Average of a subset of questions (Assessment Metrics)
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎05-03-2017 03:50 PM
Hi everybody,
I was looking for a way to get some "calculated" kind of metric, based on other responses belonging to the same assessment.
For example, if your questions are like this:
- How satisfied are you with your technician's skills? 1-5
- How satisfied are you with your technician's politeness? 1-5
- Was your technician's response time good? 1-5
- Comments: (this shouldn't be used as part of the calculated metric)
I'd like to have a new (fifth) metric that won't be asked to the user, for each survey taken, that works as something like "Overall customer satisfaction" and calculates the average score of the first three metrics.
Is there a way to achieve this so that the results will appear just like any other question in the asmt_metric_result table?
The final purpose of this is using these results to create a monthly average of the user satisfaction in an indicator with targets and thresholds and all that good stuff.
Thanks.
- Labels:
-
Analytics and Reports
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎05-03-2017 04:39 PM
That's a good idea of handling all survey results. Yes it is Possible with Performance Analytics.
Firstly, you should have view between assessment metric type and assessment metric result and if it is for incidents include incident, asmt assessment instance in view.
Then you will create Automated Indicator's and one Indicator Source for each of 5 Metric types you mentioned to calculate SUM of Result/Actual value which should be in decimal/integer.
As you have Sum now. This time you will create same above set of Automated indicators for Count
Finally, Create a Formula Indicator that has formula with all above automated indicators SUM/COUNT{Sum1+sum2+......./Count1+Count2} This gives you Average of Survey Responses that you created above.
Create a widget using above formula indicator and then show it on Dashboard
Hope this helps, Let me know if you have any questions.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎05-03-2017 05:21 PM
Thank you for the quick reply.
Yeah, I already made one of those, but the reason to do it for each survey and keep it as a separate metric is that I need another indicator as well where I need to count the number of surveys (for each individual survey) and break it down into buckets where the avg score was below 3.5 or above 3.5. And then a formula indicator displaying the percentage of scores above 3.5, so it'd be (100 * (records above 3.5 / the total count ) ).
Is there some way to keep the individual avg score for each survey?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
‎05-03-2017 10:40 PM
when you say each survey I assume like for each question.
so if you have 4 different metric questions
and for each question you create one automated Indicator to aggregate Sum on 'Actual Value(Or whichever field that stores result)' and One to get Count of total surveys for that metric question.
Like that you will have set of 2 automated Indicators for each question.
if you want individual average Just create a formula indicator that has indS1/indC1 which is SUM/Count = Avg for each metric question.
if you want Collective average Just create a formula indicator that has all indS1+indS2+.../indC1+indC2+.. which is 'SUM of values of all metric questions/Count of total surveys for all metrics' = Total Avg
and remember to give 1 as 'Precision' in your formula indicator to allow the ratio value to one decimal place. like 2.5,1.7 etc
To use formula indicator to calculate surveys for only above 3.5 .you can do this when selecting indicator in your formula, also select your PA bucket Breakdown with also an element >3.5 that you should already have in your Pa Bucket. if not create a PA bucket breakdown by referring Bucket groups
That gets you: records above 3.5 / the total count