contact centr e top kpis for your · top kpis for your contact centre first contact resolution...

Top KPIs for yourContact Centre

By Rod Jones

Top KPIs for your Contact Centre

BOOST PRODUCTIVITY & EFFICIENCYWITH THE RIGHT CONTACT CENTER KPIs

By Rod Jones © 2019 Rod Jones Consulting (Pty) Ltd

Ver 2.0 at 4 June 2019

IntroductionAny call centre, contact centre or Cx-enabledcustomer interaction centre is faced with �vepowerful challenges in 2019.. and beyond. Theseare the �ve most important expectations that theorganisation’s most senior executives will demandof the operation.

■ Cut, Reduce or Cap Operational Costs

■ Increase Operational Productivity and E�ciencies

■ Increase Revenues (Including Sales, Retention, Up-Sell/Cross Sell, Collections etc.

■ Reduce Risks (Including �nancial as well as brandand reputational risks)

■ Increase Customer Satisfaction (Including allcontemporary Cx measures)

To maintain management focus on operationale�ciencies and productivity, it is vital that theorganisation puts into place, and closely monitorsand interprets appropriate Key PerformanceIndicators. Whilst there are literally hundreds of well-de�nedstandards and KPIs that have been written about fordecades, after almost 40 years in the industry I havereduced these to the Top Ten that I personallyconsider to be the most critical in the context ofdriving e�ciencies and productivity whilst takingcognisance of Customer Experience and commercialor �nancial issues.

Page 1Top KPIs

https://app.designrr.io/projectHtml/291967?token=8f479126c4498920b19cb291ea19e9cd&embed_fonts=

Top KPIs for your contact centre

First ContactResolution (FCR)For most well-run contact centres, the focus overthe past decade has moved from driving e�ciencymetrics to striving for positive interaction outcomes.

Resolving customer issues quickly and e�ciently isthe fundamental principle of any call centre orcontact centre. It is precisely what customersdemand and expect of our operations. The quickerthat agents process high quality customerinteractions with positive outcomes, the greater theoperational e�ciency and productivity. And if this isachieved in conjunction with CX-orientated QualityAssurance practices, there will be an additionalpositive knock-on e�ect in terms of CustomerSatisfaction as well as positive business outcomes.

Clearly state exactly how your speci�c operationde�nes ‘First Contact Resolution’. No two contactcentres are alike, and no two interaction types areexactly the same. Just as di�erent interaction typeswill have di�erent SLAs, it is also vital to clearlyde�ne FCR in this context. Ensure that yourde�nitions contain both inclusive and exclusivecriteria. For example: Does your operation de�ne‘Resolved’ as being when the main issue has beendealt with (FCR from the operation’s perspective)and also, only if the customer speci�cally expressesfull and complete satisfaction with the entireinteraction?

How do we know that we have achieved FCR?Putting it simply, we need to ask the customer. Well-trained agents (supported by e�ective QA, coaching,mentoring and on-going training) will conclude calls(or any other interaction type) with a short set ofstructured probing questions. Ideally, these arescripted or system-driven and will form part ofmandatory call control and post-call wrap-updisciplines. Analysis of this data will provide clearindicators of actual FCR; this, subject to correlationwith Repeat Contact Analytics.

The second dimension of FCR is the challenge ofidentifying Repeat Contacts. Repeat contactsrelating to previously reported issues meansigni�cantly increased customer e�ort on the onehand, and this has a consequential negative impacton loyalty and the overall customer experience.

In many cases, identi�ed repeat contacts willhighlight �aws in operational processes or evenunderlying organisational policies and othersystemic drivers.

Repeat contacts also consume extremely costlyoperational resources and are therefore asigni�cant drain on e�ciency and productivity. Usethe contact centre technologies and reportingcapabilities to analyse data to identify repeat callsusing CLI numbers (Calling Line Identi�cation).Typically, call or interactions going back seven to tendays might well be related to a single issue. But notalways. Correlate this data with the CRM data andcraft policies and processes to address identi�edissues.

Page 2Top KPIs



Resolve the most commonreasons for low FCRLow FCR (and that means anything under about 70%FCR) is a clear indicator that there is room forconsiderable improvement in the contact centre,and urgent attention is needed to remedy issuesthat fall into four primary categories. Fix theseissues and your true FCR will rise signi�cantly; as willcustomer satisfaction

Empowerment – Ascertain what percentage of non-FCR compliant calls could agents have e�ectivelyhandled but were not empowered by theorganisation to do so. For example: Policies,processes, procedures or the absence of speci�cmandates or empowerment may preclude agentsfrom achieving FCR for certain interaction types.

Competency – Identify how many calls orinteractions are either not resolved or escalatedbecause all or some speci�c agents have notreceived the appropriate training, and are thereforenot competent to assist customers with certaintypes of interactions.

Processes – Identify which interaction types andwhich processes are responsible for non-FCRcompliance. Make adjustments (where appropriate)to policies, procedures and processes.

Systems – Review current technologies in terms ofagent support, knowledge management, dataharvesting, analytics and reporting. Implementappropriate changes.

A word of caution: Great care must be taken whende�ning precisely what FCR means in any speci�coperation. FCR isn’t a “One-Size-Fits-All” KPI. Investconsiderable time and e�ort to craft the exact,measurable de�nitions speci�cally for youroperation.

Best Practice GuidelinesMost internationally recognised benchmarkingreports and consultants’ views recommend thatFirst Contact Resolution should be between 70%and 75%. However, percentages will most likelychange depending on how FCR is de�ned in thecontext of SLAs across di�erent channels, bydi�erent interaction types or call reasons and by themethods used to measure FCR.

