Benchmarking allows you to specify benchmarks that can then be reported in Feedback and Cohort reports. They are based on all previous responses for an assessment and are refreshed nightly. For each benchmark, an average response score, section scores, subsection scores, and segmentation scores are automatically calculated and available for reporting in all feedback and cohort reports associated with this assessment.
As well as "all responses" you can define benchmarks based on Classifiers. These allow benchmarking at any level, down to and including fellow team members. If you, for example, had classifiers of job roles and geographies, you could show benchmark scores for responses from the same role and the same geography as the respondent. Any chart can show up any number of benchmarks alongside the response in a feedback report or responses in a cohort report.
Multiple assessments can be included one benchmark provided they are structured exactly the same, i.e. the sections and subsections are aligned. This means that multilingual assessments or copies created for different brands can optionally be rolled into the same benchmarks.
Steps to specifying Benchmarks
- Indicate that Benchmarking is required by simply ticking the checkbox on the assessment on the Assessment Edit form. This will initiate the automated nightly calculation of benchmark data. On this form, you can also
- associate assessments with a parent assessment. This is intended for use with multilingual or copies of assessment made for cobranding purposes etc. The assessments must be aligned on section, subsection, and segmentations. The responses for all related (assessments that have a parent specified or are the parent) will be totaled into the benchmark.
- specify a label for "all data". This might be "All companies" or "All employees", etc. depending on the context of your assessment. This label is used if you report "all responses on your feedback or cohort report.
- If the further breakdown of benchmarks is desired, you should define Classifiers and enable benchmarking on the Classifiers to be used.
- Benchmark data is created for each value of the classifier found in the data
- The classifiers have a label associated with them that will be used when reporting the data.
- The classifiers have a minimum number of responses that are required for each data value before benchmark data is generated. This is to avoid comparing a response to a "benchmark" that is based on only one or two previous responses.
- The Classifiers must be associated with a question in the assessment using the Question Edit. The question may be hidden if you are predefining the classifier when creating the response request.
Benchmarks are reported by including the "BM" parameter in merge strings in the Feedback or Cohort reports. Benchmarks can be shown on their own using charts like bar charts, or alongside the response (s) using charts like multi-bar charts.