SpeechMiner comes with a number of standard report templates. These report templates are described in this section For information about configuring and generating reports from templates, see Reports.
|
This section explains the general layout of report templates. The Common Parameters section, describes parameters that appear in many report templates. These sections are followed by the Templates section, a catalogue of all of the report templates, with descriptions of each report and additional information about the reports. When specific templates use significantly different layouts and parameters from the most common ones explained below and in the Common Parameters section, they are explained in the Templates section in the description of the relevant report template.
The report templates in SpeechMiner have a maximum of three rows of parameters, each row dealing with different functions in the report. In addition, most report templates have a Data Set filter section on the left side of the screen. The layouts of most templates are similar to the Agent Comparison template shown below.
In this section, we will explain the layout of report templates using the Agent Comparison template as an example, since it has the most common layout of parameters you will encounter.
The following rows appear on the right side of the template:
Contains controls for working with report results (see Creating a Custom Report), followed by the Template field, in which you can select the type of report, and the Report Name field, which you can use to name the report. If you choose to save the report parameters as a saved report (see Managing Saved Reports), the Report Name is used to identify it in the system.
Contains fields for defining the report title and an optional report description. These items are displayed at the top of the report results. By default, the name of the report template is used as the report title, but you can modify it as necessary. Some templates also have a Version parameter in this row. If so, you can use this parameter to select the size or format of the report output.
Contains the Items on Report parameters - the fields that determine which items will appear graphically on the report. In some reports, one or more of the parameters may also have statistical functions.
The left side of most report templates contains the Data Set filters. These filters specify which data will be included in the report's analyses. For information about the filters and how to configure them, see Common Parameters.
When you are not configuring the Data Set filters, you can hide the left panel of the screen if you wish.
To hide or show the Data Set filter panel:
The current filter settings are displayed on the right side of the template below the first row:
This section contains detailed information regarding each report template, including a general description, an overview explaining how to customize the report (how the parameters function), and a sample report. For some templates, additional information on common usages and actions you can perform on the reports is provided.
Most reports are interactive: you can drill down within a report to view additional details about any graphic component (graph bars, lines, or table headers) by clicking on it. In addition, you can play a call in the Media Player directly from an Interaction List in the report results. These interactive options are called actions. In some cases, report results offer additional action options. In these cases, the description of the report includes as Actions section describing the additional actions that are available. For more information about using the standard report actions, see Creating a Custom Report.
In some cases, reports have a number of available versions, such as a Full version, a MINI version, and a Wide version. In these cases, a single template is used to generate all of the versions, and a Version parameter is used to select the desired version. (For additional information about the Version parameter, see Common Parameters.)
|
The following standard report templates are available. Select a report template for more information.
The Agent Bubble Chart report compares Agents and/or Workgroups by plotting them on a graph according to the percentage of calls in the horizontal (x-axis) and vertical (y-axis) Categories. Each Agent is represented on the graph by a bubble. The size of each bubble represents the average call duration.
The X-Axis Only and Y-Axis Only versions present the data as a histogram indicating percentage of calls per Agent and/or Workgroup, rather than a bubble chart.
| This template is only available in an Analytics UI mode. |
For all other parameters, see Common Parameters.
See Common Parameters.
For all other parameters, see Common Parameters.
The Agent Comparison report consists of a bar graph comparing Agents' performance against each other as well as against the average value of all Agents represented.
For all other parameters, see Common Parameters.
See Common Parameters.
See Common Parameters.
The Agent Trend report is a line graph that displays agent performance on one or more selected categories over time.
For all other parameters, see Common Parameters.
See Common Parameters.
See Common Parameters.
The Capacity Trend report contains information that can be used by the system administrator to monitor the system load and see the extent to which the system can handle new incoming calls. It has two parts:
| This template is only available in an Analytics UI mode. |
See Common Parameters.