To ensure that management is provided withmeaningful reports and actionable insights, it isrecommended that the operation de�nes individualFCR targets for the top ten to �fteen interactiontypes and contact reasons, across all availableinteraction channels.

The Bene�ts of First ContactResolutionBy focussing attention and making rationalinvestments into achieving ever-improving FCRstatistics, the operation will bene�t signi�cantly.Here are �ve ways that the organisation will bene�t.

Reduced Operational Costs. Achieving high (75%+)FCR will have a dramatic impact on the reduction ofrepeat calls. This will not only signi�cantly improveoverall operational e�ciency, e�ectiveness and costcontainment but it will also go a long way to reduceagent stress levels when interaction volumes spike.A 15% increase in FCR can result in between 50%and 60% reduction in repeat calls.

Improved Customer Satisfaction. FCR is one of thekey factors in the overall Customer Experience (Cx).Improved FCR has a direct and positive impact onthe operation’s Customer Satisfaction Index (CSat)

Page 3Top KPIs



Increased Revenues. When customer issues areresolved quickly and e�ciently on the �rst contact,it immediately opens up the opportunity for agentsto up-sell or cross-sell products or services.

Increased Customer Loyalty. When customer issuesor concerns are resolved quickly on the �rst contact,customers feel valued and appreciated. High FCRmeans signi�cantly increased long-term customerloyalty and consequentially, increased CustomerLifetime Value (CLV)

Improved Employee Engagement. By implementingprocess, technology and systemic changes to driveimproved FCR, overall agent engagement and jobsatisfaction will improve signi�cantly. By speci�callycategorising certain types of FCR that are within theagents’ ability to control, well-structured incentivesand bonuses can drive signi�cantly improvedperformance, longer tenure, reduced attrition andreduced absenteeism.

Quality AssuranceDe�nitionsQuality: A degree of excellence as measured againstsimilar things

Assurance: Con�dence or Certainty

It means being con�dent or certain that your callcenter is achieving excellence.

Balancing ActThere is no shortage of information supporting theprinciple that implementing and maintaining robustquality assurance policies, procedures andprocesses is vital to the well-being of any contactcentre or Cx operation.

When crafting such protocols, it is vital that the twoprimary dimensions of QA are kept in balance: Theissues that are important and matter most toCustomers. The issues that matter most to thebusiness; these will include compliance andgovernance.

The QA FrameworkIt is a misconception that contact centre QA ismerely a matter of developing or evolving a multi-faceted ‘tick sheet’, perhaps applying a fewweightings and after that, it’s all plain sailing. By notrecognising the distinctions and the importance ofthree very di�erent strategic approaches to QA,managers are not only missing the point, but alsomissing out on the opportunity to use QA to feedthe organisation with extremely valuable businessintelligence; the very actionable insights that arenecessary to achieve de�ned business outcomes.

The three dimensions of QA are: Operational,Tactical and Strategic. Each of these, by their nature,will have uniquely di�erent sets of KPIs ormeasurables and therefore, very di�erentassessment criteria and weightings. This principletherefore requires that the operation crafts at leastthree distinctly di�erent call assessment checklistsand that assessments are carried out with the viewto arriving at speci�c business insights in each ofthese areas.

It is vital that the operation constantly monitorsagent performance against the establishedbaselines or benchmarks in the context of the threedimensions mentioned above. Thereafter, this datacan be used for analytics, coaching, mentoring andtraining on an on-going basis; a closed loop process.This to ensure that the contact centre is �rmly kepton the path of continuous improvement.

Page 4Top KPIs



Some Best PracticeGuidelinesOpinions vary greatly about how many agent callsor interactions should be assessed in any givenperiod, At the one end of the spectrum, there arereasonably successful operations assessing fewerthan one agent call or interaction per week, whilstothers are assessing as many as 10 to 12 calls peragent per week, or as high as 50 per month. The�gure of 8 to 10 calls per agent, per month appearsto be as close as we can possibly get to what may becalled ‘global best practice’.

It must be noted, however, that the widely acceptedpractice of ‘random call selection’; for the purposesof assessments, more than likely assesses mostly‘Average’ calls.

Statistically, out of the typical 8 calls per agent, permonth, only one is likely to be a ‘good’ or HighOpportunity call, and two calls will be ‘poor’ or High-Risk calls. The other six calls will be spread fairlyevenly as ‘Average Calls’ that provide little or no realbusiness value.

Contact centre operators are advised to implementtechnology-enabled interaction identi�cation andselection solutions utilising data analytics, ‘wordspotting’ and other automated QA techniques.

Outsourcing the QAFunctionGiven that contact centre QA has become one of thecore drivers for improving Customer Experience,Business Performance and contact centreOperational E�ciencies, there is a fast-growingtrend for operations to outsource all or a portion oftheir QA to specialist third party service providers.

Historically, typical contact centre QA has consistedof basic checklists comprising of many Do’s andDon’ts, with signi�cant ambiguity and mysteryaround the scoring and reporting, andcomparatively little value being added to theorganisation at a strategic level. At worst, QA isviewed in the contact centre as a form of ‘discipline’– ‘Big Brother is Watching’.

In today’s world of customer-centric organisationsand Cx-enabled contact centres, there are manybene�ts to outsourcing QA. These include:

Cost E�ective. In most cases, it is signi�cantly morecost-e�ective to outsource the QA function to aspecialist third party provider. Measurable costsavings are signi�cant. In many instancesoutsourced QA will be as low as 40% of internallymanaged QA.

