One of our most frequently asked questions and areas we train our users the most in is how to identify and investigate the themes that are impacting their core business metrics like NPS, CSAT, App Reviews, etc.

Today we're going to share some best practice tips around how to identify and explore high-impact themes using your Kapiche Dashboard!

In this handbook:

  1. Knowing what types of questions your data is capable of answering

  2. Getting to know your different Dashboard Tools

  3. Benchmarking Themes against core business metrics

  4. Drilling down into a specific theme

  5. Using Observed vs Expected to create actionable insights

  6. [Bonus] Tips for creating a killer insights deck


Know what types of questions your data is capable of answering

Before diving in, we recommend sanity checking that the feedback you're analyzing is actually capable of answering the questions you've been tasked with answering.

For example, if you need to know what is having a negative impact on your key Business Metric (like NPS or CSAT) but the survey question you're analyzing asks customers "How was your checkout experience?" vs "Why did you give us that score?" then the responses will generate insights specific to the checkout experience.

Don't worry...if you find yourself in this position, it isn't the end of the world!

You can still draw super meaningful insights out of your customer feedback even when the question being asked isn't directly tied to your business metric, you just may need to slightly adjust the expectations of the team you're reporting to.

For example, you might not be able to say "The main reasons our customers are giving us low scores is XYZ" but you'll be able to surface insights like "Customers talking about Payment Failures during the Checkout Experience are more likely to be Detractors".

In addition, this may also be a great opportunity to chat to the relevant folks internally about whether the survey questions need to be adjusted to be more aligned to the key business questions you need to answer!


Familiarize yourself with your Kapiche Dashboard

Your Kapiche Dashboard has a lot of different tools that are there to help you identify and explore your different theme drivers but it can be a little overwhelming when you're just getting started.

We have a dedicated handbook for getting to know your Dashboard but we've also popped a summary below of the tools we use the most when wanting to understand how different themes are impacting key business metrics like NPS & CSAT!

Date Ranges

You can apply a date range if you're only wanting to look at a specific time frame - by default your Dashboard will be set to show all data!

Dashboard Filters

Dashboard filters are perfect if you only want to look at how themes are impacting specific segments of data, like a customer segment (e.g. "Feedback only from those aged 18 to 24") or by things like NPS Category (e.g. "Show me the most common themes among Detractors").

Tip: To clear the Segment filters you’ve applied, simply click on the Segment Filter dropdown, choose the ‘Clear' button, then hit 'Apply'.

Saved Queries & Concepts

Your Saved Queries & Concepts tool is a summary of both the key Themes (Queries) your team has identified for you when analyzing the open ended feedback in a specific data set as well as the key Concepts Kapiche has identified when analyzing the data.

By default we'll show you each Theme sorted by frequency but you can use the relevant dropdown menus to toggle between Themes and Concepts and to sort based on key business metrics (like NPS, CSAT, Revenue, etc.) etc.

Learn more about your Saved Queries & Concepts tool here!

Quadrant Chart

Use the Quadrant Chart to gain incredibly quick high level insights into how specific themes and customer segments are impacting your core business metrics like NPS, CSAT, App Reviews, etc.

The Y axis represents Frequency, the X Axis Represents your chosen metric and the vertical line represents the overall score for that metric.

For themes having a negative impact on your business metric, you're looking for themes that are higher on the Y axis and to the left of your Overall Score - these are the themes that are high in frequency and have a lower score than your overall.

For more tips, check out the Quadrant Chart handbook.

Timeline

Your timeline chart is a great way to discover trends and monitor key business metrics over time. By default, we'll show you the frequency of your themes over time, but you can use the drop-down menus along the top to select different data points, metrics, and time periods as well!

To view how your themes are impacting your key business metrics over a period of time, simply change the Display dropdown from Coverage % to one of your business metrics and from there you'll be able to see how each theme is impacting your business metrics over time.

For more tips, check out the Timeline Chart handbook!

Drilling down into a specific query

Once you've found some themes that you'd like to explore (e.g. perhaps they're having the greatest negative -- or positive -- impact on your overall NPS) you can click into it via the Saved Queries & Concepts widget which will take you to a dedicated page for that theme.

On this page you'll be able to see:

  • General stats for the theme (e.g. how many records belong to it)

  • The overall NPS & Sentiment breakdown for that theme

  • A timeline chart to show how the theme is trending over time

  • A verbatim summary for all verbatims that belong to the theme

  • The Observed vs Expected chart to highlight outlier segments

  • A context network to highlight related terms customers use when talking about this theme

The team at Kapiche primarily uses the Query Drill-down page to highlight any potential segment outliers in the data by looking at which customer segments belong to specific themes more than others (more on this below) and to surface related verbatims that help humanize the data.

Learn more about your Query Drilldown page here!

Observed vs Expected (Segmentation Chart)

Perhaps one of the most powerful tools on your Kapiche Dashboard is the Observed vs Expected chart which allows you to identify key segment outliers in your data by comparing the Expected Frequency (the distribution of a segment across the entire data set) with the Observed Frequency (the distribution of a segment when looking at a specific section of data, like a theme).

To avoid doubling up in today's piece, we've gone into a lot more detail about the Observed vs Expected chart down below (because it's that important!) but we also recommend checking out our dedicated handbook and video here!

Emergent Concepts

This tool is super handy if you're doing longitudinal analyses as it's designed to show you the concepts that have changed the most during your chosen time period (the default is the past 30 days).

