Understanding our Consumers to Drive Improvement in Tokopedia
Have you ever filled in a survey asking:
“Seberapa ingin Anda merekomendasikan Tokopedia kepada teman atau keluarga”?
If the answer is yes, then are you wondering what we will do after we get the answer?
Firstly, let me introduce myself, my name is Tika. I am a Data Analyst for Operational Tribe at Tokopedia. The Operational Tribe is responsible for interacting with consumers which plays an important role in one of Tokopedia’s DNA, Focus on Consumer.
To make it happen, we collaborate with Customer Experience team to conduct a research in order to understand more about our consumers’ needs based on their experience, so we can learn how to improve our products or services. One of the methods is Net Promoter Score (NPS).
What is NPS?
Net Promoter Score, or NPS is a survey that measures the willingness of your customers to recommend your company, product or services to others based on their overall experiences. The typical question of this survey is
“How likely are you to recommend our product to your friends or family?”
and why you gave that score.
As the user, you will answer the survey on a scale of 0 to 10. The lower the score, the lower your eagerness to recommend the product to others. The score will then be grouped into three categories: Detractors, Passives, and Promoters.
And the NPS Calculation is
Why is Tokopedia doing NPS research?
With “Focus on consumer” and “Make it happen, make it better” as two of Tokopedia’s core DNAs, here are the top three reasons on why NPS research does matter to the company:
- To know how likely our consumers are to recommend Tokopedia to their friends or family
- To precisely understand consumers’ needs based on their personal experience
- To improve our product or service based on consumers’ feedback
How do we conduct and gain consumer insights from our NPS research?
There are some steps we need to follow in doing NPS research, starting from compiling and conducting a survey, building dashboards to monitor responses, until giving a recommendation on how to improve the products or services. In this part, I’d like to share how NPS research is conducted and how valuable insights are produced by integrating the survey result with analysis from internal data. Note that all data in this article are dummy data.
1. Build the survey and define the criteria of respondent
The first step of NPS research is to design and conduct an NPS survey. Understanding the objectives of the research is the key at this point, as the survey will be considered to be good if it satisfies all the objectives.
Based on our experience, here are some basic tips to create a good survey
- Always focus on the objectives of survey
- Ask simple and to-the-point questions
- Make sure the questions are easy to understand
- Avoid misleading or biased questions
Another important thing in conducting the survey is the criteria of respondents. We need to make sure that we select those who will be appropriate for the goal of the research.
2. Categorize consumers feedback using text analytics
In order to handle the number of responses of NPS survey that reach thousands per month, we apply text analytics by using n-gram analytics and then visualize the result to get insights quickly.
The result of n-gram analytics also helps us define categorization rules. The categorization is divided by products and services in Tokopedia, so we can focus on product or service that needs improvement most.
3. Validate consumers’ feedback with data
From that survey, we always focus on consumers’ feedback and how to make it happen (if needed) or make it better. But before that, as a part of a data-driven company, Data Analyst always validates consumers’ feedback with data.
For example, we heard some feedbacks from users relating to an error occurred when paying their credit card bills. Instead of directly executing the idea, we initially validate on whether or not the issue does exist.
From the above chart, it is clear that credit card bill payment had the lowest success rate from all digital products. So the feedback is valid.
4. Brainstorming with product or business unit
After feedback validation, we elaborate the issue with the related product and business unit to understand business process and identify the problems. In context with the same use case above, we discussed the credit card bill payment issues with the digital unit stakeholders. One of the takeaways we learned was an increase of complaint tickets on Operations was caused by failure of promo usage. The next fact that we learned was an increasing number of promo within the credit card bill payment transaction.
To limit and focus on the root cause, we create a hypothesis that states
“an increase in frequency of promo causes an overload within the payment aggregator which eventually detracts transaction success rate.”
5. Further analysis
Then the next step is to do further analysis based on the elaborate results with the stakeholders. There are many ways to do further analysis such as deep-dive analysis, hypothesis testing by using statistical method, competitive analysis, qualitative research, etc. Deep dive analysis and hypothesis testing can be done with internal data while other methods needed more external data.
In this case, we do deep dive analysis using our internal database — the relation of occurring promo and credit card bill payment success rate within the same month. To understand more about the situation and also get insightful results, we need to visualize the data.
From the visualization above, we can see the success rate of credit card bill payment transaction using promo has lower compared to the one who checkout without promo. After some observations from internal data, there exist some pending transaction in payment aggregator which lead to a root cause that the aggregator can’t handle the demand from Tokopedia. Then the final result is presented to stakeholders for discussion and improvements.
To sum it up, in collaboration with Customer Experience research team, we are able to understand more about our consumers’ needs through NPS research and drive improvements to related products or services.
So guys, if you want us to fulfill your needs, just fill out our survey and contribute to our improvement because your voices matter a lot!