Specialised QA Tools. Most reputable outsourcequality experts have highly advanced, specialisedQA and on-line assessment and coaching tools thatallow managers or supervisors to see results and�ndings immediately and in real time.

Error Detection. Good QA service providers alsoassess operational front-line data entry and use ofCRM systems, and if desired can correct or ‘�ag’areas identi�ed as “risk/incorrect”

Page 5Top KPIs



Drive Business Improvement. Specialists QA serviceproviders are able to drive business improvementsby assessing First Call Resolution problems, newopportunities, potential risk of sales ‘stickability’ andpropensity to pay (PtP) for collections environments.Experienced interpretation of these type of �ndingsprovide valuable business insights and businessintelligence to the organisation.

Near Real-Time. Certain professional QA serviceproviders can deliver almost ‘real time’ assessments,reports and coaching or training recommendations.

Impartiality. Independent, objective assessmentsand feedback is essential when linking QA �ndingswith incentives/KPI’s.

Multilingual. Specialist QA service providers canassess calls in accordance with customerdemographics and in line with TCF (TreatingCustomers Fairly principles)

Best Practice Knowledge & Guidance. Experts in QAservices will guide your organisation to implement‘Best Practice’ both locally as well as internationally.This can include consulting services as well asspecialised training for managers, team leaders andother key operational support sta�.

On-Line Coaching. Specialist QA services canprovide the ability for agents to log into the qualityand online coaching tool to identify �rst hand theirresults and personal development opportunities.

Flexibility. The operation can scale up or down QArequirements as required for strategic, tactical oroperational purposes

Reduce Headcount. Reduction in �xed HRoverheads frees up operational capacity (Contactcentre seats) for increased customer service or salesproduction.

Business Intelligence. Assessment analytics andexperienced interpretation and expertise generatesreliable and credible insights for decision-making.

Compliance Monitoring. Specialist resources tomonitor and report on governance, regulatory andcompliance issues.

Bene�ts of ProfessionalQuality AssuranceIt will be appreciated that implementing andsustaining professional quality assurance processesand protocols in any customer-centric contactcentre, is not only vital to the overall performanceand success of the operation, but it should also bethe source of vital Business Intelligence andBusiness Insights necessary to drive certainboardroom strategies and tactics.

The following are merely a few of the many bene�tsthat the operation will derive.

Cut Costs and Increase Pro�ts. When designed tofocus on all key areas of the contact centreoperations (not merely assessing call recordings),QA analysis and feedback will highlight numerousopportunities for cost reduction and revenueimprovements.

Driving improved FCR, for example, will signi�cantlyreduce repeat call or interactions, a majorcontributor to contact centre infrastructure andoperational costs. Similarly, increased agentsatisfaction and engagement will drive longertenure, reduced attrition and absenteeism as well asreduced recruitment and training costs.

Page 6Top KPIs



Achieve Consistent Service Delivery. Today’scustomer is becoming fully immersed in the day-to-day utilisation of omnichannel interactions withproviders. The total customer experience, across alltouchpoints – the Customer Journey – has becomethe focus of most ‘world class’ organisations.

As Dan Kiely of Voxpro so eloquently puts it, “Thecustomer journey is only as delightful as its leastdelightful experience.”

A robust QA strategy, framework andmethodologies will highlight �aws and weaknessesin the service delivery chain and provide theorganisation with the necessary insights to remedythese quickly and e�ectively.

Measure Progress. Scorecards and checklistsappropriate to the Operational, Tactical andStrategic goals and standards of the operation willallow management to e�ectively monitor andmanage the continuous improvement of processes,individuals, teams and other de�ned andmeasurable business outcomes.

Service Level & QueueTimesTraditionally, the term ‘Service Level’ refers to thepercentage of calls or interaction that are answeredwithin a speci�ed time. For example: 90/20 or 90%of calls answered within 20 seconds. A ‘generic’service level benchmark such as this is a goodindicator of actual operational capacity in terms oftechnology, infrastructure and manpower

Managers (particularly those managing modernmulti-channel or omnichannel operations) shouldcarefully re-de�ne ‘Service Level’ contextual to eachof the di�erent types of interactions in their centre,and also in terms of high tra�c and o�-peak times.

When de�ning Service Level, cognizance should alsobe taken of actual customer expectations in thecontext of the type of operation. For example, anemergency services contact centre will have ansigni�cantly higher service level requirement – say95/5. (95% of all calls answered within 5 seconds)whereas in other circumstances, certain types ofcustomers, such as those calling technical supportcentres, will be happy with 90/30.

Secondary Service Levels. A problem with a typical90/20 Service Level is that it doesn’t consider the20% of callers that aren’t answered within thede�ned threshold. A good way to ensure that callersaren’t subjected to inordinately long wait times is foreach Primary Service Level (PSL), set a SecondaryService Level (SSL). For example: Set 80/20 as thePSL and 99/300 as the SSL. This means that only 1%of calls will be waiting longer than 5 minutes.

Establishing, monitoring and managing the mostappropriate KPI for Service Levels will be vital fordriving customer satisfaction, operationale�ciencies whilst addressing cost management.

Queue Times and Impact on Customer Satisfaction.Most analysts agree that whilst it is important foroperations to maintain their individually de�nedservice levels, when service levels drop to below60%, customer satisfaction starts decreasingsigni�cantly even when FCR is achieved.

