How did you improve your metrics? We want to know!


  • Bram Verswalm

    We have made our own Lync tool which helped a lot! Read more here:

  • Andrew Soderberg

    We have been running a 98% or higher CSAT score for the last 3 years. We've had a 100% CSAT score for the past 7 weeks. We currently solve about 600 new tickets a month (double the tickets from 2 years ago). We solve 41% of all tickets in under 4 business hours. 53% are solved the same business day. I have a front line team of 6 agents, the second tier and other departments make up the difference.  The front line team solves over 80% of all our tickets. We make heavy use of Zendesk/GoodData reporting.


    We do extensive training of both our support staff and of our key support contacts at our customers (Colleges and Universities). We support the admins, trainers, and web developers at our customers (they in turn support the over 50,000 end users of our CMS).


    So, when I say that we are currently solving about 600 new tickets a month, there are no Level 1 ‘I lost my password’ type questions among them. We are only getting industry equivalent level 2 and 3 support requests from our customers.


    We historically got an average of ~35% satisfaction survey response rate per month. We are now seeing 50% response rates the past several weeks (we took the advice of a blog post from Zendesk about when CSAT surveys are sent to customers). We now have the survey emails sent in the middle of the night. Since that change our weekly response rates have increased from 40% to now 50% per week.   

  • Jennifer Rowe

    Andrew, that's pretty impressive! Thanks for sharing your story. Out of curiosity, which blog post did you read?

    Keep up the good work! :)

  • Jennifer Rowe

    Ah, yes, an oldie but a goodie. Thanks for the link. We'll add it to our documentation to make sure others find it too. 

  • Daria Gru

    We are a small eCommerce business using a variety of sales channels (multiple Amazon markets are the biggest ones). Majority of our customers are used to Amazon type of customer support (decide in favor of customer regardless of objective situation). One of our challenges is the fact that we get negative feedback when customers get frustrated with Amazon's platform, not with our service.

    In the past year year we never dropped CSAT below 90% and most of the months we manage to keep it around 94-95%. We receive around 1200 new tickets per month on regular months (not pre-Christmas and Christmas season, when it usually doubles). We aim to keep first reply around 12 hours, but we have only 3 people doing support among other tasks so it tends to be around 20 hours.

    Reading Andrew's post, I feel that we came to the point when we need to add an agent to our support team.

    Interesting article regarding timing of feedback surveys!

  • Aesquilin

    I believe that the most important aspect of increasing customer care is the  initial greeting of an incoming call. I have 25+ years in customer care, and I have come to the conclusion that the key to success is getting all the necessary info needed to efficiently give a client a high quality of care. This eliminates wasted time corresponding with clients via email thus prolonging the actual service time. Also, it gives the perception that your really not grasping the nature of the call , lowering the confidence and trust level with the client. I work for Telego Phone services and I perform level one tech support as well as administrative support. I usually answer the incoming calls and I try to implement this mind set with every call I receive . In doing so I have seen a rise in the level of service . So, as small as it may seem, the initial contact is the most important one.

  • John Rauser

    Hi there, we improved our First Response Time by getting Phillips Hue lights and placing one above each desk of our support team members. We wrote a Python script that uses the Zendesk API to check if there are any new tickets in any of several queues. When a new ticket comes into any of the queues, the light turns on and stays on until the ticket is taken by one of our team. The system is very effective and guarantees that a ticket will get assigned to someone within 5 minutes. 

  • Tim Clark

    We are improving our customer sat by having the team-leaders run daily scrum calls to discuss open tickets and an open floor for anyone to discuss any tickets that they are stuck on or want other eyes on it.

    This has increased focus on the tickets and solves them faster. The Collective Mind of the team is greater than the Individual.

  • Clint Wilson

    I think having a Follow-the-Sun approach has worked out well for our type of software subscription business since using Zendesk with our 5 country teams.

