From Satisfaction Drivers to Sentiment Analysis: A Game Changer in Customer Insights

In the rapidly evolving landscape of customer feedback analysis, the traditional approach of using satisfaction drivers to categorise feedback is making way for a more advanced and insightful method: sentiment analysis.

As a CX specialist with a decade-long journey working with e-commerce companies, I have experienced this transformation firsthand.

In this article, I highlight why sentiment analysis is now my go-to method with all my clients.

If you're still using satisfaction drivers in your business, I'm sure you'll find this article useful!

?

The Era of Satisfaction Drivers

When I was working at Zalando about 10 years ago, satisfaction drivers were the new and hot method for extracting customer insights.

Asking specific questions about the different parts of one’s experience with the company, not just the NPS or CSat score, gave us a much better understanding of the pain points along the customer journey.

This method provided valuable quantification of the feedback, which we wouldn’t get otherwise, keeping in mind we were dealing with millions of survey answers each year.

This is what the survey would look like from the point of view of the customer (note: this is not a Zalando survey, just an example I've seen lately, it doesn't matter that it's in German, the text isn't important here):

Looking at the screenshot above, you can see this survey contains many questions, with answers mostly quantitative (not satisfied = 1, satisfied = 5).

My main problems with those types of surveys using satisfaction drivers are:

1. all the questions are created by the organisation, meaning that we are influencing the customers in telling us what we want to hear, not what customers want to tell us. Was this experience really important to the customer? We don't know!
2. As it requires the company to ask many questions, it's just too long to anwer! The cognitive effort required from customers to answer is too high, resulting either in abandoning the survey half-way, which lowers response rates, or ticking quickly each answer without thinking, which lowers the reliability of the insights generated.
3. Because we ask everything to the customer, we're not understanding what experiences were truly important to them, and what didn't matter at all. Not having this information make it impossible to prioritise actions, as they all carry the same weight in the eyes of the organisation.
4. Those surveys don't function well over time: new needs and problems will arise that are not captured by the survey. This would require continuously listening to changes from a different channel, and constantly update the surveys by adding more questions.

Now let's see how sentiment analysis solves those problems.

Sentiment Analysis: A New Horizon

Fast forward to 2023, the year of artificial intelligence!

Even though to be honest natural language processing and machine learning tools have been around from some time already.

These technologies have revolutionized how we work with customer feedback.

About five years ago, when I founded CXMania to help more businesses scale their business, I took the opportunity to make the switch to sentiment analysis.

Here is what the survey looks like using sentiment analysis:

If you're thinking "is that it?", you're right!

Just two questions, the score and a text-field for customers to tell us more about their experience.

It's simple, it's qualitative and it solves all problems listed above!

How Sentiment Analysis Works

Simply put, sentiment analysis splits all qualitative feedback (a customer’s comment from your survey, your emails, any public feedback) into multiple pieces of content.

For each piece of content, it can associate a positive, neutral or negative emotion and categorise the content in multiple levels, as per the example in this screenshot:

You can see how the tool is splitting the content.

For instance, a comment like "the food was great" is categorised into "food/quality" and "great" indicating a positive sentiment.

By automatically categorizing and quantifying feedback, sentiment analysis offers a seamless transition from qualitative insights to quantitative data, which will be crucial for everyone within the organisation to monitor as we’ll explain later.

Imagine now you've had thousands of feedback like the previous one, the tool would then quantify the insights generated, like this:

Easy, isn't it?!

Here are the main reasons why sentiment analysis has become so valuable to me and my clients:

?

1. Understand what's really important to your customers

By letting customers speak their mind without any influence from the organisation, they can express themselves freely and provide qualitative insights, making it possible to understand their true sentiments and categorize their feedback accurately.

If no customer talked about the speed of loading of your home page, that's probably because they didn't care, otherwise they would have probably bounced and gone to another website. So don't ask that question.

Sentiment analysis highlights what's important to your customer, opening to the possibility of prioritising by importance and satisfaction, which couldn't be done using satisfaction drivers.

?

2. Increasing the share of valuable comments to generate more reliable insights

You've probably heard it already: "surveys are dead, no one wants to answer them anymore!"

"Survey fatigue is a real pain point."

Well... partly!

If you can stop asking 50 questions that no customer cares about (back to the website loading speed question), and only ask 2 questions, then survey fatigue would disappear.

We regularly see surveys reaching 25-30% comments rates.

To be precise, this ratio means that for 100 surveys sent, we get about 40 responses, and 70% of those responses would have the text-field completed; that's here 28% of valuable feedback per survey, just with one survey.

Imagine if we have 4-5 surveys along the entire journey how much valuble feedback we can get, without bothering the customers (they would only receive one of the 5 surveys, in order not to spam them).

?

3. Getting to the Root Cause of a Pain Point

?While satisfaction drivers provide scores and surface-level information, they often leave organizations guessing about the root causes of customer issues.

Sentiment analysis, on the other hand, directly connects qualitative feedback to quantifiable insights. This means organizations can pinpoint exactly where to improve, by first looking at importance and satisfaction, then going deeper into the feedback left by the customers.

I’ll share more next time about how this works, for you to see how easy and efficient we can go from insights into action, and how those features transform the role of the CX insights manager from a data miner to an project facilitator within the organisation.

?Stay tuned for that one :)

Conclusion

Sentiment analysis has become an indispensable tool for understanding customer feedback.

Allowing customers to share their thoughts without bias, increasing response rates and the quality of answers, and uncovering the root causes of problems in a few clicks are just a few reasons why sentiment analysis is a game changer.

As a consultant and CX Insights manager, I can now swiftly analyse and share insights with all teams, speeding up the prioritisation process and enabling faster and more targeted improvements.

?

Sentiment analysis is a must-have for any organization looking to truly understand its customers and improve its products and services.

It's time to let your customers tell you what they want to tell you, let the insights guide you and watch customer satisfaction rise.

Thank you for reading!