Best PracticeGuidelinesThe following are regarded to be ‘SLA Best PracticeGuidelines’ of contact centres aspiring to beregarded as ‘World Class’

Telephone – 90% of calls answered by a ‘Live’ agentwithin 20 seconds

Page 7Top KPIs



Email– 95% of emails answered within four hours(the better contact centres are aiming to respond to80% of emails within 15 minutes)

Live Chat – 80% of chats answered within 30seconds

Social Media – 80% of contacts answered within 20minutes. (Some centres are attempting to servicesocial media interactions as close as possible toreal-time)

SMS/Messaging Apps – 80% of messages respondedto within 40 seconds

Letters – 95% of letters answered within three days

Abandon Rate (AR)By closely monitoring and analysing AbandonedRate, contact centre managers will be able to gleanvaluable insights relating to e�ciencies,e�ectiveness and speci�cally, into many of thedrivers of Customer Satisfaction.

De�nitions & StandardsLost Calls. There are three basic situations whereinbound calls can be ‘Lost’ and each of these needscareful consideration.

Abandoned Calls. This is discussed in considerabledetail in the text below.

Missed Calls. In some circumstances it is either theACD’s default setting or it has been adjusted, toautomatically disconnect callers once a speci�c on-hold threshold or wait time has been reached.Astute managers will investigate this setting andmake appropriate adjustments.

Dropped Calls. Dropped calls are when calls areaccidently disconnected due to technical errors. Thismetric should be closely monitored and if droppedcalls rise above about 0.5% there will be cause forserious concern and remedial action.

The de�nition of Abandon Rate has caused a greatdeal of confusion in the industry for many years.The big question is: “At what point does one startmeasuring Abandon Rate?”. Some operators recordand report on the total number of calls thatabandon before being answered by a live agent.Others, disregard all calls of under – for example – 5or 10 seconds. It is also important to check what theACD default is for reporting AR. Di�erent vendorsoften have di�erent de�nitions.

Most consultants will agree that an averageabandon rate of between 3% and 4% is acceptablein the majority of inbound contact centres. That is, itis subject to excluding particularly short calls ofunder 10 seconds.

Best Practice Guidelinein typical inbound sales environments, an AR ofanything over 1.5% to 2% is cause for seriousconcern. In most inbound customer service contactcentres, an AR of 4% or under is regarded asacceptable.

The Impact of Abandoned CallsThe value of closely monitoring and analysing ARcannot be underestimated. It is a especially clearindicator of real customer sentiments and there is aclose correlation with critical KPIs and factors suchas Customer Satisfaction, Customer Loyalty andCustomer Lifetime Value.

Page 8Top KPIs



The �rst insight that AR gives us is an indication ofcustomer tolerances. In certain environments wherecustomers have a multitude of options (such asordering fast-food or certain inbound salesoperations), tolerance levels will be particularly low.But in cases of calls to technical support deskswhere customers may have little option but to stayon hold, customers may well be considerably moretolerant, and as such, AR will be lower.

The table below clearly shows an example wherebyas high as 60% of inbound calls were abandoned inunder 60 seconds and over 80% when the hold timereached 4 minutes. It is a startling reality that it is ahigh probability that most calls abandoned in thisscenario probably resulted in a lost sale andtherefor lost revenues and lost pro�ts.

Now compare Abandoned Rate of typical inboundTechnical Support contact centre.

What is clearly notable in this scenario is thatAbandon Rate only reaches 50% afterapproximately 2 minutes and that these types ofcustomers are considerably more tolerant.

Abandon Rate andCustomer SatisfactionWithout going into too much detail, there isconsiderable well-researched evidence to show thatas a rule-of-thumb, as long as the calling customer’sissue is resolved quickly and e�ciently (i.e. HighFCR), even slightly lengthy hold times of up to 45seconds, can result in relatively high CSat (CustomerSatisfaction) scores. Conversely, when the initialhold time (or wait time) is longer than 90 seconds,CSat diminishes greatly, even if the caller’s issue isresolved.

Managing Abandoned CallsTo be in a position to implement meaningfulinterventions it is vital that representative historicaldata relating to AR is analysed to establish normsand patterns. This data will provide valuable insightsand a baseline for measuring the impact of anyoperational improvements.

Increase Capacity. Most obvious intervention toimprove Service Level and decrease AR is byincreasing contact centre capacity by adding moreagents. Perhaps by forcing down Average Handlingtime it may also have the e�ect of increasingcapacity. Whilst these may be ‘Quick Fix’ solutions,they will probably be expensive options and mayhave very little positive measurable e�ect onCustomer Satisfaction. In fact, forcing down AHTmay well have a serious negative impact on CSat.

Page 9Top KPIs



Redesign the IVR. A thorough investigation into thestructure and content of the IVR system may wellprovide a number of valuable insights and solutionsto improve both AR and CSat and the overallcustomer experience.

IVR First Message. If the analysis of AR data clearlyshows that a high percentage of abandoned callstake place within seconds of calls being answeredby the IVR, perhaps the content and the style of theinitial IVR messaging needs attention.

IVR Menu Structure. Long, complex IVR menus and‘unenthusiastic’ voice recordings are a guaranteedway to annoy callers and to drive up AR.

Use ‘Intelligent’ IVR Messaging. Most modern ACD’swill allow for the use of intelligent alternativemessaging. Instead of merely repeating the typicalphrase “Your call is important to us, please hold theline” every 20- or 30-seconds ad nauseum, try‘escalating’ alternative messages. For example, thesecond message (after, say, 45 seconds) may be “Weapologise for the extended delay. An agent will takeyour call shortly”. It is also well worth experimentingwith di�erent time spacing between on-holdmessages. In some environments 20 or 30 secondintervals may be appropriate whilst other types ofcallers would only expect an on-hold message every45 seconds.

