers don’t know what they want. When they see a visual representation, there are many things they realize. We will look at Tableau today for the same.
I can’t share details of my experiments due to PHI reasons. But, I will create generic scenarios to explain the same. Let us assume you have a grocery store with multiple types of items (let’s say chocolates and cakes) and you want to know both product and business outcomes of the sales. You could create analytics with Tableau.
What are the pre-requisites for your experiments?
No prizes for guessing! - Data As of now, let us assume you have an internal portal, used by your employees to checkout the items for the customers. The data would be stored in the datastore eg: AWS DynamoDB. For analytics as your data volume increases, we need a Business Intelligence platform such as Snowflake. Usually, the timing of the population-level analytics might not be upto the second. If you are incharge of creating the analytics, then you can use Tableau desktop or Tableau Cloud.
Features of Tableau I found useful
You can find various types of charts here. Customers found the bar, line and pie chart the most intuitive.They liked the fact that they could filter by various categories. (eg: How many chocolates did they sell in Feb 2025? How many customers bought cakes?)
Another type of chart for novice users is Text table. The highlighted text aspect was useful.
Customer responses
Customers liked the fact that we started off with a draft and showed them the data. They realized a few things and 2 important ones were -
Better clarity on numbers - Many times there was bad data. They realized that they had to clean the internal portal’s validations so that the data was good in the tables. As an example, let’s say that your store has a discount code and you apply it for the incorrect customer. So, the wrong customer will get the discount.
Better actions based on numbers - Many customers realized that they are not using their employees effectively. So, they reassigned the employees effectively for better results.
Final Thoughts
If you have a developer create a data store then you can create visual representations even as a non-developer. So, when customers see the representations then they actually come up with suggestions that work for them. We could run quick experiments with them.
Depending on your industry, Tableau provides you a detailed report on how it has helped companies in that industry. As an example and someone working in healthcare , you can make use of Tableau features!