Call Center Metrics: Key Performance Indicators (KPIs) - ICMI

1 Call Center Metrics: Key PerformanceIndicators (KPIs)Edited by Brad Cleveland and Debbie HarnePost Office Box 6177 Annapolis, Maryland 21401800-672-6177 410-267-0700 Fax: TutorialsTU00002 Call Center Metrics: Key Performance Indicators (KPIs)Edited by Brad Cleveland and Debbie HarneContentsIdentifying Key Performance Indicators .. 1 Call (Contact) Quality .. 9 First-Call Resolution/Errors and Rework.. 12 Definitions and Use of Service Level, .. 14 Response Time and Quality .. 14 Average Speed of Answer .. 18 Abandoned and Blocked Calls.. 20 Forecasted Call Load vs. Actual .. 23 Scheduled Staff vs. Actual .. 24 Adherence to Schedule .. 25 Average Handling Time (AHT).. 27 Occupancy and Productive/Nonproductive .. 29 Cost per Call (Cost per Contact) .. 32 Average Call Value.. 33 Revenue .. 34 Budget/Cost Objectives.. 35 Objectives for Outbound .. 36 Customer satisfaction .. 39 Employee satisfaction .

2 41 Turnover .. 42 Overall Call Center ROI .. 44 KPIs as Interrelated Outcomes .. 45 How KPIs Relate to Customer Expectations .. 47 Reference Bibliography .. 49 About This ICMI Tutorial.. 50 About the Editors .. 51 About Incoming Calls Management Institute .. to ICMI, Inc., 20031 ICMI TutorialCall Center Metrics: Key Performance Indicators (KPIs)Identifying Key Performance Indicators Key Points Key Performance Indicators (KPIs) are high-level measures of callcenter Performance . KPIs generally include: Call (Contact) Quality First-Call Resolution/Errors and Rework Service Level and Response Time Average Speed of Answer Abandoned and Blocked Calls Forecasted Call Load vs. Actual Scheduled Staff vs. Actual Adherence to Schedule Average Handling Time (AHT) Occupancy and Productive/Nonproductive Cost per Call (Cost per Contact) Average Call Value Revenue Budget/Cost Objectives Objectives for Outbound Customer satisfaction Employee satisfaction Turnover Overall Call Center ROIE xplanationA significant amount of information is required to effectively manage a callcenter.

3 You need, for example, data on caller needs and expectations, the queueand caller tolerance, the load on the system, agent satisfaction and Performance ,call patterns, cost components, the activities of other parts of the organization,and conditions in the external environment. But you also must be able toclimb above the detail and assess overall Performance without having to reviewdozens of reports. Acquiring the data required to track some key Performance Indicators isstraightforward and comes directly from system reports; other KPIs requiremany more data sources and much more assembly. Given the variety to ICMI, Inc., to ICMI, Inc., to ICMI, Inc., 20032 ICMI TutorialCall Center Metrics: Key Performance Indicators (KPIs)systems, reports and statistics available, the emerging solution is to useCTI/middleware to pull disparate sources of information into question, then, is, What measures adequately summarize the numerousactivities of a call Center ?

4 While any measure by itself has the potential tomislead, the following key Performance Indicators typically provide a goodpicture of the call Center s Performance when they are interpreted following table summarizes: Common call Center Performance objectives The formula or approach for measuring each objective Sources of data and tracking methodologies A summary of suggested applications and relevant roleObjectives Related to Quality:Call (Contact) QualityApproach: Assigns a value to the quality ofindividual typically comes from samples viamonitoring and/or recording contacts;however, some criteria may also be generat-ed from ACD-based call coding or reportsfrom customer information systems. Eithersupervisors or quality assurance specialistsreview monitored/recorded contacts for ini-tial data : Appropriate in all environ-ments as both a high-level objective (overallsummary of the results of individual con-tacts) which is generally tracked monthly,and as the basis for specific objectives foragents and supervisors, contact by : The quality of each contact is essen-tial to successful call Center should be defined to reflect theneeds and objectives of both the organiza-tion and customers.