The Category Bubble Chart report compares Categories by plotting them on a graph according to the percentage of calls in the horizontal (x-axis) and vertical (y-axis) Categories. Each Bubble Category is represented on the graph by a bubble. The size of each bubble represents the average call duration. The x-axis value indicates the percentage of calls that were in both the Bubble Category and any of the selected x-axis Categories. The y-axis value indicates the percentage of calls that were in both the Bubble Category and any of the selected y-axis Categories.
The X-Axis Only and Y-Axis Only versions present the data as a bar graph indicating percentage of calls per Category, rather than a bubble chart.
| This template is only available in an Analytics UI mode. |
For all other parameters, see Common Parameters.
See Common Parameters.
| For all of these parameters, only Categories that were not excluded from the data set by means of the Categories parameter of the Data Set Filters section are included in the report. Thus, if you selected specific Categories in the Data Set Filters section, only Categories that are selected both there and here will be included in the report. |
The Category Distribution report is a graph or chart that shows the distribution of calls by Category for a specified domain. The distribution can be calculated by percentage of calls or by percentage of resources (total duration of all calls included in the analysis).
| This template is only available in an Analytics UI mode. |
Full Report, Mini, Wide: Bar graphs listing the selected Categories and showing their proportion in the data set. For additional information about these versions, see Common Parameters.
For all other parameters, see Common Parameters.
See Common Parameters.
For all other parameters, see Common Parameters.
The Category Trend report shows changes in Categories over a period of time. The trend line can represent number of calls, percentage of calls, average call duration, or total resources. There is one line per selected Category.
| This template is only available in an Analytics UI mode. |
For all other parameters, seeCommon Parameters.
See Common Parameters.
See Common Parameters.
The Coaching Summary report summarizes the current status of coaching sessions in a pie chart and table, and lists active coaching sessions by user. Items in the list can be expanded to display details about coaching sessions.
For all other settings, see Common Parameters.
The Metadata Frequency report is a histogram that shows, for a selected metadata field, how many calls in the data set had each of the values of the field. The report has two levels. The top level is a summary that shows how many calls they belonged to each value and the percentage of all the calls in the data set that had the values. For example, if the metadata field "department" is selected, the report shows how many departments (values) were represented in the data set, and divides the departments into groups based on how many calls they had. The lower level gives details for one of the groups in the top level, listing the values that had at least the number of calls represented by the group, and indicating how many calls they had. The Wide version of the report also shows the most common Category of those calls.
In the sample reports below, the top-level report contains a bar graph that indicates there were three departments ("3" at the top of the first bar) for which there were three calls (3x at the bottom of the bar), one department for which there were four calls, and one department for which there were five calls. Above this, the report shows that 14.8% of the calls in the entire data set were in the first group, 21.3% of the calls were in either the first or second group, and 29.5% of the calls were in the first, second, or third group. The Mini version of the report does not include the percentage data. The lower-level report was opened by drilling down on the first bar of the graph in the top-level report. It shows the breakdown of calls for all the values.
For all other parameters, see Common Parameters.
See Common Parameters.
For all other parameters, see Common Parameters.
In the top level:
In the lower-level:
The Performance Profile report is a bar graph and table that compares Workgroup or Agent performance in a specified Category or group of Categories to the benchmark (the Workgroup’s or Agent’s average). You can choose to analyze performance based on the number of calls, the percentage of calls, average call duration, or total resources. The table includes data about the number of calls, the difference from the benchmark, and the percentage above or below the benchmark.
| This template is only available in an Analytics UI mode. |
For all other parameters, see Common Parameters.
See Common Parameters.
The QM Agent Summary report summarizes the responses on forms that were filled out by quality monitoring staff after they reviewed calls.
For all other settings, see Common Parameters.
See Common Parameters.
