Search

If you torture the Data long enough, it will confess to anything..

Updated: Dec 21, 2021


Picture source: https://www.nrgcycles.com.au/

It was Ronald Coase who said data would literally speak when tortured. LOL.


For this project, I used data provided by Sprocket company to analyze their existing customer dataset and discover customer trends and behavior. I was then able to make recommendations to help the business optimize resource allocation for targeted marketing.


Step 1.

Like every dataset, I started the preliminary data exploration and identified ways to improve the quality of data received.

The importance of optimizing the quality of customer datasets cannot be underestimated. The better the quality of the dataset, the better chance you will be able to use it to drive company growth.”

I was also sure to document my observation and provide feedback to the client. The reason was simply to ensure that the Client (Sprocket) and I agree with the data.

Below is a quick glimpse at some of the data quality issues I found and corresponding strategies to mitigate them.

  1. Issue: Multiple columns have empty values in certain records

  2. Mitigant: A drop-down list (data validation) can prevent future occurrence as this will help capture all the required data

  3. Issue: Inconsistent values for the same attributes. e.g. the gender column has "female", "male", "f", "fe".

  4. Mitigant: A drop-down list (data validation) can prevent future occurrence as this will help capture all the required data

  5. Issue: Inconsistent data type for the same attributes

  6. Mitigant: Format field according to the data type. Having different data types for a given field make it difficult to interpret results later.

  7. Issue: Unclear Column Header

  8. Mitigant: Use clear column headers and proper formatting, to ensure that data is captured correctly.

Click here to see email to the client.

Step 2

I prepared a detailed approach for completing the analysis by understanding the data distributions, feature engineering, data transformations, modelling, results interpretation and reporting. How you may wonder? For example, I was able to transform the transaction dates into the 4 seasons experienced in Australia. I group age into categories as per the world population age structure constructed for the period 2000-2025.


Step 3

I used Tableau to group and visualize the datasets. To do this, I needed to answer the following questions

  • What are the trends in the underlying data?

  • Which customer segment has the highest customer value?

  • What do you propose should be Sprocket Central Pty Ltd ’s marketing and growth strategy?

  • What additional external datasets may be useful to obtain greater insights into customer preferences and propensity to purchase the products?

Step 4

The below recommendations were made to the client


  • Define the WIITFM (What is in it for me) for each wealth segment specifically the affluent and the high network segments. Are the products top of the range and the best buy for these groups?

  • Brand repositioning: Identify the main competitors in each wealth group. What advantage do they have over Sprocket and what would be the switching cost? How does lead generation work for the competitors? What are the things that remind the customers of why the Sprocket brand serves them?

  • Create a tailored marketing plan for each wealth segment. How does Sprocket plan to reach the customers? Using the 80-20 rule, identify and prioritize the wealth segment and products with the highest return on investment (ROI).

  • Create a go-to-market plan for the other products and segments. How can Sprocket tailor it campaigns and promotions and WIIFM to its advantage? What is the core message Sprocket is communicating with the customers and how is it different from competitors?

  • DAGA- Demographic And Gender Advantage. Create a target market using age & gender demographics to boost sales.

Click here to view the presentation deck to the client. 

Remember, if you torture data long enough, it will confess to anything ~Ronald Coase






22 views0 comments

Recent Posts

See All