Digital Conqueror - Logo
Search
Close this search box.

Using Watson Analytics for Customer Analysis & Effective Marketing

In past few months, we have covered How to use Watson Analytics platform in-depth, and you can relate the analysis much properly with following use case about using this platform for effective marketing and understanding your customer.
As a small business user or as a marketing manager at an enterprise, having an effective marketing strategy can make or break your whole campaign. And the effectiveness must be constantly measured and altered according to customer feedback. The customer behavior and how they drive to your business needs to be thoroughly understood in order to maximise the results. To eliminate the need for external or internal IT team, and get this analysis done quickly let’s use Watson Analytics for our use case here.
To get started, sign in to your Watson Analytics account or Sign up for free 30 days trial account here.
We are using campaigns data from customer survey by importing a local file. If you’re new to Watson Analytics, Learn more about how you can import and refine data using Watson Analytics.
Once the data is loaded, click on it to explore suggested data points. You can ask even ask the questions about your data using natural language, so let’s try out the break down of how our customers have responded to various offers we have put out in given data range.
As we can see here, a number of customers showing interest in available offers is significantly less. We need to identify what drives the customers to respond positively to our offer. So I will click on Yes, and right click and click ‘Keep’ to explore the customers who responded as Yes.
Before drilling this data further, I have renamed the discovery from ‘Untitled’ to ‘Customers by Response to offers’.
As we can see from this graph that offers 2 is most effective compared to others. Let’s find out the customers from which states have responded the most to offer 2.
As we can see here with TreeMap created by Watson Analytics that Maharastra is the top state of all followed by Karnataka. We’ll dig more details on what policy type got sold better in a different state, to achieve this I’m adding ‘Policy Type’ to our smart filtering.
This modified chart identifies ‘Personal Auto’ has outperformed every other policy type. Going down into the data, also showcases that ‘Personal 2’ and ‘Personal 3’ are among the most bought policies by these customers across states of Maharashtra and Karnataka states.
Here we are concluding the current discovery, but we need some more information to identify more information about our campaigns. So, now I will create a new tab here, and use one of the recommended starting points by Watson Analytics. I am selecting Income by State Map visualization. I am adding customer lifetime value and number of the customer to the smart charting, and it shows some insightful information about the concentration of customers across different states.
While Karnataka has highest customer value, Maharashtra has the highest number of customers. On the right smart visualization is showing us some interesting data about these customers, their vehicles and policy details.
By Clicking on top drivers of customer lifetime values, I will get more detailed data on what factor drives most customer value. This data in spiral visualization shows me factors as we strength of each factor affecting customer value. I can use this data to upsell more offers or cross-sell different offers to particular customers with the highest rate of returns.
Let’s dig deep into the ‘Vehicle Class and Number of Policies’ as its showing 65% of strength among the other factors. Clicking the ‘+’ icon gets me more data on this insight.
This visualization shows us that sweet spot for our campaign has been the intersection fo Luxury cars & Number of policies. These are very detailed insights into the campaign and same is hard to achieve with traditional tools like spreadsheets. With this experiment, we have identified the top performing offers, corresponding states, customer LTV and what drives the same in a granular manner. We can now create a display that combines all of these visualizations and shares it with respective campaign designers or marketing team.
I hope you’ve enjoyed reading our 6-part series on IBM Watson Analytics. Feel free to share if you have any queries about the platform and how it can be better used in your particular business in our comments section below. If you’re interested in IoT or Real-time analytics check out our article: Real-time Mobile Data Analysis using Watson Analytics.

Share:

Facebook
Twitter
Pinterest
LinkedIn