Monday, May 4, 2015

Data Collection for contact center reporting

Reporting in the contact center begins with collecting the data from various sources.  All of us use these applications every day and therefore the data should be available to combine into another reporting application. Unfortunately that is not always accurate.  Data is often locked down or access is very limited by the data provider.

Another challenge with data collection is the method that is used to capture the data.  Data is stored in a certain fashion and collecting that data is going to be limited.  The reporting software that you use must be able to collect the data in the fastest and most secure method possible.  Some data can be captured using older technology which might be cheaper but it opens the door to problems.

Spectrum’s software, NeXorce, collects data using different methods.  Briefly these methods are:
  1. Delimited File Collectors
  2. ODBC Collectors
  3. Screen Scrape (Data Emulation) Collectors
  4. Socket Collectors
  5. Web Page Collectors
  6. XML File Collectors
Rather than get extremely technical the purpose of this overview is to go over the advantages and disadvantages of each collection method.    This article is meant for the call center manager and not for the IT department.  For that reason the explanations are kept to a basic level and do not go into details or complete explanations.  For further details on each of these collectors it is recommended that you search through

Delimited File Collectors
A delimited file stores numbers and text in plain text.  Plain text means that the file is a sequence of characters that is readable with a standard text editor.  Each line of the file is a data record and each record consists of one or more fields separated by a delimiter such as a comma, tab, colon or a pipe.
  • The files are easy to read so collecting the data and caching or storing the data shall not take software efforts.
  • The data can be parsed quickly.  The structure of the file allows for easy access of the specific data.  Collecting specific data from a delimited file can be very quick.
  • Access to the file requires local or a network share.  This can be a security issue for some companies.
  • Lockfile can occur. When the application is writing to the file and the collector is reading the file a collision can occur. These collisions or conflicts can result is a <NULL> value or a blank KPI.  
  • Some Excel Files are saved as delimited files and may contain formulas, date/time values, blank rows and columns or color coded text.  These Excel features cannot be read by most collectors. 

ODBC Collectors
Open Database Connectivity (ODBC) is a programming language API that is used for accessing database management systems.  An ODBC Driver is used as a transition layer between the application and the database management system.  Collectors are used to capture the data as defined by the specific query (data request).  Although not accurate, many beginners are told to think of a database as an Excel file with each tab being a table within a database.
  • Fast method of collecting the data from the database.  
  • Efficient ability to collect a single piece of data or all of the data requested.
  • Multiplatform collectors “joins” can capture data from multiple tables.
  • At times can be slow if the volume of data is very large.  
  • Large tables may require multiple collectors to improve on refresh rates.
  • Some ODBC’s require a driver that is not automatically included with the application.  

Screen Scrape (Data Emulation) Collectors
Screen Scrape (Data Emulation) is a phrase that describes the operation of obtaining data or text from a simple text file.  A program will monitor the text file and have a MAP defined that tells the application where each piece of data or text resides.  This application will then scrape this file and obtain whatever information is in each location (coordinates) that are defined in the “map”.

  • Easy to initially set up – usually very quick to setup and run when a canned report of file is used.  
  • Low impact on the system – Screen scrapes can be run on minimum PC requirements.
  • Cost – low cost third party applications can be used to capture the data.
  • Lockfile – the PBX/ACD refreshes its text file report; the screen scraping software (data emulation software) needs to scrape the data from the text file report and write it to its own file.  These two refreshes can happen at the same time and when it happens there can be a collision or a conflict.  The conflict will result in the screen scrape software not collecting some of the data or returning data as <null> values.  
  • Windows permissions are needed to write the text file to the local machine or creating the file somewhere on a network machine.
  • Screen scraping requires an actual screen report to be captured from a window.  This must be done from within a user launched application.  In a server environment this is not ideal as the application performing the capture will need to remain open and running under a user account with active session.  Security concerns and if the session is ever closed the reporting application will also close and the data will no longer be updated.
  • If the report has to change the screen scraping will no longer be accurate.  Something as simple as adding a new column or removing a row will immediately affect the accuracy of the data being collected and reported.
If you are using data emulation or screen scraping or do not know how the data is being collected your reports may show the following results: Missing data on the report.  The report may have a blank column or row or just a few of the KPI’s are blank.  This does not happen at the same time or day.  Screen scraping is very antiquated and because of security issues should no longer be used.

Instead of a numeric value you have <NULL>.  This is the same result of the screen scraping software expecting there to be a value but there is nothing there so the result is <NULL>.  Finally there may be times when the value is “0” when as a manager you know that value cannot be accurate.  What some data emulating software will do is if the value is a <NULL> it will exchange it for a “0”.

Socket Collectors
A socket is used for communications between programs.  However the data source application requires that the incoming application socket conforms if it does not the connection cannot be made between the two applications.
  • Fast method of collecting the data from the source. The data source could be a delimited file or database table.  
  • Efficient ability to collect a single piece of data or all of the data requested.
  • The collector may need to be restarted if the Port is closed.  Antivirus software may close the port.
  • Application compatibility.  

XML File Collectors
Extensible Markup Language “XML” defines a set of rules for encoding files that are readable.  The goal of XML is simplicity and easy to use across the internet.  XML format is used by many API’s such as RSS, MS Office, and Apple iWork.
  • The files are easy to read so collecting the data and caching or storing the data shall not take software efforts.
  • The data can be parsed quickly.  The structure of the file allows for easy access of the specific data.  Collecting specific data from an XML file can be very quick if you know the design structure of the file.
  • The file structure has its advantages and disadvantages. Knowing the file structure is a requirement and can also be a speed bump in the data collection process.  
  • A large file can be very slow to collect data from because of the data collection methods required.
There are other disadvantages to XML files which make collecting data from the file challenging.  If the amount of data is rather small then XML files can work. If the data is a large amount then staying away from an XML file is the best path to take.

Web page Collectors
HTML is the standard markup language that is used to create web pages.  Web pages are used for countless reasons from company information to traffic and weather.  The web page data is collected and shared with others that did not have the time to find that website.
  • You do not have to create the data it is already done for you.
  • A wide variety of “data” is available.
  • You are not responsible for maintaining the “data”
  • The data format can be changed or removed without any notification to your data collector.
  • The data can be locked down and future access denied.
  • The data access is no longer free.
Collecting data for Call Center Reporting can be challenging for the applications that are being used and for the manager who has to decide on the data to be collected.  One of the issues that are rarely thought of is the collection methods.  This seems like an IT issue but not selecting the right collection method can result in problems with the data or the access to the data. Security, reliable data, accurate data, easy access and fast replies are all issues that the call center manager must be made aware of before deciding on the data collection methods.

Spectrum is a leading provider of Unified Contact Center Reporting.  Contact Spectrum today to discuss data collection methods.  For more examples of reporting visit our website and the products page.

Follow Spectrum Corporation:
  • SlideShare
  • LinkedIn
  • Blog
Dan Boehm
VP Sales and Marketing
+1 713 986 8839

No comments:

Post a Comment