iSixSigma

Unable to solve the issue … pls help..!

Six Sigma – iSixSigma Forums Old Forums General Unable to solve the issue … pls help..!

Viewing 10 posts - 1 through 10 (of 10 total)
  • Author
    Posts
  • #43565

    Rajesh Mohandas
    Participant

    Hi,
    I’m held up with a very odd issue with respect to the sample size.
    I have 25 agents in the team.
    The Average C-Sat score of the team is 85%
    All agents dont have the same number of surveys which makes it difficult to identify which agent is doing good and who should be in focus or who should fall in improvement group.
    Say there is an agent who has a Csat of 95% but the number of surveys are only 10
    and other agent has a Csat score of 90% with 15 surveys
    and another one has a Csat of 85% with 30 surveys and so on.. the sample size is not constant.
    I wish to understand if an agent needs to be considered underfocus group or If I have to point out saying Agent A is good compared to Agent B how will I do that …..?
    Rajesh Mohandas

    0
    #138352

    Pradeep Sharma
    Participant

    In case of CSAT, focusing on agents is not advisable, focus should be on the process. Identify issues or calls which are getting you DSATs, focus on them. Because your agent population can change any time but your call type will remain same. Do root cause analysis on the calls/ issues which are driving DSATs. Tools used for such scenario can by
    1. Pareto Analysis 2. Fish Bone 3. 5 Why Analysis

    0
    #138356

    Rajesh Mohandas
    Participant

    Thank you Pradeep, I get your message, from the project prespective I will definately follow what you told. But this was a concern raised by my QA and is the same problem when it comes to Monitoring calls, Say an Agent gets 80% on 4 calls monitored and 85% on 2 calls monitored and 70% on 10 calls monitored. ….
    I wish to come up with a neutralization formula which can help me treat all agents on a common plane and over come the crib in the team…
    Is it possible…, if not then is there any alternative for the same.
    What I suggested my QA is to monitor 5 calls each and accordingly form the focus group it was then the C-sat score and all related issues occured…
    Regard,
    Rajesh.

    0
    #138361

    Ken Feldman
    Participant

    In the case of evaluating call center agents, your first task is to determine whether the measurement system is good enough to distinguish between good and bad agents.  How are you determining your cust. sat. scores?  Percentages calculated on small sample sizes can be very deceiving.

    0
    #138362

    Felix
    Participant

    To my Knowledge,
    It is clear that the Accuracy % (CSAT in your case) of the agents are visible, and you should always see these data in terms of Accuracy delivered rather not the number of calls attended (no of calls attended is measured as Productivity).
    For eg, If ‘A’ assembles 5 Cars with the Accuracy of 80% and ‘B’ assembles 3 Cars with the Accuracy of 95% then ‘B’ should be honored in front of Accuracy %. Although number of cars assembled is less by ‘B’ that totally falls under Productivity and not Accuracy.
    So my suggestion is don’t confuse with the Productivity (No of calls attended) and the Accuracy (CSAT score). If Productivity is less try improving that by a 6sigma project and if accuracy is less do the same (Use Pareto to Prioritize and do root cause analysis to attack on major issues)
    May be you can design a Point system for the agents which will integrate Accuracy & Productivity and project a neutral point for the agent.
    Sorry could not discuss much as I’m hurrying for a meeting, do get back incase of anything else.
    Regards,
    Felix
     

    0
    #138367

    Peppe
    Participant

    Hi Pradeep, it could be useful to run a test for samples homogeneity and then a test for differences between proportion.

    0
    #138368

    Rajesh Mohandas
    Participant

    Hi Felix,
    I guess, I have hit at the point. Can u pls help with a formula or a standard so that the neutralization happens.
    One of the friends on the fourm suggested to run a homoginity of variance and then perhaps identify if difference exists and action accordingly.
    Regards,
    Rajesh Mohandas.

    0
    #138376

    Pradeep Sharma
    Participant

    Hi Peppe, would like to know more about it. How do we do it?
    Thanks in advance

    0
    #138379

    Pradeep Sharma
    Participant

    Hi Rajesh,
    In this scenario keep following things in mind
    1. Have standard number of calls monitored for all the agents
    2. Align your QA form with your CSAT questions (mostly QA score does not corelate with CSAT scores as QA form is not alligned to CSAT questions. Foremost and important thing is to allign your QA form with CSAT questions)
    3. Based on your CSAT or QA scores identify your focus group, increase the numner of monitorings for them, you can reduce number of monitorings who are consistently scoring good to tackle resource availability)
    4. When you have identified focus group, you should have focus area identified for the focus group (again I will say based on issues you idetify which are driving DSATs), need not to do project for this simply map your DSAT with call driver, run pareto on it. Do root cause analysis for top 30% of the drivers, we use 30/70 instead of 20/80 in call center scenario (tried and tested).
    5. Do focus on Issue Resolution and then soft skills
    I hope this will help, if not then go for full fledge SS project

    0
    #138393

    Romel
    Member

    You need to identify first the margin of error and standard deviation and check if you are having the right sample size. This approach can be done two ways — either as a whole group comprising of different agents or on a per agent scenario.
    If you want to see the significance of each agent of concern irregardless of varying sample size, you can hypothesize using ANOVA or have chisq test of independence to see if agents are enabling the csat scores

    0
Viewing 10 posts - 1 through 10 (of 10 total)

The forum ‘General’ is closed to new topics and replies.