The Workbench is a dynamic view of your customer feedback. The workbench allows you to customize the display of widgets, navigate easily between widgets and filter your data and insights according to different customer segments.

The Workbench gives you access to our new AI summaries:

In addition, our new tree-view Themes widget displays your theme framework with Theme Groups and Themes. This view allows you to easily explore key statistics at the high level, as well as dive into sub-themes to understand your customer feedback at a more granular level.

Compare Mode & Insights Cues. Compare Mode will allow side-by-side comparisons on the workbench. Our soon-to-be released Insights Cues feature will automatically identify peaks, troughs, and outliers in your trend/timeline data.

The future state of the Kapiche workbench will enable configurable Automatic Reports with key business stakeholders at top of mind.

You can share your Workbench in a range of ways:

Filter your Workbench to have the whole Workbench reflective of a subset of your data. Filtering to particular customer segments or demographics allows you to find insights for that particular segment, as well as make comparisons. Saving your filtered views is recommended to save time with reporting and give your stakeholders a customized view of the data and insights.

You can save any configuration of widgets, show and hide widgets, and show and hide themes, in addition to any filters you have applied.

The Workbench features a clickable sidebar that allows you to quickly navigate between each of the following sections:

The Dataset Overview leverages AI to identify the Top Paint Points and What’s Working Well across your overall dataset. The Dataset Overview is automatically generated each time the underlying data changes, for example, when new data is added or Workbench filters are applied.

The Themes widget displays the complete Theme framework for your dataset. By default the high level Theme Groups are visible, with themes nested underneath their respective Theme Group label. Themes can either be generated using Kapiche’s AI Auto Themes or be custom built via the Theme Builder. See the Theme Builder to view, merge, re-categorize and re-name themes or theme groups.

You can view your themes by a range of values as well as the impact of the theme on the selected number, score or metric:

To understand how impact is calculated please see: How does Kapiche calculate Impact?

The Themes widget also allows you to compare values for a customer segment to the overall values. To do this add a customer segment filter to the Workbench. The above values will then be displayed for your chosen customer segment. The ‘expected’ value is based on the overall dataset. By looking at the differences between the value for the particular customer segment and the value for the overall population, you can identify which Themes are more important to that particular customer segment. This is because it identifies which themes, topics or issues are spoken about more than what we would expect.

The Net Promoter Score widget is reserved specifically for the NPS metric. It shows a breakdown of promoters, detractors and passives and the NPS score. In addition, a timeline chart is shown to visualize and track this important metric on your Workbench.

The Score widget is a dedicated widget for any numerical fields you set as a score field during project creation. This widget can be configured to display the score (e.g. CSAT) as the:

Just like the NPS widget, the Score widget includes a timeline view of your metric. The Score widget shows one score at a time and can be configured to show the most important one for your analysis.

See: How to configure score fields & How to use the Score widget

Sentiment is calculated by Kapiche using Natural Language Processing (NLP). We use our own proprietary sentiment analysis algorithm that assesses sentiment at both the overall verbatim and sentence level.

The sentiment expressed in your verbatim data is shown on the Workbench at the verbatim level. For example, if a single verbatim contains both positive and negative sentiment it will be assigned mixed sentiment. The Sentiment widget shows the percentage of verbatim in your dataset that contain positive, negative, mixed and neutral sentiment. The Sentiment widget also tracks changes in sentiment with a timeline chart.

Sentiment is a great tool to measure and track what your customers think and feel, without the need for a rating or score question. This means you can load any text data into Kapiche and get a metric that you can use to measure, track, determine impact and prioritize.

The Segmentation widget gives you visibility of the customer segments in your data, showing the frequency of verbatim data by customer segment as well as the NPS for each customer segment.

The Segmentation widget is most powerful when you have drilled into a Theme. It allows you to identify which customer segments are over-represented for that particular Theme, providing you with a customer segment to target or deep dive into to find further insight. To do this, drill into a Theme from the Themes widget. Refer to the Segmentation widget to find segments that have a much higher observed frequency than expected frequency.

