Author Topic: Infomart / StatServer AHT discrepancy  (Read 4305 times)

Offline Adam_W

  • Full Member
  • ***
  • Posts: 171
  • Karma: 0
Infomart / StatServer AHT discrepancy
« on: April 01, 2011, 03:30:33 AM »
Yet another Infomart post; you can tell I'm slightly out of my depth here!

I'm working with a customer who uses WFM and Infomart/Interactive Insights for reporting.  There is an issue over the difference in Average Handle Time calculation over the course of a day.

WFM obtains its statistics from StatServer whereas Infomart works with events from T Server, so to an extent I would expect a difference in the figures, but its a sticking point in this case.  We've narrowed the difference down to 15-20 seconds per day by trying to make the calculations the same between the two sources, but the desire is that they will be exactly the same.

It's interesting that it's 15-20 secs each day and suggests there may be something I'm missing in the calculation.  I have no idea what.

Just interested to see whether anyone else has every tried to overcome the same challenge?

Offline mhsoomro

  • Newbie
  • *
  • Posts: 11
  • Karma: 0
Re: Infomart / StatServer AHT discrepancy
« Reply #1 on: February 26, 2016, 04:37:49 AM »
I am also facing the same issue with my customer. If IWS KPI from Stat server is showing 2:52 AHT then at the same time interactive insight report showing 3:58 AHT.
Please Help

Offline mhsoomro

  • Newbie
  • *
  • Posts: 11
  • Karma: 0
Re: Infomart / StatServer AHT discrepancy
« Reply #2 on: March 05, 2016, 04:38:37 PM »
R we going to get any response on this.

Offline nonny

  • Full Member
  • ***
  • Posts: 205
  • Karma: 0
Re: Infomart / StatServer AHT discrepancy
« Reply #3 on: March 06, 2016, 01:35:54 PM »
http://www.sggu.com/smf/index.php?topic=9099.0

Sent from my SM-N9005 using Tapatalk


Adam G

  • Guest
Re: Infomart / StatServer AHT discrepancy
« Reply #4 on: March 06, 2016, 06:04:17 PM »
Yet another Infomart post; you can tell I'm slightly out of my depth here!

I'm working with a customer who uses WFM and Infomart/Interactive Insights for reporting.  There is an issue over the difference in Average Handle Time calculation over the course of a day.

WFM obtains its statistics from StatServer whereas Infomart works with events from T Server, so to an extent I would expect a difference in the figures, but its a sticking point in this case.  We've narrowed the difference down to 15-20 seconds per day by trying to make the calculations the same between the two sources, but the desire is that they will be exactly the same.

It's interesting that it's 15-20 secs each day and suggests there may be something I'm missing in the calculation.  I have no idea what.

Just interested to see whether anyone else has every tried to overcome the same challenge?

OK.... here we go again.... deep breath, aaaand... :)

A StatServer gets it's info through being connected to a TServer.  Your statement that "WFM uses a StatServer and Infomart uses a TServer" is correct but Infomart uses an element called ICON to extract statistical data from a TServer.  So, in actuality, the source of the stats for both solutions is a TServer - but the path to extracting the stats is not quite the same.

There isn't a matrix of comparable units for the statistics across CCA|GI2|CCP|WFM|IWS because they all serve different purposes;

CCA, CCP and IWS generally count Volumes
GI2 generally counts Interactions
WFM generally counts Work Effort

Next, the way in which stats are calculated across these different solutions is not the same.  With one, it may calculate AHT based on whole numbers only (13 seconds) - whereas another may count the absolutes (12.56 seconds).

Finally, as a rule of thumb and a Best Practice, even using the same Platform which has 2 separate reporting sources, you should allow for up to 3% differential between stats.  You are looking for everything to be mathematically equal and - whilst it may be very satisfying to see that - it will not be.  In actual fact, without even knowing what your volumes are, if you have a differential of only 15-20 seconds in AHT for day, that is extremely accurate!

If you take anything away from this post, let it be this;

Genesys is not a Solution.  It is a Platform.  It can contain many different Solutions which were built with specific tasks in mind.  Ask yourself if you are using those Solutions for the individual purposes they were intended - or you are trying to manipulate data sets from different Solutions purely because you think you can.  The volumes, counts, statistics, averages and other calculations and algorithms within each Solution are applied to thousands - sometimes millions - of units for measurement, so there are bound to be some small scale differences.  So, really - don't sweat the small stuff, it will be a fruitless effort for very little gain.

Adam

Offline mhsoomro

  • Newbie
  • *
  • Posts: 11
  • Karma: 0
Re: Infomart / StatServer AHT discrepancy
« Reply #5 on: March 16, 2016, 03:47:17 PM »
OK, 3% differential is acceptable but 20% to 30 % difference is not acceptable anywhere in industry. let me tell you what we are facing in difference.

