Reference9:IQM Statistic Help English: Difference between revisions

From innovaphone wiki
Jump to navigation Jump to search
mNo edit summary
mNo edit summary
Line 134: Line 134:


The loading and processing of a week's data takes about 7 seconds.
The loading and processing of a week's data takes about 7 seconds.
== Build 80311 ==
Starting with Build 80311, the individual data sets can also be displayed numerically:
[[Image:Stat02.PNG‎]]
In addition, a monthly statistics display is possible: It takes about 30 seconds to load and display.
The loading time is shown with a progress bar for both the weekly and monthly statistics.


=Related Articles=
=Related Articles=

Revision as of 11:10, 22 February 2022


Description statistics iQM Dashboard (from build 80308)


Overview

DashbGraf 01.png

1 Visualization of today's data, in real time every 15 minutes a measurement will be added. You can switch between a date view and an online view at any time

2 The data of the date set below will be displayed, if there is no data the windows remain empty.

3 Increase the day

4 Date displayed (14). The date can be modified and confirmed by sending or by pressing key (2)

5 Decreases the day

6 Display of 24 hours (from 00:00 to 24:00) and 96 samples or 12 hours (from 07:00 to 19:00) and 48 samples.

7 View the number of agents (21)

8 Displays the number of calls transferred (not answered and closed by the agent but transferred to other extensions)

9 Displays the number of outgoing calls from agents

10 Total time of waiting in the queue of all calls in the period. The time is indicated in minutes and hours, with a key it is possible to switch to display in seconds.

11 Peak waiting time

12 Average waiting time

13 Peak number, the graph automatically aligns (auto-scale) to the values.

14 Date (day) displayed. The language depends on the operating system settings

15 Total number of calls viewed (= 100%)

16 Total number of calls answered (QoS), absolute and in relation to the total number of calls

17 Total number of missed calls (QoS), absolute and in relation to the total number of calls

18 Average calls received in the sampling period (15 minutes)

19 Average calls answered in the sampling period (15 minutes)

20 Average missed calls over the sampling period (15 minutes)

21 Indication of the data shown in the graph below

22 Open settings, in English, password protection option

23 Peak calls, the graph automatically aligns (auto-scales) to the values

24 Switch between statistic and dashboard views

25 Minimize in the task bar

26 Close application

27 Call threshold exceeded

28 Average number of calls per agent

29 Average call duration

30 Week data display


Below is a description of the data displayed in the statistics and how the processing takes place.

Sampling period

The iQM server accumulates statistical data every 15 minutes and stores the sampling. For each day there are up to 96 records (24x4). Writing always takes place in delay. The data displayed at 8:00 am is the accumulated data from 7:45 am to 8:00 am, the data displayed at 8:15 am contains the accumulated data from 8:00 am to 8:15 am etc. At the time of sampling, the counters are reset to zero with the sole exception of the peak value (see below). The average call values ​​(15, 16, 17) are the total calls divided by the periods where at least one call is present. The value represents the average number of calls received in the “active” sampling periods. The division of the total number of calls is therefore not fixed at 96, but equal to the number of sampling periods with traffic.

Example: With a total number of calls equal to 26 and sampling periods equal to 14, the sampling average is equal to 26/14 = 1.857. Rounding to one decimal takes place and the displayed value will be 1.9 (average number of calls in active sampling periods). Analogously, this also applies to the average calculation of answered and missed calls.

Auto Scale

The size of the graphs always remains the same while the scale varies according to the maximum value detected. This maximum value is shown in the graph (23, 13). The ladder and the support lines are calculated automatically. If for example the highest number found is equal to 52 this value will be displayed and there will be 5 support lines indicating 10, 20, 30, 40 and 50. For example, if the value is equal to 2.996 there will be a line at 1.000, 2,000 and 3,000 etc.

Incoming calls

Each call that enters the queue increments the counter of incoming calls shown in the upper window. If the call is not answered but the caller hangs up, the missed calls counter is increased. The number of answered calls is the mathematical difference between incoming calls and missed calls. Consider that the counter of the number of incoming calls is incremented at the moment of entry into the queue while the counter of the lost calls is incremented at the moment of release. As explained the difference of these two values ​​is interpreted as called answers, it follows that there can be differences.

Example: All the counters are zero, a call comes in at 08:14, the operator does not answer and the caller continues to wait. At 08:15 the graph displays a call answered for the sampling period 08: 00-08: 15, even if in reality the outcome of the call is still uncertain because it is still in the queue. If the operator then answers at 08:16 this call will not be displayed in the subsequent sampling period 08: 15-08: 30 because it has already been quantified in the previous period. If, on the other hand, the call is not answered in the following period, there are two cases: if in the following sampling period (therefore from 08: 15-08: 30) a call enters, it will be counted as not answered regardless of whether it is actually answered or not. If, on the other hand, no further calls come in, they will not be counted / displayed. The total number of calls is therefore always correct because calls are never counted twice.

With a large number of incoming calls, the effect is balanced and the purpose of assessing the quality of the service is still given. The system typically counts in favor of answered calls, counting only in special cases unanswered calls as answered and only if there are no calls in the next sampling period. If, on the other hand, there are calls in the following period, the non-answer eventually passes through the following period on an answered call.

Active agents (7)

Active agents are logged into the hunt group at the time of sampling and also have not set any presence status. Any log-in and log-out or changes in the presence status within the sampling period are not counted. The value shown shows the number of agents available for the service in the last second of the sampling of every fourth hour.

Calls transferred (8)

Calls transferred by agents coming from the call queue are counted. Typically they are calls where the Agent was unable to answer questions or the caller needs other extensions and therefore the Agent passes the call.

