Outbound - which VDN on ACD

[astrakid]

Hi,
regarding outbound we are facing some problems: the wait time for agents from call release until next incoming call is too high, sometimes up to 5 minutes.
it seems, that everything is configured the right way. do you have any idea, which parameter could influence this behaviour? currently we have about 20 agents doing outbound in parallel per location, and each location is handled separately (realized by a "location-skill").
regards,
andre

[Fra]

Andre,

lots of infos is missing:
# which mode are you running? Predictive, progressive?
# how is the routing done?
# the fact the agents get a call every 5 mins doesn't mean OCS isn't dialing. What's the hit ratio?

Fra

[astrakid]

Hi, sorry for missing infos:
we are running campaigns in predictive mode, the strategy is searching for the most idle agent at that location (each location has the same vdn).
The hitratio is rather low, but we are not sure why, because the datasource is the same for the last months, and before using ocs another dialer was used with much better results...

regards,
andre

[bcyk]

All outbound calls entering VDN / ACD should have statistics
Check CCPulse+ if there is any connected calls dropped in VDN/ACD.


A better way to identify "slow" dialing pace is described as follows.


1. Identification of bottleneck
  - create Campagin Group view in CCPulse+; the default view is fine
  - In Performance category, fields are good indications
SystemError 00:00:00
DialMode 12:23:11
WaitingAgent    00:00:00
WaitingPort    00:00:00
WaitingRecords  00:04:33
    - above example indicates the dialer is waiting for dialing records

  - the real-time panel at the LHS is another shortcut
  - select interested Campaign Group(s), right-click mouse button and select
    "Monitor Extended Current Status"
  - In Extended mode, comprehensive campaign states are shown
StatusDeactived
StatusActivated
StatusRunning
  ModePredict
  ModeProgress
  ModePreview
  StatusWaitingRecords
  StatusWaitingPorts
  StatusWaitingAgents
  SystemError


2. Waiting Resources
  - Waiting Agent, Waiting Port and Waiting Records
  - As their names implied, do whatsoever to make OCS dialing

3. Low hit rate
  - A common issue for 'not-so-good' calling list (invalid contact number, etc)
  - Assuming "Predictive" mode is being executed, dialing pace can be adjusted
    online in OCM
  - The "Optimized parameter includes "Overdial Rate", "Agent Busy Factor" and "Average Waiting time"
    the "Average Waiting Time" is easier to explain; all parameters may be restricted by other option
    -> to keep the agent "Average Waiting Time" to n-second, e.g., 40 seconds

  - a compaign-level option "predictive_max_overdial_rate" may limit the ceiling overdial-rate
Specify the maximum allowed over-dial rate for dialing in predictive mode
Checking is performed separately from optimization parameters in predictive mode
If set to 0, OCS does not perform separate control of the over-dial rate

  - OCS server uses 'overdial' rate (i.e., % of dropped calls due to connected calls terminated by dialer or ACD
    when there is no ready agent) to control dialing pace
    => enable full debugging log in OCS and see these statistics

  - in some countries, auto-dialer's "overdial" rate and duration of detection
    (see option call_timeguard_timeout at campaign level) are restricted by local authority!
    e.g., overdial <= 2% and duration of detection <= 3 seconds

  - if overdail rate reaches the maximum value, OCS changes dialing mode from Predictive to Progressive
  - OCS server keeps campaigns statistics in memory as long as it is loaded/running
  - unload and load/run campaign reset these internal counters and may help to dial faster

  - If overdial rate is not a concern, set "predictive_max_overdial_rate = 20" (or higher!) and
    set "Average Waiting Time" to 30 (or smaller) seconds and observe results and behavior
    ***WARNING*** you may have "Racing" condition to front-end application/CRM if
                  CPD server is using 2-step transfer (i.e., CPD controls call_wait_in_queue_timeout value)


Regards

[astrakid]

Thanks for all these infos.
I checked already several things.

we have an overdial rate of 80%, but waiting times up to 5 minutes for the agents. but I made the following observation in the last two days:
we have 3 locations for outbound. one campaign is running on each locations, calls are handled by each location, that means, if an outbound call is done by tserver in location A, then an agent in location A has to get the call (it is done by a "location-skill"). In the morning we have lower idle times for the agents, most of the time around 1 minute or less. After launch-time, when more agents are logged in, the waiting time increases.