Implement Automated Call-Backs. Many of the moresophisticated contact centre platforms provide thefacility to o�er callers a call-back option, either as aprimary IVR option, or after a certain ‘on hold time’threshold has been reached. Using CLI (Calling LineIdenti�cation) some platforms will be able toidentify and di�erentiate calls from mobile phonecalls landlines. It is only practical to o�er call-backsto mobile phone callers. Landline callers can beo�ered a typical message mailbox option.

The Impact of Music on hold. The inappropriateselection of the music that callers are compelled tolisten to whilst on hold can have a signi�cant impacton abandon rates as well as on Customersatisfaction. Customer demographics will also be amajor in�uencer. Best practice will be to carry outregular customer surveys, not only probing forattitudes to on-hold music, but to the overall IVRexperience.

Average Speed toAnswer (ASA)Monitoring and managing ASA in conjunction withAbandon Rate is often a critical part of certaincontact centre’s SLAs, particularly those operating inthe BPO environment. It is also important torecognise that like many KPIs, ASA should bede�ned speci�cally in terms of individual contactcentres’ operating realities. Similar to Service Leveland Abandon Rate, ASA should have a considerabledegree of design �exibility, dependent on the typeof services provided and the type of customersbeing serviced.

Virtual Queues with Call-Back Some platforms canalso hold call-back customers in a ‘virtual’ inboundqueue without losing their position after hangingup. Other features include sending text messages tocustomers con�rming the number that they wouldprefer to be called back on, as well as allowing themto select the actual preferred time for the returncall.

Increase the Ring Time. Although it may only have asmall impact on improving both Abandon Rate andService Level, it may be worth trying to increase theactual Ring Time before the inbound call is pickedup by the IVR. When one considers that the averagemobile phone connection rings for approximately18 seconds before it passed through to voicemail, itis a clear indication that customers are used to andcomfortable with a slightly longer up-front ring time.

Page 10Top KPIs



For example, customers calling technical supportservices will be content to wait quite a lot longer tobe served than those calling typical consumer linesor an emergency services number.

One should also take care when crafting ASAde�nitions and reporting models to speci�callyexclude abandoned calls from calculations.

Just as it is critical to monitor and report on ServiceLevel on an hour-by-hour basis (if not even morefrequently) ‘averaged ASA’ can hide critical failuresin service delivery. In a typical operational situationwhere -for example - over the lunchtime period –poor scheduling coupled with poor scheduleadherence by agents and a spike in inbound callvolumes, could force callers into a two- or three-minute queue time. Although this would certainlyraise the average across the day by only a fewpoints, it will probably still be within the de�nedSLA. Yet the negative impact on CSat, NPS and othercustomer metrics during the contact volume ‘spike’,could be devastating.

Best Practice GuidelinesThe most widely accepted standard for AverageSpeed to Answer is the same as that for ServiceLevel, namely 20 seconds. (90% of calls answered bya ‘live’ agent in 20 seconds). This can be used as apractical baseline. But bear in mind that the type ofoperation (for example, Emergency Services,Inbound Sales etc. will have an in�uence on howyou establish this KPI)

Cost per Interaction(CPI)It is the case in many organisations that the callcentre or contact centre is regarded by many in thehigher levels of executive management (particularlyamongst the ‘Old School’) to be somewhat of areluctantly-funded cost centre; a ‘necessary evil’.Whilst this situation is starting to change for thebetter, we contact centre professionals have toshoulder some of the blame for this situation.

We have tended to shroud much of our contactcentre activities in what the ‘Old School’ mayconsider to be a Mumbo-Jumbo of hard-to-comprehend acronyms and piles of statisticalreporting that only makes business sense to a fewof the more enlightened.

As an industry we may well have been able to getaway with it in the past. We applied for and weremostly provided with merely adequate budgets tokeep the operation going, year-after-year. It’s now‘Game Over’. The new game in town in the “Battlefor Budgets in the Boardroom” is quite simply,Return on Investment. Unless contact centremanagement can provide irrefutable evidence andcredible arguments, we are not going to be able towin the support for �nancial and other resourcesvital to drive customer service excellence. We haveto start talking the language of the boardroom… The Business Case and the organisation’s strategicintent!

Page 11Top KPIs



Sadly, far too few contact call centres or centresactually measure and report on operational costs ata granular level. This is a shocking reality;particularly when one recognises that all �ve of thewidely acknowledged ‘Key Challenges’ (see theopening statement of this article) or demands bythe executive team on the operation refer to�nancial issues.

It is my position that particularly as the industryrapidly enters the omnichannel era, and as aconsequence, will require signi�cant investments inincreasingly sophisticated (and expensive)technologies and enhanced human skills, it hasbecome critical that every interaction type (and subcategory) is speci�cally measured in terms of ‘all-in’costs and reported on in terms of cost-per-second,per minute or per completed process.

Not only will this granular-level reporting highlightexcessive costs, wastage and opportunities for costreduction, these reporting models will allow contactcentre management to present rational and highlycredible ROI arguments when lobbying theexecutive team for �nancial support for newtechnologies, increased sta�ng, additional training,incentives and bonusses or any other interventionsrequiring funding.

Best Practice GuidelinesOther than the more sophisticated Business ProcessOutsource (BPO) providers and matured telesalesoperators whose very existence hinges onmaintaining accurate costing models, it has been myexperience that most service-orientated contactcentres only have a vague idea of overheads anddetailed operating costs.