The observed frequency is based on how often the particular theme is spoken about by each customer segment; the expected frequency is simply the representation of each segment in the overall population (e.g. x segment represents x% of all respondents).

In the below example, we see that the 'Low' spend category is overrepresented when it comes to the theme of 'Pricing'. This customer group's feedback includes the theme of pricing more than we would expect. 'Low' spenders also happen to be our largest demographic.

The Timeline widget is the place to track your Themes, identify trends, large changes over time and large differences between themes at any given point in time. The timeline chart not only shows specified values over time but Kapiche’s impact calculation, which tells us the impact of any given theme on a particular metric (displayed on the timeline chart).

This widget can display a range of numerical fields by theme that include:

For detail on the Timeline truncation (e.g. 'Top 5') and the options to do this by (e.g. 'Highest absolute change' see: How to use the Timeline Chart on you Workbench

The Top Correlations widget is designed to identify strong relationships across any of the fields in your data. This widget casts a wide net across all of the fields in your data to allow any unexpected relationships to emerge. This means that it does require some configuration to hide fields that aren’t of interest or don’t make sense to explore.

The pivot table gives you an aggregated view of one or more categorical fields in your data. Fields can be aggregated by the below:

The pivot table is a great way to compare your themes across the full range of categories or segments for a particular field. It also serves as a convenient way to obtain the above values for reporting purposes. You can export the Pivot Table data as a CSV, PNG or PPT as well as copy to clipboard as a PNG image file.

The Quandrant Chart in Kapiche is your classic prioritizer. It is a powerful visualization to showcase Themes that are high impacting and highly negative (red quadrant), high impacting and highly positive (green quadrant), and topics that are negative (orange) or positive (blue) but spoken about less frequently.

The Quadrant Chart also displays the overall value as a line to clearly visualize which themes are having a positive impact on the score and which themes are having a negative impact. Numerical fields do not display as quadrants since they are not specified in Kapiche to have a positive and negative end of a scale.

In order to show the top Themes, the Quadrant Chart can be configured to display the top x by:

In addition, the Quadrant Chart can be used to visualize all Themes within a Theme Group to allow for prioritization within a specific business area, for example 'Customer Service and Staff' (as shown below).

Drilldowns via the Workbench are a powerful tool to deep dive into your Themes and customer segments to:

You can click into a number of elements on the Workbench to drill down further.

These include:

You can also go another layer deeper by creating intersections between the data in your drilldown and another Theme, customer segment, key phrase or concept. In the below example you can see we clicked on the ‘High’ spend category to reveal what this customer segment in particular thinks about 'Product Quality and Freshness'.

Here you’ll find an AI generated summary of what is being spoken about for a particular Theme or customer segment that you have drilled into.

You will also find Sample Verbatims that come in handy for reporting.

Here you can see what other Themes are spoken about alongside the Theme or Theme Group you have drilled into.

The context Network is a visualization of the Theme you have drilled into using key words that represent concepts. The strongest relationships are connected with a line, with the thickness of the line showing the level of strength. The size of the bubble indicates the frequency the word and variations of that word appear in your data.

Use the context network to get an understanding of the narrative surrounding a theme.

You can also use it to go one layer deeper by clicking on any of the words in the context network. This will filter the widgets on your drilldown to a subset of the Theme where the particular topic or concept is mentioned.

The Verbatims widget is where you will find the verbatims that make up the Theme. By default you will be shown the most relevant first but can also sort by recency, and NPS. If your dataset contains NPS, you will be shown the NPS alongside the verbatim. Click ‘SHOW INFO’ to see categorical values associated with the verbatim.

As mentioned earlier in this guide the Segmentation widget in drill down mode allows you to identify which customer segments are over-represented for that particular Theme.

The observed frequency is based on how often the particular theme is spoken about by each customer segment; the expected frequency is simply the representation of each segment in the overall population (e.g. x segment represents x% of all respondents).

The Key Phrases widget shows the phrases that are most strongly correlated with the Theme you have drilled into. Key Phrases give you additional context to understand what the Theme is about and what language your customers are using.

Use Key Phrases to drill even deeper by looking at the intersection between the Theme and a particular key phrase.