To identify Emerging Concepts, Kapiche looks at all of the concepts that existed in your data prior to your chosen time frame (it excludes the data within your chosen time frame) and then it looks at all of the concepts that exist across the entire data set (up to and including the data in your your frame).

From there it can see identify any key changes that have appeared in the data based on your chosen timeframe (e.g. what concepts are new, what concepts are increasing in frequency and what concepts are decreasing in frequency).

You can learn more about Emergent Concepts here.


Benchmark your Themes against core business metrics

So now that you know what types of questions your data is capable of answering and you have good working knowledge of the different tools available to you on the Kapiche Dashboard, you're ready to start benchmarking your Themes against your core business metrics.

At Kapiche, we typically start by changing the Display view of the Saved Queries & Concepts tool from Frequency (the default view) to the business metric we want to benchmark against, like NPS:

By default Kapiche will order Themes by the highest NPS to the lowest NPS but if we're wanting to look at which Themes are impacting our metric the most, we'll choose to sort by Impact instead:

Tip: Click the little arrows to toggle between Highest - Lowest and vice versa!

We'll also take a look at the Quadrant Chart which provides us with a similar level of insight into how each theme is preforming against core business metrics, just in a more visual way (which is also a fan-favourite chart in insights reports!)

From here you can easily identify which themes are having the greatest Positive or greatest Negative Impact on NPS / your chosen business metric (where impact means that if all of the records that belong to that theme were removed from the data set, your metric would increase or decrease by that amount).

These two charts alone provide you with high level insights into what your key Theme Drivers are! 🤯

But let's go deeper...


Drill down into a specific theme

Once we've benchmarked our Themes against our core business metrics we need to decide which themes we want to explore further and report on.

Our typical workflow is to pick the Top 5 positive themes that are increasing our business metric (the things we want to ensure we keep doing well!) and the Top 5 negative themes that are decreasing our business metric (the things we want to fix, improve or avoid doing).

We'll also generally try to avoid choosing high level/general themes like Customer Service and instead focus on the more specific Sub Themes (like Customer Service - Wait Time, Customer Service - Knowledge, etc.) as these are the themes that tend to lead to more actionable insights!

Once you know which Themes you're going to explore further, you can click on them individually to be taken to the dedicated Query Drilldown Page we spoke about a bit earlier which you can use to surface things like related verbatims to help humanize the data and where you can also access the Observed vs Expected chart to highlight key segment outliers (more on this in the next tip below!)


Use Observed vs Expected to create actionable insights

As mentioned earlier, the Observed vs Expected chart is one of the most powerful tools you have access to in Kapiche as it enables you to identify key segment outliers in your data, allowing you to turn high level data points into much more specific and actionable insights!

For example, lets say we've identified that Product Freshness with our Vegetables is the theme that's having the biggest Negative Impact on our company's NPS - this is a great insight to have at a surface level but we know that it would be impossible for our company to go away and improve this across every Store in every State all at once.

However, if we can identify that there are specific Customer Segments who are experiencing Product Freshness more than others (e.g. specific stores, customer touch points, age groups, etc.) and that it's these segments that are the ones who are bringing our NPS down, we make it much easier for our organization to take action.

When looking for these specific insights in Kapiche, we use three things:

  1. The Customer Segments/fields included in a data set

  2. Dashboard Filters

  3. The Observed vs Expected Chart

We then get started by:

  1. Applying a Dashboard Filter to focus on the most unhappy customers (usually NPS: Detractors or customers with low scores attached to their feedback)

  2. Clicking into the Theme we'd like to explore

  3. Navigating to the Observed vs Expected Chart

  4. From here, we click through the different Customer Segments/Fields we have access to to see if there are any Segments that are over-indexing on the OE Chart.

For example, here we can see that overwhelmingly most of our Detractors (Dashboard Filter) are experiencing issues with the Quality of their Veggies (Theme we clicked into) in Western Australia (Segment) than any other state:

Immediately we know that freshness isn't as much of an issue in any other states which allows us to go back to the company and let them know that if we focus on improving the Freshness of Vegetables at the stores in Western Australia, we have a good chance at improving our company's overall NPS!

The depth in which you can get your insights will depend on the types of fields you include in the data when uploading it into Kapiche (e.g. in the example above we would have loved to be able to surface which Stores were impacted by this the most but we didn't have access to that data).

With this in mind, we highly recommend having a think about what level of insights your company needs to get to in order to be able to make meaningful decisions, ideally before the data gets uploaded to Kapiche!


Rinse and Repeat!

From here, we'll typically rinse and repeat the workflow outlined above until we feel we have a good representation of the most impactful Themes (both positive and negative) and know which of our Customer Segments are the most influential against our core business metrics.

This usually gives us enough to start pulling these insights together into a report or dashboard that we can share with the specific teams in our organization.


5 Tips for creating a killer insights deck

If you're looking for tips on how to create a super engaging insights report for your audience then this is the guide for you!


That's all we have for you today!

We hope you found this Handbook on how to investigate high-impact themes using your Kapiche Dashboard to be super helpful ❤️

Questions? Feedback? Ideas? 🤔

If you have any questions or feedback about today's topic or even if you just want to say hi, you can get in touch with us any time by using the blue chat button towards the bottom right corner of any Kapiche screen! 👉

We'd LOVE to hear from you!!

Did this answer your question?