    • First reply time : time between ticket creation and the first public comment from an agent
    3.3 hrs
    2.6 hrs
    27.8 hrs

    Your first reply time is 24.5 hours shorter than your industry average. This is better than 93.4% of other Zendesk customers.

    • One-touch resolution : percentage of tickets that were solved with one agent reply or with no agent reply

    Your one-touch resolution percentage is 8.5% higher than your industry average. This is better than 66.1% of other Zendesk customers.

    • Customer satisfaction : overall customer satisfaction rating for your Zendesk, including all users and groups

    Your customer satisfaction rating is consistent with your industry average. This is better than 73.0% of other Zendesk customers.



  • Camila Correia


    Aqui na nossa empresa temos utilizado os SLAs para first reply. Acredito que isso tenha ajudado bastante na melhora dos nossos prazos de resposta. Cada agente consegue enxergar na sua fila quando o próximo ticket vai estourar.

    Além disso temos diversas macros para perguntas frequentes, além de gatilhos que disparam respostas automáticas para perguntas repetitivas.


    Edited to add English translation:

    Hello! Here in our company we have used the SLAs for first reply. I believe that this has helped a lot in improving our response times . Each agent can see in your queue when the next ticket will burst . Also we have several macros for frequently asked questions , and triggers that trigger automatic responses to repetitive questions.

  • Asger Andersen

    We work specifically towards identifying the common issues with our clients and deliberately create standard replies that fit the common troubleshooting steps so that we do not type out similar answers time after time. 

    Also, we're using the help center to redirect the client to self-help articles where we thoroughly describe the issues and the common ways of resolving them.

  • Kapil Dev

    Hey, we don't do anything special but analyze ticket trend, nature of tickets, frequent inquires etc. & train the staff accordingly. Ensure to acknowledge new ticket with macro comment & provide answers turn by turn.

    Prioritize tickets high to low instead of first come & first serve basis. Prompt escalation handling process. And above all, read tickets carefully, in some cases multiple times to be sure you understand the request & then respond. Hence save time & improve resolution time. Be specific & probe in case of unclear request.

  • Jason Fordham

    Since the first response time metric is meaningless and unmanageable, we don't do anything to manage the reported metric at all.

  • hypercatjohn

    I can't say I'd be happy waiting on a non-existent response, so I'd happily say it's up there as one of the most important elements. Perception is everything, if a customer get's a good service (that includes a quick and meaningful response) then you'll have their confidence for a long, long time.

  • Laurie Anderson

    We're anal/retentive about responding as fast as possible to customer inquiries, requests, and, yes, even complaints!

  • Matt McLean

    I agree with Jason Fordham that average first response time is unmanageable, at least out of the box.

    Using Insights, we were seeing a Median first reply time of less than 30 minutes, but the Average first reply time found in the Monthly Snapshot emails was hovering above 10 hours.

    I have had great success bringing our Average first reply time down in the past month by preventing tickets from "falling through the cracks":

    1. I created a Trigger with [all] of the following conditions:

    [Ticket Status▾] [Less than▾] [Closed▾]

    [Ticket: Agent replies▾] [Is▾] [1 ]

    [Ticket: Comment is…▾] [Public▾]

    that performs the action:

    [Ticket: Add tags▾] [first_reply_has_happened ]


    2. I created a View called "Tickets With 0 Replies" with [all] of the following conditions:

    [Ticket: Status▾] [Less than▾] [Solved▾]

    [Ticket: Tags▾] [Contains none of the following▾] [first_reply_has_happened ]


    When I first did this, that view was of course filled with the entire backlog of tickets that existed before I created the new tag.

    There was a tedious process of "Playing" that view, and manually tagging tickets that had had public replies. During this process, however, I also found that there were many unsolved tickets, some of them days or even weeks old, that truly had 0 public replies.

    In most cases, these were tickets opened by the agents themselves - either on an end-user's behalf, or a ticket from one agent to another. Some of the internal-only tickets even had Private Replies going back and forth for days! Tickets that should have helped our 1st reply time metric, since we could have literally replied to them publicly within seconds, had been hurting us.

