When you're analyzing data from multiple surveys in the executive dashboard, the way questions and response properties are structured doesn't always support the visualization you have in mind. For example: you might want to combine two similar questions from two different surveys. Perhaps you want to segment responses based on a bespoke time-frame that your business operates on. Or you want to create a custom scoring system for responses.
That's exactly what Custom Groups is designed for.
Custom Groups is a new feature that lets you create your own dimensions — all without touching the underlying response data. You decide how your data should be grouped, labeled, and calculated, and those groups instantly become available across all your widgets.
In this article, we’ll walk you through the steps of creating custom groups.
1. Inside the executive dashboard, click on the three-dot icon, and select Custom Groups.2. When the pop-up appears, click on Create Custom Group.3. In the general configuration section, provide a name for the group, then choose the type of group by clicking on the first drop-down menu under Utility (default option is data group).
There are three kinds of custom groups you can build:
Data Groups - you can combine multiple response options into meaningful buckets. For example, a data group which merges Likert scale ratings into 3 columns: Bad (1, 2, 3 and 4 stars), Average (5, 6, 7 and 8 stars) and Good (9 and 10 stars). Or a data group that merges options from similar MCQs into columns. You can even assign a numeric score to each group for weighted calculations. Inside widgets like bar graphs, this data group can be used as your X-axis dimension.
Merged Questions - you can combine similar Likert scale questions from different surveys into a single, unified dimension. This gives you a bigger sample size and lets you make cleaner comparisons. In bar graphs and related widgets, this can only be used as a Y-axis dimension.
Time Period - you can define timeframes in any manner that matches your business’s operational schedule and then use them as dimensions for historical breakdowns. For bar graphs and similar widgets, this can be used as an X-axis dimension.
Note: You can make your custom group available for the current dashboard or all dashboards, by clicking on the drop-down.
Data Groups
(For the purpose of this article, imagine you're an electronics retailer running identical surveys for Apple and Samsung, and you want to consolidate the responses into a single view. You’re going to create a data group that aggregates two identical MCQs from both surveys into one)
1. If you’ve decided to create a data group, the next thing to do is to select your source survey. Click on the drop-down under Source and choose from the options. You can add multiple sources.2. After adding your sources, click on Define Groups.3. Inside the configuration for the new group, you must first set the data properties. For data groups, you can choose from questions, variables, contact properties, response properties and share channels. 4. For the first group, select the property and its specific instance, from any one of the source surveys. In this case, we’re selecting an MCQ from the Samsung survey. 5. Next, you must define the qualifying values for this property. In this example, we want to create a bucket to aggregate specific options. Select the ratings and click Apply.6. If you’d like to add additional properties, click on the plus icon next to the property and follow the same steps as described above. In our example, we also need to combine the same choice from the Apple survey, so we’ll be repeating the same steps.Note: When chaining multiple properties from different sources, the default logic is And - meaning that responses must satisfy both properties. If you’re creating properties from different sources, use Or logic instead. You can change this by clicking on the drop-down menu between the properties.For further specificity, you can also create multiple chains of properties inside a group, by clicking on the Filter Group icon at the top. The default grouping logic is And, which you can change to Or.To name your group, click on the Rename icon next to the Filter group icon.7. To add multiple groups, click on Add Group. In this case, we have to create groups for the rest of the options. Once you’re done, click on Save.How can data groups be used?
Imagine that the retailer from our example wants to understand feature preferences across Apple and Samsung buyers. Both surveys have an identical MCQ for feature selection (‘Your favourite feature?’) which can be clubbed into a custom group. (See the group created above.)
This group can be used as the X-axis in a bar chart widget. But in order to visualize it, you also need to combine both sets of survey responses into a Y-axis dimension. You will have to create a merged question group, which we’ll explore below.
Merged Questions
This custom group lets you merge Likert scale questions (from one or multiple surveys) into one dimension, giving you a bigger sample size to measure across your X-axis dimensions. Let’s look at the setup. (In our example, we want to combine the ratings of Apple and Samsung buyers into one dimension.)
1. Just like above, after naming your group and choosing the merge questions type, proceed to select your survey sources by clicking on the drop-down menu under Source.2. After adding your sources, click on Define Groups.3. Now, select the questions you want to merge, by clicking on Add Questions and choosing from the drop-down menu. You can only select one question at a time, so repeat the step again.Note: You can only merge questions with the same Likert scales. If your first question has a 10 point scale, you will only be able to choose questions with 10 point scales. Questions with different scales will not even appear.4. After adding the questions, click Save.Coming back to our previous example, we have a data group (X-axis) to combine the feature preferences of Samsung and Apple buyers, as well as a merged question group (Y-axis) which combines both sets of responses. When a bar chart is created, this is what the breakdown would look like.Here, the feature preferences are plotted against the combined total of buyers. This lets the retailer understand which features are preferred by buyers, irrespective of brand.
Reporting values inside Data Groups
When creating data groups, you can also add a numeric score in place of a text value, if you wish to perform calculations within the widgets for those values. Let’s explore this with the same example.
Suppose you have a matrix scale question in both surveys that asks users to grade features (screen, battery, camera, performance) as ‘Bad’, ‘Okay’ or ‘Good’. You want to understand which features are performing the best. But to understand how buyers of both brands perceive the quality of these features, you must assign reporting values to these ratings to get an average score.
1. Start by creating a data group combining the matrix question columns from both questions into 3 buckets. To assign reporting values, click on the checkbox at the bottom.2. A reporting value label appears next to each group. Click on it, enter your value and click Save. In our example, we want to assign the following values to each rating: 1 for Bad, 2 for Okay and 3 for Good.3. After assigning those values, save the data group. Next, create another data group that combines the rows of the matrix questions (screen, battery, camera, performance) into different buckets. This becomes our X-axis dimension.4. Now, create a bar-chart widget with the Y-axis dimension being the group with reporting values, and the X-axis as the group aggregating features. But in the Y-axis selection, click on Count, and change it to Average.This is what the widget would look like. The best-performing features have the average score that’s closest to 3, letting us understand what’s performing well among customers.
Time Period
In this question type, you can define your own timeframes on any scale. You can then use your time period groups as X-axis dimensions in graph type widgets. Let’s see the process.
1. After naming your group and choosing the time period type, proceed to select the property that you want to use by clicking on the drop-down menu.There are 3 categories you can choose:
Submission Time: when the response was received.
Variables: any date-based variables of a specific survey.
Contact properties: you can select the contact creation date.
2. Once you have selected the property category, click on Define Groups.3. Just like in Data Groups, you can create multiple groups (columns), with filter groups within. All the filter groups are units of time. (In this example, we want to split responses based on financial quarters, so we’re going to create groups of months.)4. After adding your groups, click on Save.Let’s revisit our example. Suppose the retailer operates on a quarterly cadence (Jan - Mar, Apr - Jun, Jul - Sep, Oct - Dec) instead of a monthly cadence. They want to understand which quarter generates the most consumer activity.
So they first create a time period group with buckets for each quarter. (See the group created above.) They then create a merged question group to combine the responses for both surveys. Then they create a bar chart widget where the merged group is the Y-axis and the data group is the X-axis.In this case, the quarterly segments are plotted against the overall sample size, so it is clear when customers were the most responsive and engaged.
Now you’re all set to use custom groups in your widgets. Now, your data can be shaped to match your needs right inside your dashboard, not in spreadsheets or BI tools. Anyone on your team can build the views they need — without touching the underlying data or waiting for someone else to do it for them.
Feel free to reach out to our community in case you have questions!
