To get the best insights from Net Promoter surveys, you have to read your customers’ responses to the open-ended “What is the primary reason for the score you just gave us?” question. We call this “verbatim feedback” or just “verbatims”.

To turn your verbatims into actionable insights, you have to categorise the feedback so you can see what themes are being mentioned most by your customers. In that way you’ll learn what you are doing well that your customers love, and what you’re not doing well that exasperates them.

The most effective way of categorising verbatims is to do it manually in Excel.

Word cloud apps won’t cut it because they just look for word count frequency. You need a human to understand that…

“I wish you had solved my problem quicker”

… is the same thing as…

“You were too slow to fix my issue”

… and the opposite of

“You couldn’t have fixed my problem any quicker! Thank you!”

Machine learning/AI software is getting close to being able to do a pretty good job at identifying themes, but it’s not there yet. Until then, here’s you can do it in manually.

To identify themes, load your scores and verbatim feedback into Excel. You can do this using the Survey Responses report, specifying the format=csv parameter.

The two main columns in your spreadsheet would look like this:

You will also want to keep the other columns, e.g. the Team (Support Group) and Segment (Customer Segment). Later, this will enable you to understand where themes might apply to specific teams or customer segments.

So your table would look something like this:

To keep things simple, we won’t show those extra columns in the rest of this article.

Let’s add some more feedback to our example.

Before beginning to tag the feedback, remove feedback that is not specific enough to derive insights from. In this example, we'd ignore “You guys rock” and “The whole experience was terrible”.  Maybe stick “You guys rock” on a post-it note on the fridge ;)

As you read through the feedback, add columns for each theme that emerges. When that theme is mentioned, put a “1” in the corresponding column, like this:

Often, you’ll find one verbatim talks about multiple things. For example, “Sam was super friendly. I just wish my problem could have been solved quicker” refers to two separate things:

1. Sam was friendly.

2. The ticket took too long to solve.

In this case, add two columns to your spreadsheet, one for each theme and put a “1” in each.

Because you’ll be tracking themes that are both good and bad, make sure your good and bad themes are clearly differentiated. We like to call positive themes, “Reinforcement” themes. And the stuff we need to improve, we call “Corrective” themes.

This will make sure you don't mix-up easy-to-use vs hard-to-use ‘Portal Usability’ themes!

By the time you’ve tagged the remainder of your feedback, you’ll have something like this:

Identifying the themes that are most important to your customers is simply a case of counting the tags.  The tags with the highest counts show you where you’re strong and where you need to improve.

To really understand your strengths, count the Reinforce tags just for Promoters.  And to understand where you need to improve, count the Corrective tags for Detractors and Passives.

You might end up with something like this:

This gives you a pretty clear view of what you’re doing right (friendly staff, and your new portal) and what you need to do to create more Promoters (speed up resolution times, better communication on ticket statuses).

By using the additional columns in your spreadsheet (e.g. Team and Segment), you will be able to get even more detailed insights.

For example, if you knew that Sales had the lowest Net Promoter Score of all your Customer Segments, you could repeat the above analysis just using feedback from customers who work in Sales. The resulting analysis would then tell you which factors you have to improve to make Sales happier. Similarly, you could repeat the analysis for a single Team to learn what that particular team's strengths and weaknesses are.

Finally, although we’ve used feedback from a transactional survey in this example, the same principles apply if you were analysing feedback from Relationship Surveys, e.g. about the whole of IT (great for informing IT strategy) or about a single project (great for improving future project capability).

Once you’ve done your analysis you’ll know where you need to improve.

You can then prioritise your service improvement initiatives taking into account other relevant factors such as cost, resource availability and dependencies.

Importantly, don’t forget to communicate with your customers using whatever channels you have available to you.

We recommend you communicate three messages at three different times:

• Thanks for your feedback, here’s what we’ve learned from it (good and bad).
• Here’s what we’ve got planned as a result of your feedback (to build on the good and address the bad).
• Here’s what we’ve done as a result of your feedback (show you’ve delivered what you said you would).

Net Promoter surveys are customer friendly (super short) and deliver actionable insights when you mine the verbatims.  Although it can take a bit of time, it’s simple. Customer feedback is pure gold if you want to improve service. Your customers are like an army of free consultants. Go use them!