iSixSigma

Sample size for call center transaction monitoring

Six Sigma – iSixSigma Forums Old Forums General Sample size for call center transaction monitoring

Viewing 11 posts - 1 through 11 (of 11 total)
  • Author
    Posts
  • #38365

    jb2107
    Participant

    If agents in a call center handle 100 phone calls per week and the center handles 10,000 calls per week, how many phone calls should be monitored to accurately determine agent-level and center-level performance?

    0
    #114684

    Ken Feldman
    Participant

    Depends on the nature of what you wish to measure.  Is it continuous like AHT? or discrete like utilization?

    0
    #114686

    Mike Carnell
    Participant

    jb2107,
    I don’t understand how you can evaluate agent or center performance by number of calls. That is a volume metric not a performance metric. I am sure if you center has the volume I can take 100 calls per day. you probably won’t have those 100 customers again but does that make me a high performer?
    Any chance you can work on eliminationg the cause of the call – not the customer – the reason the customer had to call?
    Just my opinion.
    Good luck.

    0
    #114687

    Ken Feldman
    Participant

    There are some characteristics of the call that can be a reflection of the Associate performance.  You are absolutely correct that a great project is to collect data on the reason for the call in the hopes of eliminating the reason for the call and thus reducing call volume.  Gee, it would be a shame to lose all those overseas call center jobs.
    Mike, hope all is good with you but you still haven’t called when you have been in town.  8th Street is waiting for us.

    0
    #114697

    Tajjammul Hussain
    Member

    I think your question was
    Numbers of calls that should be monitored to determine the Agent and Center level Performance (roll up of agent level performance).
    Option 1:
    If you have Historical Data , You may use the Sample Size Calculations
    Option 2 :
    If Historical Data is not available, then the best practice is to monitor atleast 1 call per agent per month (if the monitors bandwidth is limited) else monitor 1 call per agent per week.
    Once you have the data, then you determine the optimal sample using this data.
     
    Regards,
    Hussain.
     

    0
    #114726

    Mike Carnell
    Participant

    Darth,
    The reason I made the comment was because of a story from our friend that got married in the Dominican Republic. He was working a call center and watched a guy spend 9 minutes (may be poor recall here) with a customer and left them completely satisfied (in terms of the call). The problem was they had a metric that was built around an average 5 minute call (again the numbers may not be correct). When PT asked him what he was going to do he picked up the next call and immediately disconnected it. Now the metric said he was ok but the result is 1 satisfied customer and 1 dissatisfied customer.  The metric was driving the wrong performance and the person wanted to do the job correctly – the long call.
    The question comes down to what is the CTQ on a call center. There does need to be some level of speed otherwise you end up with a shadow organization the same size as the organization that delivers products. You may also get the same result if you ignore the pure volume of calls because they are a Y not an x.
    I haven’t made it to Me-Am-Ee in a while. We are closing out Wave I in RSA and we have some great talent. The CEO just had a one on one with the class for about 2 hours and then had dinner with the entire class. No long speechs – just one on one conversation. They are pumped and what more can I ask of the CEO. I need to stay focused for another 2-3 months to get this Wave over the finish line.
    Christmas was magic. I hope consulting is treating you well.
    Regards,
    Mike

    0
    #114727

    Mikel
    Member

    Knowing what I know about call centers, I will refuse to let the rep off of the phone if they are not satisfying me (formally know as the Lupienski method).
    PS – Too bad you did not make Miami, me either, but I am making at least part of bike week. I just had a real nice conversation with Scott’s dad. Sounds like you are doing a good thing there.

    0
    #114733

    Mike Carnell
    Participant

    Stan,
    A fate worse than death – John L calling into a call center. It would be interesting to see someone actually review those “recorded for quality purposes” tapes. It could be a significant social event.
    I believe guy in the story never actually said anything – just connected and disconnected – he knew it would bring the average down. I guess if you have to decide between doing a poor job with 2 customers to meet the metric or satisfy one and hope the other one just thinks it was an equipment malfunction was a decent strategy. Ultimately the metric creates those choices. Without the metric he might have actually satisfied both? There is a concept.
    It might be interesting if you had 2 call centers to take that BS metric off of one of them and see how the numbers shift. Is there a significant difference in time, number of dropped calls and then if someone was really energetic they could do follow-up calls and see if the customers were better satisfied. Going back to the original post – can a call center person actually satisfy 100 people per week? Where did that number come from or is it just a number that they came up with (dollars) that they were willing to throw at problems so they didn’t actually have to figure out what was wrong and fix something.
    You would think that anyone with an ounce of common sense would drive to reduce the root cause of the calls so the volume would stay low enough so they could let people actually do their job – satisfy the customer. The metric around duration of call or number of calls seems more like a management imposed metric to contain the cost around a call center. They dress it up and stick it in the window and call it performance but at the end of the day most call centers are rework for some other part of the organization that is delivering crap and is managed by people who are as shallow as spit. Just my opinion.
    My scooter hasn’t been out in so long there wasn’t even enough juice in the battery to light the ignition lights. I am just hoping for a long hot run to Naw Olins over the 4th.
    I tried to escape to RSA through Miami but it was booked solid – every seat (the US airlines struggle while Virgin sells out). Hopefully some time this spring in Miami before the no seeums and mosquitos show up.
    RSA is going well. Scot is in his mid twenties and has 4 deployments under his belt. He does a great job.
    Regards,
    Mike

    0
    #114748

    Ken Feldman
    Participant

    Mike,
    I can see PT telling that story with his hands flying and voice full of animation and exasperation.  Of course, the key is balance between the handle time and the satisfying of the caller.  In most cases, both metrics are gathered.  The primary might be AHT but the secondary has to be customer sat.  Nowadays, sales is a heavily weighted measure.  Management wants the call length to be short, the customer delighted and some attempt made to sell them something while on the phone.  All of this from usually the lowest skilled, lowest paid and least trained employees in the company.  Ah, but you have to believe in miracles.
    Glad Xmas was better than you expected.  Santa was good to you.  Sounds like RSA is still going strong but wouldn’t it be nice to have a great client like that a little closer to home.  Maybe Santa will give you one next year….only if you are good.  The new job is interesting and I am keeping busy.  I’m learning something every day.

    0
    #114759

    sumant
    Member

    Hey, you can use some scientific sampling method. like for attribute data, you can use MIL STD 414D. it will give you no of calls per day you need to monitor based on the population(no of calls handle per day). there are 3 factors you need to identify. you can use all 3 standard levels of sampling….
    hope this clear….

    0
    #114780

    Mike Carnell
    Participant

    Darth,
    Glad it is working out. If you are learning something every day then you must be spending time with someone other than the owner.
    Actually we just picked up a new client in the nothern hemisphere but still about half way around the globe. The international stuff is a good time. I like the attitude you get in most companies – willing to do something for the good of the company.
    PT was animated as always with just the right touch of sarcasm. Great story teller.
    If management wants the call length to be short then the shoetest call is no call at all. That means instead of creating this new career call “call centers” they need to understand why people are calling and eliminate the need for the call. We have a device in one of our processes that separates little rocks from big rocks. Little ones move on and big ones go back to the mill to be made smaller. We have a constant battle over is it a VA operation. Kind of – because it is the most efficent way to run an incapable mill. We are optimizing a incapable process. Same scenario with call centers.
    Just my opinion.
    Regards,
    Mike  

    0
Viewing 11 posts - 1 through 11 (of 11 total)

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