GII Report output:
Agent Name      Accepted   Avg Handle Time         Engage       Hold           Wrap
Agent 1   Inbound   81           0:02:17                   0:02:11   0:00:00   0:00:06
Agent 2   Inbound   54           0:02:08                   0:01:42   0:01:09   0:00:23
Agent 3   Inbound   71           0:04:35                   0:03:48   0:00:42   0:00:39
Agent 4   Inbound   24           0:03:03                   0:02:15   0:01:11   0:00:16
Agent 5   Inbound   103         0:02:31                   0:02:21   0:03:14   0:00:08
Agent 6   Inbound   81           0:03:31                   0:02:55   0:02:13   0:00:15
Agent 7   Inbound   80           0:02:56                   0:02:36   0:00:53   0:00:07


Agent IWS KPI's:
Agent Name   Calls   AHT
Agent 1            81   2:16
Agent 2            54   2:03
Agent 3            71   3:52
Agent 4            24   1:58
Agent 5           103   2:29
Agent 6            81   3:00
Agent 7            80   2:23


Above both are only Inbound results as this call center does not have outbound option. Some agents are having minor difference but most of them having major difference.

Below i will tell you the formula used in GII report for AHT and formula used in IWS_StatServer Stat of AverHandleStatusTime

GII Report AHT Formula:
sum(AG2_AGENT_DAY.ACCEPTED) )+(
 sum(AG2_AGENT_DAY.CONSULT_RECEIVED_ACCEPTED) ))<>0
 then 1.0*( (
 sum(AG2_AGENT_DAY.ENGAGE_TIME) ) + (
 sum(AG2_AGENT_DAY.WRAP_TIME) ) + (
 sum(AG2_AGENT_DAY.HOLD_TIME) ) + (
 sum(AG2_AGENT_DAY.CONSULT_RECEIVED_ENGAGE_TIME+AG2_AGENT_DAY.CONSULT_RECEIVED_HOLD_TIME) ) + (
 sum(AG2_AGENT_DAY.CONSULT_RECEIVED_WRAP_TIME) ) + (
 sum(AG2_AGENT_DAY.CONSULT_RCV_WARM_ENGAGE_TIME+AG2_AGENT_DAY.CONSULT_RCV_WARM_HOLD_TIME) ) + (
 sum(AG2_AGENT_DAY.CONSULT_RCV_WARM_WRAP_TIME) ) )/((
 sum(AG2_AGENT_DAY.ACCEPTED) )+(
 sum(AG2_AGENT_DAY.CONSULT_RECEIVED_ACCEPTED) )) else 0 end


IWS_StatServer Stat of AverHandleStatusTime:
Category: "AverageTime"
MainMask: "CallInBound, AfterCallWork, CallInternal, CallConsult, CallOnHold"
RelMask: "CallInBound"
Subject: "AgentStatus"

Now please help us matching atleast with 3% difference. Please note that i can only change the IWS Stats values.

« Last Edit: March 16, 2016, 04:10:26 PM by mhsoomro »

Adam G

  • Guest
Re: Infomart / StatServer AHT discrepancy
« Reply #6 on: March 16, 2016, 08:30:16 PM »
You might want to take a look at the definitions for the IWA "ACW" versus the GI2 "Wrap".

Even better, you can extrapolate each parameter (Inbound vs. Engage, Hold vs. Hold and ACW vs. Wrap) and compare them on their own, to see where the anomaly is.


Offline mhsoomro

  • Newbie
  • *
  • Posts: 11
  • Karma: 0
Re: Infomart / StatServer AHT discrepancy
« Reply #7 on: March 16, 2016, 09:07:04 PM »
I have been trying but no success.

Adam G

  • Guest
Re: Infomart / StatServer AHT discrepancy
« Reply #8 on: March 16, 2016, 10:06:01 PM »
Not much else I can advise but keep bringing things back to their basics and compare pure stats from each solution... you are bound to come across the anomaly at some point.

One thing I did notice: CallInBound is not a generic/pure stat name - it's CallInbound (no capital "B").  You might want to take a look at that...
« Last Edit: March 16, 2016, 10:11:02 PM by adamgill »

Offline cavagnaro

  • Administrator
  • Hero Member
  • *****
  • Posts: 7172
  • Karma: 56330
Re: Infomart / StatServer AHT discrepancy
« Reply #9 on: March 16, 2016, 10:34:47 PM »
Just did and doesn't work. Help me fix it.

Offline mhsoomro

  • Newbie
  • *
  • Posts: 11
  • Karma: 0
Re: Infomart / StatServer AHT discrepancy
« Reply #10 on: March 16, 2016, 11:01:26 PM »
Capital B was a typo here. On production, it's small b.

Adam G

  • Guest
Re: Infomart / StatServer AHT discrepancy
« Reply #11 on: March 17, 2016, 12:18:19 AM »
Just did and doesn't work. Help me fix it.

???

Offline cavagnaro

  • Administrator
  • Hero Member
  • *****
  • Posts: 7172
  • Karma: 56330
Re: Infomart / StatServer AHT discrepancy
« Reply #12 on: March 17, 2016, 12:24:54 AM »
Just did and doesn't work. Help me fix it.

???
Lol anticipating the OP answer

Adam G

  • Guest

Offline cavagnaro

  • Administrator
  • Hero Member
  • *****
  • Posts: 7172
  • Karma: 56330
Re: Infomart / StatServer AHT discrepancy
« Reply #14 on: March 17, 2016, 01:49:04 AM »
Told you, nothing new on how to help him ;) Do magic stuff  ;D