Best practice recommendation is for contact centremanagers to engage on this subject with colleaguesin the likes of the Finance and HR and any otherappropriate business units, and to jointly develop agranular level costing model. This is an extremelychallenging task. Many of the individual line-itemcosts will be ‘buried’ in the corporate �nancialmanagement structures and will need the tenacityof a CSI detective to unearth the real �gures. Acomprehensive contact centre costing model willcontain as many as 300 or 400 individual cost line-items.

Success lies in knowing, understanding and beingable to correlate true operational costs with realisticcustomer value data, and thereafter to producecredible, �nancially-based reports or motivations forresourcing.

It is also vital that the true cost and the measurablevalue of individual customer interactions arereported along with credible projections of value-based customer satisfaction statistics, customerloyalty and revenue or pro�t projections.

Average HandlingTime (AHT)It is widely recognised that AHT is probably the keymetric (from an operational or productionperspective) for all contact centre capacity planning.In a nut-shell, AHT tells us precisely the amount oftime an agent is working on or committed to anidenti�able task or work item. Based on this data,we have the ability to accurately plan for capacity tomeet or surpass customer service expectations.

The formula for calculating AHT is as follows:

Page 12Top KPIs



Whilst Total Hold Time will have a signi�cant impacton AHT, care must be taken to ensure that this ismeasured and analysed separately (See below)

AHT is also one of the most commonly used metricsfor overall contact centre operational performanceas well as for agent or team performance. Thefundamental problem with using AHT as aperformance metric is that in many cases it forcesagents to focus on speed of service rather thanquality of service. Also, AHT metrics tell us nothingabout call quality or call outcomes. There is agrowing trend for management to closely monitorAHT but agents are measured on it in terms of CSat.

Watch out for AgentShenanigansAstute contact centre managers will be aware of andtake precautions and actions to mitigate sometypical agent ‘tricks’ that can seriously skew AHTstats.

The following are examples

■ Deliberately cutting o� long or complex calls asthe call duration approaches the AHT target

■ Unnecessarily transferring or escalating long,di�cult or complex calls

■ Rushing callers to complete the call

■ Not truly listening to caller’s problem, challengesor information requests

■ Not probing for additional customer issues afterthe primary issue has been successfully addressed

■ Making promises to callers that cannot be kept

■ Blaming or �nding loopholes in IT

■ Avoiding or not completing after-call wrap-upprocesses

■ Not following up on undertakings to customers

CautionsExtreme care should be taken to ensure that drivingagents too hard to maintain prescribed AHT ‘norms’can drive down the overall Customer Experience,Customer Satisfaction and Net Promotor Scores.

Average Talk Time(ATT)It is critical for contact centre managers to recognisethe di�erences between Average Talk Time andAverage Handling time. Although ATT and AHT bothmeasure the time an agent is ‘committed to aninteraction’ ATT speci�cally excludes any hold time.

As an example: A contact centre may wellconsistently meet its de�ned Average HandlingTime. On paper, it looks good. But what if ATTanalytics were to show that a signi�cant portion ofthe handling time was made up of agents habituallyputting customers on hold? The overall CustomerExperience (CSat, NPS and other metrics) wouldcertainly su�er. This will also have a negative impacton Service Level and the accuracy of scheduling.

A deep-dive and a closer look at real ATT analyticscan reveal serious �aws in processes, recruitment,training, QA, coaching, technologies and supportsystems.

Page 13Top KPIs



Occupancy, Utilisationand Internal ShrinkageOccupancy, in its most succinct form, is calculatedas a percentage representing the actual amount oftime that agents spend on call-related (orinteraction-related) activities while they are logged-on and expected to be handling interactions.‘Related Activities’ include – for example – all talktime, hold time and wrap-up time. (Or similaractivities for non-voice interactions) Measuring,monitoring and managing Occupancy will providemanagers with a powerful E�ciency Index. In otherwords, are agents fully utilised or could they behandling additional tasks or work items?

Bear in mind that Occupancy targets should be setin terms of organisational goals for continuousimprovement and take cognizance of �nancial,sta�ng practices and Cx metrics and goals.

A good guideline for establishing OptimumOccupancy will be between 80% and 85%. This isprobably ‘the sweet spot’ in terms of productivity,e�ciency and rational manpower utilisation.Managers are cautioned against pushing foroccupancy rates higher than this as it will almostcertainly lead to agent burn-out, reduced Cx andhigher absenteeism and attrition. The knock-on ofthese factors includes diminished CSat, reducedcustomer loyalty, declining revenues and escalatingcosts associated with recruitment, training andassimilating new sta�.

Utilisation, on the other hand, di�ers fromOccupancy in that it represents the amount of timethat agents spend on call- (or interaction) relatedactivities whilst they are present at work in thecontact centre. Utilisation calculations thereforeinclude all the time spent by agents on internalactivities – not just their logged-in time.

Monitoring and measuring Utilisation is a criticalelement of overall contact centre management as itconsiders all non-customer-related activities thatcomprise a large portion of overall centre costs andoverheads.

Best Practice GuidelinesOccupancy. A good guideline for Occupancy will bebetween 80% and 85%. Managers are cautionedagainst pushing for occupancy rates higher than thisas it will almost certainly lead to agent burn-out,reduced CSat and higher absenteeism and attrition.

ShrinkageMany industry stalwarts may disagree on thedetailed composite elements or wording when itcomes to de�ning ‘Shrinkage’, but most will agreethat it is fundamentally anything that takes an agentaway from their ability to handle customerinteractions.

