Incident Quality Review in SN

[Glenn H]

One of our departments here is doing a quality review on the tickets handled by an off-shore team.  We are looking for best practices on how to manage the whole process in ServiceNow.  I was leaning toward Visual task boards but the user is unable to create from the task time worked table.  Any advice or thoughts would be greatly appreciated.

Current Process

Our user is currently looking to do QA on all tickets touched by the vendor. The user currently exports the task time worked table into a report and then the report into Excel.  The user then selects a random number of those tickets and reviews them. Afterwards, the user must return to ServiceNow to reviews the work notes. After the review, he assigns a rating on the information and loads that information into a different spreadsheet. That spreadsheet has month over month data that shows how well the vendor is doing.

We use the task time worked table because that team usually does not end up with the ticket closed by their group.  Most of the work is handled by our service providers of another group but stays within the same service offering.

[Uncle Rob · Accepted solution]

I'll think you'll find this is a pretty manual process in most places, otherwise SN would have put a simple version of it in the incident management tool.  Here's how I might build it.

GOALS:
- process must be as completely within SN as possible.  Data processing, analysis, actions outside tool are to be avoided
- as little customization of product as possible.  Salvage what we can from OOB

KEY QUESTIONS:
- How can we identify 'potential' tickets without exporting?
- How can we rank ticket quality?
- How can we show month over month performance?

Identifying Tickets to QA

It gives me the heebeejeebees that hte ticket doesn't actually transition groups.  If it did, I'd select a random sampling from the Metric Instances for Group Assignments (available to you in the Incident Metric table).  Then it wouldn't matter if the task ended in the group, only that it was there at some point.

The next problem is simulating the "random" selection.  Your choices are to either eyeball select a set of X or get the computer to do it.  In the latter case you'd want a "rated" yes/no field, then a scheduled job that runs every rating interval. Each interval select the population of incidents you could pick from, then pick X randomly.  That'll need an If statement with a math coin toss like Math.random()... but the thing is you may get unlucky and lose a coin toss each time.  Someone else could fill in the blank on javascript that gets you "X random records" from a gliderecord set.  If a ticket is selected for sample, mark "rated" to yes.

Ranking Quality

There is a "rating" field in ServiceNow but it doesn't seem to be exploitable yet.  You'd consider building a u_rating number field.  QA person looks at set of all Incidents where Rated = yes and u_rating "is not empty", then adds ratings.

Tracking Month over Month performance

The best tool for this job is performance analytics, which is purpose built for summarizing data and storing "scores' period over period.