Is this a normal behaviour and would it make sense to change the parameter from overdial rate to average waiting time (30s or 60s)?

we are talking about 40 agents in maximum per location, in total not more than 80. for each location we have 10 CPD Ports configured. maybe this is relevant.

kind regards,
andre

[bcyk]

There are 10 dialer ports for 40 logged-in agents.
It may affect average waiting time.

The best way to confirm is to check CCPulse+ Campaign Group view.
Accumuated value "WaitingPort" is a good indicator.

BTW, historical Campaign Group view, if not disabled, reports contain similar statistics.


So the optimized parameter in predictive mode is set to 80% in OCM.
It is a pretty high value!

Please also check the option "predictive_max_overdial_rate".
OCS server internal changes predictive mode to progressive if the accumulated dropped rate is equal to or greater than "predictive_max_overdial_rate".

No matter what optimized parameter is selected in OCM, the above algorithm is applied. (Genesys OCS 7.2 or higher).


=> morning session: 1 minute Average Waiting Time (AWT)
  afternoon session: 5 minutes AWT

If campaign(s) are stopped and unloaded in OCM at Day-End, above observation may be explained as follow.

End-day (or before next working day): stop and unload campaign
  -> OCS server zeroes all internal statistics, including the dropped call counter)

Morning session: load and run
  -> dropped call count is zero; thus, predictive mode is mostly in action
Afternoon session:
  -> dropped call count is being accumulated to reach "predictive_max_overdial_rate"
  -> predictive mode will be changed to progressive internally (although OCM / CCPulse+ are shown as predictive)


To identify the root-cause, OCS server debugging log files are very helpful.


regards

[astrakid]

Thanks again for those great infos. How can I enable full debug mode?
In CME we have configured verbose="all" and all="/.....". I thought that verbose-mode all would mean debug-mode, but in ocs-log I can't see any debug-messages...


regards,
andre

[astrakid]

I rechecked documentation and genesys support site, according to that everything is right. Here also an extrat of startup-logfile:

              'log'(list) 'trace'    ''
                            'expire'    '10day'
                            'buffering' 'true'
                            'segment'  '10000'
                            'interaction'      ''
                            'debug'    ''
                            'verbose'  'all'
                            'standard'  ''
                            'messagefile'      'OCServer.lms'
                            'all'      '/***************/ocs'

But not one debug-message in logfiles...
kind regards,
andre

[astrakid]

Ok, maybe I found some kind of statistics you meant:
CurrentStatType: CallDialed; State: CallUnknown; CPDCompletedFlag: 0; CompletedFlag: 0; CallResult: 28;

Is this the right stat? That would be fine, because it is easy to extract to perform some own statistics...

kind regards,
andre

[astrakid]

And it's me again... sorry.
I found out that someone changed CPD ports from 10 to 50 2 days ago already. After I read this
http://solutionsearch.genesyslab.com/selfservice/microsites/search.do?cmd=displayKC&docType=kc&externalId=187079529&sliceId=1&docTypeID=DT_SOLUTION_1_1&dialogID=1023319&stateId=0%200%201019763
I checked our logs and could find error logs mentioned on genesys support site (lack of CPD ports). those messages are gone now. Unfortunately I am not able to see the hit ratio in CCPulse anymore - it shows 0.00, although agents are logged in and getting outbound calls... beside this I don't have the standard views for ccpulse available. Another department creates the workspaces and so I have to check if it is possible to get standard-views...
however, i have one further point:
campaign groups in CME are configured with "average waiting time", but in ocm "overdial rate" is configured. is this a matter, or does it make no difference?

regards,
andre

[bcyk]

Dear Andre

Standard OCS statistics and templates are installed during Genesys suite installation/confiugraiton.
Besides CCPulse+ template, statistics fields from Stat Server must be pre-configured.
  e.g., "StatProfile_OCS Reporting Templates.cfg" in the realtime metric installation CD media

CCPulse+ campaign, campaign group and calling list views are useful to monitor healthiness of outbound activities and results.
BTW, once "StatProfile_OCS Reporting Templates.cfg" has been configured in Stat Server, standard CCA historical layouts for campaign, campaign group and calling list will not be available (unless CCA/DMA is configured to disable new templates layout creation!)