Whilst Shrinkage is an essential factor for overallplanning and a critical input to medium and large-scale Workforce Planning tools and disciplines, it isnot necessarily a KPI that is measured and trackedon a day-to-day basis. Shrinkage is normallyexpressed across a complete 12-month period butmonitoring and reporting shrinkage would certainlyprovide management with valuable insights.

Page 14Top KPIs



External Shrinkage■ Sick leave / Absenteeism

■ Family Responsibility and Compassionate Leave

■ Planned / Scheduled Leave

■ Public/Statutory Holidays

■ Paid Breaks (Lunch and Tea Breaks)

■ Unscheduled Absenteeism

■ Lateness / Tardiness

Internal Shrinkage■ Coaching and training sessions

■ Team meetings

It is my personal opinion that a contact centrerunning consistently at a 30% (or lower) shrinkagefactor is either super-e�cient and in the hands of awell-seasoned and experienced contact centremanagement team, or alternatively, it is borderingon being a sweat-shop. In this latter case there willbe consequential negative impact from highattrition, low sta� engagement, low CSat, NPS andother customer metrics.

How to Apply ShrinkageIn its most basic de�nition: Shrinkage is thereduction of the number of sta� (from a theoreticallevel) of full-time equivalent agents (FTEs) able to befully productive 100% of the time throughout theyear.

As mentioned in the context of ‘Utilisation’ (above),there are two clear components to overallShrinkage. These are, Internal and ExternalShrinkage.

■ Contact centre social activities

■ One-to-one reviews/chats

■ Toilet, smoke and other unplanned facility breaks

■Time spent helping other departments or businessunits

■ Special and creative projects

■ System downtime

Best Practice: ShrinkageIt will be recognised that unless all Internal andExternal shrinkage factors are de�ned exactly thesame way for any two or more contact centres, theoverall Shrinkage Factor will di�er widely. However,it is generally accepted by many specialist contactcentre consultants that a global ‘Industry Standard’for shrinkage is 30% to 35%.

That would mean that were we to use an ErlangCalculator to work out sta�ng requirements for anygiven period, and the answer was (say) 70 agents,and our known Shrinkage was (say) 30%, we wouldactually need more like 100 agents to maintainService Level.

Page 15Top KPIs



Schedule AdherenceSchedule Adherence is the percentage of the agent’sscheduled working time when they are actuallypresent in their designated working environment.This will include all of the above shrinkage factors.

The Impact of PoorSchedule AdherenceThe table below clearly illustrates the negativeimpact on Service Level (and consequential drop inCSat) when Schedule Adherence drops by even oneor two agents. Conversely, over-sta�ng will certainlyhave a high cost impact with insigni�cant gains inactual customer satisfaction.

A serious negative impact of poor ScheduleAdherence is the knock-on impact that this has onOccupancy. Not shown in the above table, with 32agents, Occupancy will be in the region of 85%.When available agents are reduced to 30,Occupancy will raise dramatically to anunsustainable 95%+. This puts immense pressureon agents and symptoms of ‘burnout’ and highstress levels will start manifesting within hours.

Whilst the actual impact on operations should oneor two agents be merely a few minutes late to log-inmay be slight, when consistent and widespreadtardiness becomes the norm, the consequencesbecome extremely serious.

Best Practice GuidelinesMost reputable benchmarking reports andexperienced contact centre consultants agree thatSchedule Adherence of between 85-95% is generallyconsidered acceptable.

Customer Satisfaction (Including Net Promotor Score (NPS) andCustomer E�ort Score (CES)

Customer Satisfaction – or CSat, is probably themost widely used and recognised measure ofcustomer sentiment. The advantage of there beingno single or universally recognised approach to CSatmeasurement tools or methodologies, is that thebasic concept allows organisations to craft their ownbespoke models to meet their speci�c needs.

Another concept relating to CSat is that, in fact,overall Customer Satisfaction should be monitored,measured, analysed and acted on using acombination of di�erent data-gathering methodsand approaches. These could include, basicCustomer Satisfaction probes as well as NetPromotor Score (NPS) and Customer E�ort Score.

In its most basic form, surveys including a questionsuch as ..."On a scale of one to ten, how satis�ed were youwith our service today?" The result is a set of data that can easily berepresented as a percentage.

Page 16Top KPIs



Happy vs Unhappy

One of the easiest and quickest ways that surveyorscan use to gauge customer sentiment is the used ofHappy – Neutral – Unhappy icons, typically usedwith a ‘tick-box’ technique. This approach can yieldsigni�cantly higher response rates without the needfor complex analytics. It also allows surveyors to doquali�ed follow-up probes with speci�callyidenti�ed customers.

Star Rating Technique

Becoming increasingly more popular and used bymajor organisations such as Amazon and Net�ix,the Star Rating System is familiar to most customersand is easy to implement and to maintain as anintegral part of the organisation’s marketing.

The ubiquitous Amazon Customer Review model isa perfect example of how this technique can beapplied to virtually any product, service ororganisation.

The Net PromoterScoreOriginally developed in 2003 and introduced intothe marketing fraternity by Fred Reicheld of Bain &Company and Satmetrix, Net Promoter or NetPromotor Score is a widely used management toolused to gauge the loyalty of an organisation’scustomers.

The Net Promotor Score (NPS) is calculated basedon responses to a single question:

“On a scale of 0 to 10, how likely is it that you wouldrecommend our company (or product or service) toa friend or colleague?”Those who respond with a score of 9 or 10 arecalled Promotors.