Outgoing Calls (9)

Both internal and external calls are considered outgoing calls, regardless of whether or not an Agent is logged in at the time of the call in his queue pickup group.

Duration of accumulated wait (10)

This counter displays the total waiting time of all calls. It starts counting when a call arrives in the queue and ends when the call is answered or the caller is released. This concept is very simple for a single call, if for example a call enters at 08:00 and ends at 08:17 the period from 08:00-08:15 will show 15 minutes of waiting while the period from 08:15-08:30 displays 2 minutes. But in the case of multiple calls, everything becomes complicated because the waiting time of the caller who waits the longest will always be displayed. Even with more calls in the queue, the counter advances at the rate of one second. If the current call is answered, the waiting time counter displays the value of the call that is now presented as before. It follows that the total will be increased suddenly with the value of waiting for the call that it is now presented. Note that this also applies to calls that give up: at the moment of release the counter will be increased with the value of the seconds that were waiting.

Example: there are two calls, at moment zero the first call comes in, the second call comes in 10 seconds after the first. After 30 seconds the counter shows the value 30 and the operator answers the (first) call. The second call now becomes the current queue call. The second call has now been waiting for 20 seconds in the queue and then the counter is incremented from the previous value 30 to the new value 50 (30 + 20) and thereafter incremented every second. The same thing happens and at the moment 30 the second call gives up and releases.

The value indicated in the sampling period is the accumulated time of all waits. At the time of sampling, the total of the counter is then saved, showing the total in the sampling period of all waiting times for answered or missed calls.

If the accumulated time passes in a sampling period without further incoming calls it is lost.

Example: All counters are zero, a queue call enters at 16:14 and is answered at 16:17 in the next sampling period. The statistic will show a call in the period from 16:00-16:15 and an accumulated waiting time in this period equal to one minute (from 16:14 to 16:15). If in the following period no more calls come in, the two minutes of waiting will not be displayed. If in the following period at least one call comes in, the two accumulated minutes will be added.

Maximum continuous wait (11)

Continuous waiting is distinguished from all other values ​​in the transition from one sampling period to another. In the event of an active call with an increase in the peak value in progress, it is not reset in the passage of the period but shows the value in the following period. It therefore gives an indication on how much the queue is interrupted in exceeding the historical maximum peak value. If the queue at the time of passing from one sampling period to the other is not increasing the maximum wait counter, the counter is reset and then immediately increased with the current duration in progress. Since the value of the waiting duration is zero at the moment of the passage, it follows that this counter will be increased again only in case of exceeding the maximum wait of the previous period. In the passage of the sampling period without calls in progress, the peak counter then returns to zero and will be incremented in parallel with the first incoming call.

Average waiting time (12)

The average length of wait in a sample period is the number of seconds to wait accumulated in the period divided by the total number of calls (missed and answered). As already explained (10) the representation of this time is influenced by the moment of the answer or the renunciation of the call, because only in this moment the accumulation of the waiting time takes place. Considering that for the calculation of the average the number of calls in the sampling period is considered, it follows that a single bar can give an incorrect indication on the global average. The set of averages, however, gives an indication of the lived expectations. It should therefore be considered that the number of incoming calls is increased when a call enters the queue while the duration will be counted only at the time of answer or redirection if there are more calls in the queue.

Call threshold exceeded (27)

In the configuration of the iQM (server) a maximum number of calls can be set, if exceeded the counter is increased. The value gives an indication of how many times the overshoot occurs in a sampling period.

Average number of calls per agent (28)

The number of calls answered in a sample period is divided by the number of agents. It is a pure mathematical calculation. It should be considered that the number of agents will be detected at the transitory moment from one period to the next. The average shown on the right is mathematical (obviously you will not be able to answer a fraction of a call). The value will always be an integral digit, no fractions are displayed. If the mean is for example 0.4 zero will be displayed, an average of 0.6 with the value one.

Average call duration (29)

The average call duration is, like the average number of calls per agent, a pure mathematical calculation. The number is the result of the call time divided by the number of answered calls. The value will always be an integral value, no fractions are displayed. If the result is for example 0.4, zero will be displayed, the value 0.6 as one.

Display dates of the week (30)

The relative week is calculated based on the date entered in field (4). Monday is always the first day displayed and Sunday the last. In the example below, the date entered is 03/02/2021 and the reference week starts on 01/02/2021 and ends on 07/02/2021. The display would be identical if for example 2/2/2021 would have been set as a date.


DashbGraf 02.png

The loading and processing of a week's data takes about 7 seconds.

Build 80311

Starting with Build 80311, the individual data sets can also be displayed numerically:

Stat02.PNG


In addition, a monthly statistics display is possible: It takes about 30 seconds to load and display.

The loading time is shown with a progress bar for both the weekly and monthly statistics.

Related Articles

Howto:IQM_CallBack_requests

Howto:Queue_Monitor_-_Overview

Howto:PBX-QueueMonitor

Howto:Queue_Monitor_-_Setup_and_Localization

Howto:PBX-QueueMonitor_Client

Howto:IQM_Dashboard

Howto:Additional_call-ID_information

Howto:IQM_IP-Phones_setup_and_iQM_Phone_Display_features

Howto:IQM_Second_Waiting_Queue

Howto:IQM_Active/Standby_PBX

Howto:IQM_manipulative_behaviors

Howto:IQM_Statistical_Excel_Data_for_Agent_and_Queue

Howto:IQM_Agent_Hot-desking

Howto:IQM_command_line_arguments

Howto:IQM_remote_control

Howto:IQM_server_as_a_service

Howto:IQM_Colour_Agent_Table

Howto:IQM_in_large_PBXs