The System Load report includes a line graph and a grid displaying the amount of audio hours processed per day or hour. On the line graph, two lines are displayed, one for processed calls and one for loaded calls. The grid below the graph shows each point (either a date or time) that appears in the line graph, with its corresponding loaded and processed hours.
| A smaller version of this template, MINI_System Load, is also available. It is intended for use in the
Views page. This template is only available in an Analytics UI mode. |
The Team Comparison report is a bar graph that compares performance between teams and against the average value of all teams represented.
For all other parameters, see Common Parameters.
See Common Parameters.
The Top and Bottom Performers report is a bar graph that indicates values for three groups (top performers, bottom performers, and average performers) across multiple Categories. This enables you to identify what top performers do differently from the rest and thereby spread the knowledge of best practices. This report is only useful if you know ahead of time who your top, bottom, and average Agents or teams are. It is purely a graphical representation of comparison, and will NOT tell you who the top and bottom Agents are, but rather how they compare in specific areas of performance.
| This template is only available in an Analytics UI mode. |
For all other parameters, see Common Parameters.
See Common Parameters.
The Topic Analysis - Audits report is a table that shows the precision with which each Topic was identified by SpeechMiner. Precision values are given for one or more confidence levels, as specified in the report parameters, and are determined by the TP/FP/SFP (True Positive/False Positive/Sense False Positive) selections made by the call auditor (see Using an Event Grid).
Each line in the table contains auditing and precision information for one Topic, including the level of strictness required for recognizing the Topic (as defined in the Topic definition in SMART), how many times the Topic was found in the calls that were included in the report, how many calls it was found in, and how many of the events found were audited. In addition, the precision is graded for each threshold included in the report. The grades are calculated by dividing the number of TP events by the sum of the TP and FP events. The grades are presented as follows:
| Grade | Value of TP/(TP+FP) |
|---|---|
| A | A >= 0.8 |
| B | 0.8 > B >= 0.7 |
| C | 0.7 > C >= 0.6 |
| D | 0.6 > D >= 0.5 |
| F | F < 0.5 |
This report is instrumental to implementers, as it allows them to fine-tune the system to the correct accuracy level. After learning about Topic-identification performance, the implementer can adjust confidence levels as necessary.
| This template is only available in an Analytics UI mode. |
See Common Parameters.
| Confidence levels are assigned by the system to each term it identifies. They indicate how closely the term defined in the Topic matches the term that was discovered in the audio segment. "Precision" in the report means the percentage of TPs for all the audits in the Topic whose confidence levels were equal to or above the given Thresholds. This means that each of the Thresholds includes all of the audits in the Threshold levels above it. Thus, in the example below, the 40-Threshold column includes all audits with precision values of 40 or higher, the 50-Threshold column includes all audits with confidence levels of 50 or higher, and so on. |
The Topic Analysis - Trends report is a table that compares how often particular Topics were found in the selected time period with how often they were found in the preceding time period. Each line in the table contains information for one Topic, including the percent change between the current period and the preceding period, and the number and percent of calls in each period in which the Topic was found.
The baseline time period for the report is selected in the Data Set parameters (see Common Parameters). The preceding time period is automatically set by the system based on the baseline. For example, if the baseline is a two-month period, the preceding time period used is the two months preceding that baseline period.
| This template is only available in an Analytics UI mode. |
See Common Parameters.
The Topic Correlation report gives information about the co-occurrence of Topics. The report generates a list of association patterns by mining the set of Topic events for frequent and significant patterns of association. Each pattern of association is composed of a cause ("Left Hand Side") and a consequence ("Right Hand Side"); that is, each pattern identifies a particular Topic or sequence of Topics that seem to lead to the appearance of another specific Topic. For example, the report might include a pattern indicating that when the Topics Transfer and Dispute are found in a call, they are often followed by the Topic Dissatisfaction.