    For about a week, our 1st reply time average spiked UPWARD, as I finally replied to all of those tickets.

    After clearing the backlog, though, the new queue has been very helpful.

    Our agents are using my new "Tickets With 0 Replies" view in conjunction with the default "{{zd.unassigned_tickets}}" view, making sure that 1.Every ticket gets a public reply and 2.Every ticket gets assigned to an individual, not just a group.

    We've also all gotten into the habit of immediately replying to tickets that we enter ourselves - as I mentioned above, they should be "freebies" that help bring the average downward.

    The only "negative" we have seen recently has been that our One-touch Resolution went down from 26% to 21%.
    But this corresponded with a drop in First Reply Time from over 15 hrs down to less than 3 hrs, and this month I am already seeing it get down to about 1 hr… a heck of a lot closer to our Median of .4 hrs.

    It would be really nice if Zendesk had this functionality built in, but now that it is working, I'm happy.

    Warnings to anyone who wants to do this same thing:
    1.The instance where I am doing this is an internal support Zendesk, where our "end users" are all employees of the same company I work for. Our CSAT went up from 99.1% to 99.3% during this process. Your mileage may vary.

    2.If you try to automatically tag your backlog using an automation, you may incorrectly tag tickets that had 0 public replies from agents. That's what happened to me, and when I thought the backlog was cleared out, I was still seeing random spikes of 1st replies coming in. This is why in my instructions above, I recommend going through the (tedious) "Play" process for the view instead, so you don't let any tickets "fall through the cracks".

  • Jørgen Sivesind


    We are a small shop who recieve 50-80 support requests per month.  After we started using the ZenDesk app on our phones, our First reply time have dropped to a median of just 5 minutes and an average of 17 minutes (from median 28 and average 58 before we started with the app).  We are two administrators with the app, and when a new ticket comes in, at least one of us sees it almost immediately, so we can assign it to an agent.  Especially if new tickets come in on the weekend, this is very helpful with the app.  Before the app, we could get a ticket on a Saturday, that nobody saw untill Monday.  Such tickets obviously took a heavy toll on the average stat.

    The "First reply time" metric is a little skewed in our case, because it does not necessarily include a real answer to the requester, just information that the ticket have been assigned to an agent, and which department the agent works in.  (We then have Automations that monitor the tickets and make sure they get more replies.)  -  However, the fact that we see the tickets so quickly means we get them to the agent faster, and it is VERY important for tickets about critical issues.  In those situations, the two of us with the app, can talk directly with the agent and make sure he looks at it immediately, or discover that he is unavailable, so we need to find somebody else to look at it.

    It should probably be added that the App had a bug that for a long time prevented the notification of new tickets to work, but it was fixed in May, and that was when our stats started to get really good.

     -  So in conclusion: To improve the "First reply time", Use the App.  😎

  • Nicole - Community Manager

    Glad to hear that the app works so well for you, Jørgen. Thanks for sharing!

  • Mia Willener

    We simply asing one agent a day to take care of incoming emails (answer or transfer to the right person) so the others only have to deal with their own tickets (and incoming phone calls).

  • Jonathan


    For the first reply time, every new ticket is notified to our Hipchat room. Every reply to a ticket notify the agent @FirstnameLastName on Hipchat room

    We also gathered different templates and use case to improve our one-touch resolution of a ticket. We're working on publishing a KB through Zendesk Guide to help on this.


  • Gabe

    We are improving our first response but we are small and its already under 2 hours most  times M-F.

    We don't try to do a 1 touch response, we use a video tool to walk customers through the help process and ask if it solved their issue. So if re-opens a closed ticket but 98% of the time we get the issue solved in 1 ticket.


    We also use a fairly detailed Text KB in Zendesk and success video channel on yourtube we link to articles.



  • Nicole - Community Manager

    Thanks for sharing, Gabe. 


Please sign in to leave a comment.

Powered by Zendesk