[color=red]campaign groups in CME are configured with "average waiting time", but in ocm "overdial rate" is configured. is this a matter, or does it make no difference?[/color]

No. Present values in OCM override the default CME settings.



In OCS debug log, search the keyword "[b]Predictive data[/b]".
Samples are extracted for reference.

notes: campiang name was marked to XXXX
- the "two-step" transfer method was applied in this case
- the system was changed to "single-step" to avoid "racing" condition when over-dail rate is higher than 20%
- "Pacing algorithm:" can be "General" or "Small Group Enhanced"
  -> "Small Group Enhanced" setting (and Pros & Cons) is another story!
- "[u][b][font=Verdana][color=red]Overdial Limit[/color][/font][/b][/u]" is the value from OCS option (at campaign level) predictive_max_overdial_rate





Group Configuration - Type: AgentGroup/Regular; AgentsByConfig: 16; AgentsLoggedIn: 11;
GroupCampaign List:
  Campaign - DBID: 102; Name: cm_XXXX; WorkTime: 628378; Mode: Predictive; Status: GSW_CM_GCS_Running;
        Optimization Parameter: [b]Average Waiting Time; Optimization Goal: 20[/b]; ProgressiveFlag: 1;
        VoiceDest: 300; SentCalls: 9; DialedCalls: 1; SentEngageCalls: 0; NoRecords: 872;
  Configuration - Dialer: CPD Server (DBID: 120; SwitchDBID: 101; SwitchType: 29;) Dialer Type: Transfer/Value for option "call_transfer_type" is requested for //CfgCampaign/cm_XXX(102)
Option found in: //CfgApplication/OCServer(122)/Options/OCServer/call_transfer_type = "two_step"
two-step; StatServerDBID: 0; [u][color=red][b]Overdial Limit:  6.00[/b][/color][/u];
        CPD Ports: 10; Number of Free Ports: 9; Total channels: 40; Free channels: 37;
Predictive data
    Calls predicted: 1; [b]Pacing algorithm: General[/b]
    Expected values:  Busy Factor= 0.852; [font=Verdana]Abandon Rate= 0.334[/font]; Wait Time= 20.000;
    Real values:      Busy Factor= 0.802; [font=Verdana]Abandon Rate= 0.067[/font]; Wait Time= 30.316;
    Current Statistics
    Agents/Places - Total:    11; Ready:  2; Engaged:  0; BusyOut:  7; BusyInB:  0;
                    BusyPaper: 0; BusyInt: 0; BusyRing: 0; BusyDial: 0; NotReady: 1; NotAvail: 1;
    Calls Outbound - Dialed: 1; Queued: 0; Established: 8; Chained: 0;
    Calls Callback -            PersonalClbk: 0; CampaignClbk: 0;
    Calls Inbound  -            Queued:  0;  Established: 0;
    Chains - Created: 9561; Deleted: 9112;    Records - Created: 9561; Deleted: 9112;
    History - Hit Ratio: 0.693;
        Durations - Out: 114.95; In: 13.57; Total: 114.95; Dial: 16.70; Dial NoContact: 22.27;
        Traffic  - Out: 0.079; In: 0.000; Other: 0.000;  Last Forecast: 0; Really Dialed: 0;
        Inbound Calls  - Completed: 5; Abandoned: 3;
        Outbound Calls - Total: 9513; Established: 5096; Completed: 5088; Counted: 5081;
                        NoContact: 3881; NoContactCounted: 3881; Abandoned: 363; AbandonedCounted: 360;
                        Error: 175; Stale: 1;



regards