For each correlation in the report, statistical information is provided indicating how significant the correlation was in the data set:
| This template is only available in an Analytics UI mode. |
Identifying scenarios in the flows of calls that tend to produce certain results Investigating the root causes of events that take place during interactions Customizing the Report
The Usage Tracking report shows information about the activities of users over a selected date range.
| In a Recording UI mode the Applies column is not relevant and will not be included in the report. |
The following parameters are the most common ones you will encounter. Most of them have the same function no matter which template you use. Exceptions will be briefly explained in this section, and dealt with in detail in the documentation on individual reports in the Templates section.
You can choose or switch to any of the report templates available.
Text entered in this field will be used as a file name for the report if you choose to save the report. This name will appear under the Saved Reports tab for future access.
Text entered in this field will be displayed in large, bold type at the top of the report results. The report title is optional. The name of the report template appears in this field by default, but can be modified or deleted.
Text entered in this field will be displayed in fine print directly below the Report Title in the report results. The description is optional.
The output of some reports can be generated in different versions - different sizes or formats In this field, select the desired output version from the dropdown list. The available options vary depending on the template. The following options may be available:
The Data Set Filters allow you to narrow down the data from the entire database in order to base your report on a subset of the available data.
All of the Data Set Filters have an "AND" relationship, which means each filter further restricts the options available, and all call events and topics represented in your final data set satisfy all Data Set Filters. Should you choose options from two different filters that mutually exclude each other (e.g., a Program designated as "Customer Service France" with calls only in French and the Language option "Dutch"), you will have zero data in your report.
Note that the parameters that appear under Items on Report (for example, the third row of parameters) allow you to select specific information from the base of data that is selected under Data Set Filters. Distributional information presented in the report is calculated for the items selected under Items on Report, built upon the base of data that is selected using the Data Set Filters.
The Data Set Filters appear in the Filter Panel on the left side of the report template for all reports that support them. The filters are divided into groups. The basic filters appear at the top of the Filter Panel and are always displayed when the panel is displayed. The other groups of filters can be expanded or collapsed. The following groups of filters are available:
The parameters in this section allow you to select specific information and calculate distributional information on the data in the data set that was specified using the Data Set Filters. If no selections are made under Items on Report, default values will determine how the data you have narrowed down through the Data Set Filters appear in the report.
Since the parameters in Items on Report differ from report to report, only the most common ones are explained in this section, listed alphabetically. Less common parameters will be explained in their respective templates.
You can select the specific Agents or Workgroups to be analyzed and displayed in your report. The default value ("Any") includes all agents. Your selections here must match up with your selections in the Data Set Filters (e.g., the same Workgroups or Agents, belonging to the same Languages and Programs), if the latter are not all set to default ("Any"). In other words, the report output will only include Agents who were selected here and who were not excluded from the data set in the Data Set Filters.
You can select specific Categories of calls to limit the data set analyzed for your report. Your data set will include calls belonging to any of the Categories selected. Use the default value ("Any") to include all available Categories. Note that if a single call belongs to multiple Categories, it will be counted as one call for each Category selected under this parameter. Also, bear in mind that the report output will only include calls in Categories that were selected here and were not excluded from the data set in the Data Set Filters. Thus, if you selected certain Categories in the Data Set Filters section, only Categories that are selected both there and here will appear.
When used with Percentage of Calls in the Data Type parameter, the report will display the ratio of {the number of calls from the selected Categories in Items on Report} to {the number of calls from the Categories selected for analysis in Data Set Filters}, as a percentage. Note that if one or more calls belong to multiple Categories, the sum of the percentages of calls for all Categories may add up to more than 100%.
You can specify the type of data you wish to display in your report. The most common choices are:
This parameter is most commonly seen on reports that show trends. The output of the report will be displayed in increments of the selected Period Type. For each increment, the value of the selected Data Type will be displayed for each particular report item. For example, if the Data Type is "Number of Calls" and the Period Type is "Day," the number of calls for each day in the date range included in the report will be shown.
In this parameter, you can choose how many of the highest or lowest values are displayed on the report. The options may include 5, 10, 15, 20, All, All Sorted, and All Unsorted.