What metrics should you be tracking? (Using GoodData, part 1)

19 Comments

  • Stefan Wright
    Comment actions Permalink

    Interesting article Andrea. We currently track weekly and monthly reports via GoodData. Our stats include Avg First Response Time (weekly/monthly/year to date), Customer Satisfaction (weekly/monthly/year to date), New and Solved tickets(weekly/monthly), Channel stats, Number of closed tickets per agent/group(weekly/monthly), Solved tickets last week vs 12 week avg, Solved tickets last month vs 3 month avg, Plan upgrades (relative to our product)(last week/ last month). I didn't realise we could track stats from the Help Center so that will definitely be looked at soon

    0
  • Andrea Saez
    Comment actions Permalink

    Hi Stefan,

    That's fantastic! I love Good Data - so much great insight :) For the Help Center stats, keep in mind that the API is still in development, and it cannot be tracked via Good Data. You can, however, get some basic stats through the Zendesk reports (as you can see on my screenshot above.) I hear there's a ton of great stuff coming through for stats very soon (keep an eye out for parts 2 and 3 of this series!) 

    0
  • Laura Desmond-Black
    Comment actions Permalink

    Very cool article Andrea, can't wait for the next parts!!

    Stefan, as Andrea mentioned GoodData doesn't yet include Help Center data - we're working on that. In the meantime you can sign up for the Help Center Dashboards beta to get the additional dashboards in the agent interface, sign up link is https://zendesk.wufoo.com/forms/help-center-dashboards-beta. Please note you need to be on Plus or Enterprise to get access to the dashboards. 

    0
  • Stefan Wright
    Comment actions Permalink

    Thanks Laura, I have signed up for the beta program. I look forward to seeing the output on there!

    0
  • Jaime Cook
    Comment actions Permalink

    One of our key metrics is "Time To Assign" to an agent.  We need to show that each ticket is assigned to an agent within 10 minutes of it's arrival.  Unfortunately, it appears GoodData only captures this down to the hour level.  We can track the 1st Reply Time down to the minute level, which is great.  Are there plans to provide an "Initially Assigned In Minutes" data point?  Or is there another way to capture this metric?

    0
  • Graeme Carmichael
    Comment actions Permalink

    Jamie, assign time in minutes should be an available metric in GoodData. Are you not seeing that or do you believe that the metric is not accurate?

    0
  • Jaime Cook
    Comment actions Permalink

    Thanks for the quick reply Graeme.  I hope you're right and I'm just not seeing it!

     

    I see a metric for "Assign Time (min) [Avg]" but we're looking to track the median.  I was hoping we could build a formula like that using the First Assign Time number of minutes.  I see "First reply time in minutes" as a Fact, but nothing similar for First Assign Time.

    0
  • Andrea Saez
    Comment actions Permalink

    @Jaime

    You can edit the "Assign Time (min) [Avg]" or create a new one that will track the Median. Instead of it being SELECT AVG it would be SELECT MEDIAN

    Jason wrote a great blog post which you can find here: https://support.zendesk.com/entries/27509727-Medians-for-Advanced-Analytics-

    0
  • Jaime Cook
    Comment actions Permalink

    @Andrea

    Thanks, but even the AVG formula uses "Hours * 60" to calculate the minutes, which is not really accurate.

    SELECT AVG(( Initially assigned at (Date)*24*60)+ ( Initially assigned at Hour*60) - ( Created at (Date)*24*60)- ( Created at Hour*60)) BY  Ticket Id WHERE  Status<>Deleted and IFNULL(  Initially assigned at (Date), 0) > 0

     

    0
  • Holly
    Comment actions Permalink

    Hi all,

    @Andrea, this is awesome!

    @Jaime, I'm afraid you're spot on here. The Assign Time metric in GoodData unfortunately doesn't come "pre-calculated" from Zendesk like the other ticket time facts. I'm so sorry about this, as I can totally see why you'd want this data in minute granularity. It's the one ticket time created by GoodData and essentially stands in as a "guestimation" of your assign time rather than a down-to-the-minute accurate number.

    GoodData essentially gets the data rounded to the hour and is calculated by "Assigned at Hour/Date" - "Created at Hour/Date."

    For example:

    I have a ticket created at 11:50am, and it's assigned to an agent at 12:01pm. In GoodData, the Assign Time is calculated as 12-11 = 1 hour to assign.  Even though we can clearly see here that it only took 11 minutes to assign, there unfortunately isn't any way to extrapolate this from the rounded data that GoodData receives.

    A bit of a bummer, but this data can definitely give some interesting insight when looked at from a higher level.

    0
  • Ian Haigh
    Comment actions Permalink

    Great article. I really like the report on the tickets by shift/group but I can't figure out how to create this with Gooddata. Are you able to shed light on how you grouped the different times into shifts in GoodData? Thanks in advance.

    0
  • Dave Quinn
    Comment actions Permalink

    Im with Ian!

     

    Could you please provide some info on how to create the report on a tickets per shift basis if possible?

     

    Thank you so much!

    0
  • Colin Piper
    Comment actions Permalink

    I am going to provide an example here that hopefully helps you to find the solution, it may or may not be the solution itself.

    I am going to use custom metrics for the specific report rather than a global metric. I am going to assume you know how to create a report so just need help to create the metrics. To create a custom metric, in the WHAT section, under Metrics, click the "(advanced)" option next to Add New Metric. Then click on Custom Metric in the metric builder.

    Note that "# Tickets" must be selected from the Metrics and "Hour Ticket Created" must be selected from the Attributes. Also, the number values must be selected from Attribute Values.

    Note that you cannot do BETWEEN 0 AND 7 - not sure why!

    WHAT

    "Between 00:00 and 07:59"

    SELECT # Tickets WHERE Hour Ticket Created  17

     

    The HOW and FILTER can be set how ever you wish.

     

    0
  • Colin Piper
    Comment actions Permalink

    For some reason, my post was incomplete above

     

    WHAT

    "Between 00:00 and 07:59"

    SELECT # Tickets WHERE Hour Ticket Created  < 8

    "Between 08:00 and 17:59"

    SELECT # Tickets WHERE Hour Ticket Created  BETWEEN 8 AND 17

    "Between 18:00 and 23:59"

    SELECT # Tickets WHERE Hour Ticket Created  > 17

    0
  • Andrea Saez
    Comment actions Permalink

    Hi all,

    Terribly sorry for the late reply, I just saw this (somehow missed it in my inbox)

    To create the tickets by shift, you need to create a new metric for every shift.

    Example:

    Overnight shift 10pm-8am

    SELECT # Tickets WHERE Hour Ticket Created IN (22, 23, 0, 1, 2, 3, 4, 5, 6, 7, 8)

    You would then subsequently create this for every single shift. Then set the view to a donut chart, and voila!

    0
  • Gabriel Scott
    Comment actions Permalink

    Hi @Andrea, in this scenario, how do you report over multiple days? Insights only has so many methods for slicing this data and adding date into this as a filter still gives only tickets handled on one day, which is halves of two different shifts.

    I've created a similar metric (measuring ticket replies, not solves) and it will show the right hours, but adding a date metric still presents the same issue I had with many of my first iterations of reporting over midnight.
     

    0
  • Gabriel Scott
    Comment actions Permalink

    Something like this: 

    0
  • Andrea Saez
    Comment actions Permalink

    Hey Gabriel, looks like you got it. Yes you'd have to add an "AND" to the logic so that it measures both time and date of event. 

    0
  • Jennifer Rowe
    Comment actions Permalink

    Thanks for popping in, Andrea. :)

    0

Please sign in to leave a comment.

Powered by Zendesk