5 Criteria generallyinclude such things as interpreting cus-tomer requirements correctly, entering dataaccurately, providing the correct informa-tion, accurate call coding, capturing neededand useful information, etc. These criteriashould be an inherent part of monitoringand coaching : Appropriate in all environ-ments as a high-level objective, which isgenerally tracked monthly (however, withthe right systems, can be viewed at anytime). Components that lead to first-callresolution should also be built into specificFirst-Call ResolutionFormula: Calls resolved upon initial con-tact total resolution is generally trackedthrough a database system (customer to ICMI, Inc., 20033 ICMI TutorialCall Center Metrics: Key Performance Indicators (KPIs)mation system) or by ACD call first-call resolution is defined drivestracking , does it refer to a call notbeing transferred or to an issue that isresolved even if it must be escalated ortransferred?

6 Ideally, first-call resolutionshould be defined as an issue resolved onfirst contact (the caller doesn t have to con-tact the Center again or vice versa), even ifescalated or transferred during the contact;transferred/escalated calls can be tracked assupporting objectives for agents (however,because not all aspects are within their con-trol, these components must be selectedcarefully). Notes: Studies indicate that organizationsincur many types of additional expenses(some hidden and difficult to track) whencallers issues are not fully resolved with thefirst contact. There is significant value inanalyzing relative increases and decreases infirst-call resolution, in response to changesin call Center processes, systems and cus-tomer and ReworkApproach: The percent (and types) oferrors and rework that are generally comes from a database sys-tem (customer information system) and/orby ACD call coding.

7 Application: Appropriate in all environ-ments as high-level objective, reportedmonthly or (with the right systems) asoften as the manager chooses. Specificcomponents of errors and rework are oftenbuilt into quality objectives for agents(however, because not all errors are withintheir control, variables must be selectedcarefully). Notes: Errors lead to rework, unreliabledata and potential interpretation problemsdownstream. As with first-call resolution,there is significant value in analyzingincreases and decreases in errors andrework, in response to changes in processes,systems and other Related to Accessibility:Application: Service level and responsetime are key accessibility measures and areappropriate high-level objectives. Theyshould be reported by reporting intervals ( , the number of half-hours within,above and below objectives) NOT as aver-ages across days, weeks or should be able to identify recur-ring problematic intervals.

8 These objec-tives are also key planning targets, used forbase staff : Establishing concrete service leveland response time objectives is a prerequi-site to the solid planning necessary toensure that the organization is accessiblethrough whatever channel customers Level and Response Time Formula, service level: X percent of con-tacts answered in Y , response time: 100% of contactshandled within N days/ level is available directly from ACDreports. Response time reports may comefrom additional systems, , emailresponse management systems (ERMS),Web servers, workforce management sys-tems (WFMS), to ICMI, Inc., 20034 ICMI TutorialCall Center Metrics: Key Performance Indicators (KPIs)Application: ASA comes from the same setof data as service level. It is not necessaryto have both service level and ASA objec-tives. If service level data is not available,ASA can be a substitute. ASA does haveimportant operational applications; , itis a component of trunk : ASA is often misinterpreted as a typical experience, but the average isskewed by many callers who get answeredbefore ASA and some who wait far longerthan ASA.

9 Application: Abandoned and blocked callsare caused by insufficient staffing or trunk-ing resources. They should be supportinginformation to service level and responsetime reports, not primary : Abandonment rate, though often aprimary objective, is not a concrete meas-ure of call Center Performance , because it isdriven by caller behavior which the centercannot directly control; it should be of sec-ondary importance to service level. Busysignals may be due to insufficient trunks,but are often the result of inadequatestaffing and the resulting queues of Speed of Answer (ASA)Formula: Total delay total number directly from ACD reports andthe and Blocked CallsFormula, abandoned calls: Calls abandoned calls , busy signals: Number or percentof attempts that received busy calls are available directly fromACD on busy signals may come fromthe ACD (if using ACD controlled busies),the local telephone company and the inter-exchange (long distance) company (IXC).

10 Objectives Related to Efficiency:Application: Appropriate in all environ-ments as a high level objective, reported byinterval; it is also used for ongoing tacticaladjustments. Notes: Forecasting the workload is a high-leverage activity that is fundamental tomanaging a call Center demand will mask anddefeat all other efforts to provide goodservice. And overestimating demandresults in waste. As a high level objective,forecasting accuracy should NOT bereported as a summary of forecasted versusactual calls across a day, week or month,but an illustration of accuracy for eachreporting interval (typically, half-hours). Forecasted Call Load vs. ActualApproach: The percent variance betweenthe call load forecasted and the call loadactually call load is available from thesystem used for forecasting, , workforcemanagement system or call load is tracked by the WFMS,ACD, ERMS, fax servers, Web servers, etc.