These are �ercely loyal and supportive customers;the organisation’s Advocates.

Those who respond with a score of 0 to 6 arelabelled Detractors. These are potentially damagingto the organisation.

Responses of 7 or 8 are referred to as Passives .

It is generally accepted that although heavilydependent on industry sector, an NPS score of 50 ormore is either ‘really good’ or ‘excellent. It indicatesthat the organisation is highly customer centric. At ascore of 70+ the organisation will be in the companyof global brand leaders such as Apple, Amazon andWalt Disney. Achieving consistently high NPS ratingsindicates that the organisation is geared forsigni�cant organic growth powered largely bycustomer advocacy.

Page 17Top KPIs



Best Practice Guidelines forNPSIn isolation, an NPS score tells us very little.However, when monitored and measured over aperiod of six to twelve months (or even longer),positive or negative shifts in NPS trends will bestrong indicators. For example: A 10% positive shiftin NPS is a powerful con�rmation that theorganisation is heading in the right direction;becoming increasingly more customer centric andtherefore driving organic growth and consequentialpro�tability.

It is also recommended that to drive theimprovement of NPS, customer verbatims are usedto identify trends in expectations and sentiments.Also implement targeted coaching for agents in thebottom percentage of NPS performers

Customer E�ort Score- CESThere is certainly su�cient evidence to support thepremise that high e�ort on the part of customers toresolve issues with their providers, leads tosigni�cantly diminished Customer Satisfaction anddiminished Loyalty. High e�ort is one of the mostpowerful drivers of customer defection.

In the light of massive developments inomnichannel customer self-service, reduced orminimised customer e�ort is rapidly becoming a keycompetitive di�erentiator or competitive advantage.Modern society has become e�ort averse.

In early 2010 when measuring customer e�ort(particularly in the context of call centres) was �rstmooted, a scale of 1 to 5 was used as themeasurable response to the probe question “Howmuch e�ort did you personally have to put forth tohandle your request?” Responses were initiallymeasured as follows:- 1 = Very low e�ort 2 = Low e�ort 3 = Neutral 4 = High e�ort 5 = Very high e�ort

The evolution of CustomerE�ort ScoreIn many customer contact centres throughout theworld, CES methodologies have evolved slightly withthe emphasis on the words, ‘ease’ or ‘easy’.

The scale for this method would be measured sothat:

“How easy was it for you to get your issue resolvedfully, on a scale of 1-7?”1 = Extremely easy 2 = Very easy 3 = Fairly easy 4 = Neither 5 = Fairly di�cult 6 = Very di�cult 7 = Extremely di�cult

Page 18Top KPIs



Best Practice Guidelines for CESAs with other tactics and techniques for measuringoverall customer satisfaction, the end goal in mostcases, is to achieve optimised customer experience;this resulting in increased market share, sustainablerevenues and pro�t. CES methods should beadapted to meet speci�c operational requirementsand thereafter, appropriately integrated into theoverall organisation as a means of drivingcontinuous improvement.

Last words…

Fully understanding the nuances of KPIs and how tocustomise and implement them in your particularoperation, is the hallmark of an astute contactcentre professional. Essentially, there are four areasthat need constant focus and attention:

■ Customer Satisfaction (How well are we servingour customers?)

■ Operational E�ciency (How well is our operationperforming?)

■ Business Value (How well are we supportingbusiness sales?)

People Management (Are we getting the best out ofour people and are they happy?)

KPIs and Targets orGoalsIt is well worth remembering British economistCharles Goodhart’s ‘Law’…

“When a measure becomes a target, it ceases to bea good measure.” The fact of the matter is that as soon as we make aKPI a speci�c target, that is exactly what people willfocus their attention on and will strive to achieve.Once that target is reached, the motivation toexceed, diminishes considerably. This paradoximmediately distorts what is actually happening inthe operation to create that target in the �rst place.

Conversely, the operation does obviously need toachieve certain goals, objectives and target. Thesolution lies in the reality that the business or theoperation needs to craft speci�c new KPIs that di�erfrom de�ned targets. As an example: If theoperation needs to speci�cally improve overallcustomer experience, then NPS or CSat could bothbe metrics in play, but not individually or inisolation.

Page 19Top KPIs



AcknowledgementsWhen researching for this article I have reliedheavily on what I consider to be one of the �nestresources for information pertaining to CustomerExperience, Call Centres and Contact centres, theUK-based CallCentreHelper.com website founded in2003 by global industry legend, Jonty Pearce.

My thanks go out to Jonty and his team for havingcreated and for maintaining this incredibly valuableresource, and to the many writers, bloggers andcontributors whose knowledge, experience andexpertise is shared freely with us avid readers,researchers and writers around the world.

I urge all contact centre professionals to subscribeto the CallCentreHelper newsletter and to visit theon-line magazine whenever you need information,knowledge, insights or inspiration. Seewww.callcentrehgelper.com

My thanks also go to my long-standing friend andprofessional colleague, Roz Broom for reviewing theinitial draft of this ebook and for her hugely valuablecomments, suggestions and inputs. Also, to my sonEvan Jones for many valuable inputs and comments.

Without the critical eye and editing skills of my dearfriend Miemsie Visser, this document would havebeen a mess of typo’s, spelling, context andgrammatical errors. Thank you for your beady-eyeand your constructive comments.

Rod Jones

Lonehill, Johannesburg. South Africa January 2019

Page 20Top KPIs

http://www.callcentrehgelper.com/


contact centr e top kpis for your · top kpis for your contact centre first contact resolution...

Documents