Skip to Content
Technical Articles

Summarizing Analytical Capabilities of Qualtrics – SAP’s Experience Management (XM) Platform

Hello Community,

There is a lot of buzz around how we at SAP can combine the operational data (from SAP Digital Core) and experience data (from Qualtrics XM Platform) to deliver frictionless and amazing experience to our customers. This intrigued me, and I delved deep into Qualtrics XM Platform’s capabilities. Today, I am summarizing the analytical capabilities of Qualtrics which are driven by its two engines Stats IQ and Text iQ.

Briefly,

Stats IQ – Used for Statistical Analytics. It has five algorithms – Describe, Relate, Regression, Pivot and Cluster.

Text iQ – Used for analysis of textual/descriptive response questions.

In Detail,

Stats IQ

A. Describe

1. Describing your data

  • We can review existing variables.
  • Renaming of variables can be done
  • Changing data types of existing variables
  • Hiding unnecessary response values

2. Describe – based on structure of data – suitable card is displayed

3. Various question types are supported- single select, multi select (which don’t add to 100 %) ,               matrix, etc.

4. For matrix questions – we can reorder the options as well for convenient interpretation.

5. It’s also possible to display time-period based data.

6. Filter can be set on all cards or individual cards to restrict the analytics to certain responses.

*********************************************************************************

B. Relate

1.  Exploring Relationships Relate – Links two variables and gives output in plain English

  •  Key – It’s the output variable
  •  Other – It’s the input variable

2.  Can see statistical test results

3.  Based on data analysis cards will differ and are automatically determined

4.  Runs a correlation test, chi-squared test, etc.

5.  User can switch rows and columns as well

6.  Time based values can be also be shown with an adjustable bin size.

*********************************************************************************

C. Regression

1. Key insights using Regression

  • Used when many co-related variables for the key and dependent variables are present in        the system.
  • Helps in tearing apart complex multi-dependent relationships and understanding it.
  • Result is in plain English with metrics and prediction features as well. Users can change          values and determine the output.

2. It has multi-input, one output (key)

3. Add/Remove variables directly in results and output updates itself.

4. The algorithm is analogous to normal regression technique.

5. An important metric is relative weights – which defines how heavily the output is influenced                by the particular inputs. The ones with highest relative weights impacts the most.

*********************************************************************************

D. Pivot

1. Pivot Tables

  • Use it for heavily customized cross tab views

2. Similar to Relate technique but used when we intend to add more variables in the output

3. User can even change it’s metric at runtime (supported are – average, percentage, etc.)

4. Addition of variables at runtime is also allowed.

*********************************************************************************

1. Creating variables using logic, buckets and formula – works based on structure of data

  • Logic – Create new variables with complex statements – E.g. – tag respondents by age.
  • None of the above – Manually, Missing or via other variables
  • Bucketing – groups values together – numerical or categorical values – e.g.– countries          grouped by regions/continents
  • Replace existing variables with new or another existing variable
  • Variables are not deleted, just hidden, also hidden variables can be used in analysis
  • Formula – mathematical formula – using variables – only numeric responses allowed

*********************************************************************************

Text iQ

1. Analyzing text responses

  • Text analysis – user can create topics to search and categorize the responses in                  descriptive questions.
  • Responses get automatically tagged with topics.

2. Lambdas and spelling errors are ignored, and responses are suitably tagged. E.g. baged,                   bag, bags, bagged, bag, etc. are all valid responses for topic – bag.

3. Relational operators are supported in building complex topics with parenthesis, keywords                    (and, or, not, etc.), etc.

4. Pie chart shows the breakup with status of response classified and not classified

5. Can update, remove and add new topics any time during processing.

In nutshell, Qualtrics offers rich and flexible analytical capabilities with its own integrated user interface.

Happy to have your feedback and suggestions.

2 Comments
You must be Logged on to comment or reply to a post.