Skills Based Routing - Call Center Service Provider · PDF filefamiliarity with specific...

0

© Specialty Answering Service. All rights reserved.

1


Contents

1 Introduction .......................................................................................................................................... 3

2 History and Evolution ............................................................................................................................ 4

2.1 What Is Skills‐Based Routing? ....................................................................................................... 4

2.2 Applications of Today's Skills‐Based Routing ................................................................................ 6

2.2.1 Overflow Routing .................................................................................................................. 6

2.2.2 Highly Targeted Routing ........................................................................................................ 6

2.3 The Future of Skills‐Based Routing ............................................................................................... 7

3 Common Mistakes in Skills‐Based Routing Models .............................................................................. 8

3.1 Mistake 1: Every skill set should have a dedicated pool of agents ............................................... 8

3.2 Mistake 2: If the forecasting window is shorter, the accuracy is better ...................................... 8

3.3 Mistake 3: A lower call handle time indicates better customer service ....................................... 9

3.4 Mistake 4: Agents can acquire new skills with training ................................................................ 9

3.5 Mistake 5: Skills‐based routing can be done away with if agent desktops have sufficient

background information ........................................................................................................................... 9

4 Practical Applications .......................................................................................................................... 11

4.1 Blended Call Centers ................................................................................................................... 11

4.2 Multiple Queues .......................................................................................................................... 11

4.3 Application Processing Servers ................................................................................................... 11

4.4 Virtual Call Centers ...................................................................................................................... 11

4.5 Skills‐Based Simulators ............................................................................................................... 11

5 Advantages of Skill Based Routing ...................................................................................................... 11

5.1 Increased First Call Resolution .................................................................................................... 12

5.2 Reduced Agent Work Time ......................................................................................................... 12

5.3 Decreased Customer‐Hold Time ................................................................................................. 12

5.4 Reduced Operational Costs ......................................................................................................... 12

5.5 Increased Call Center Revenues .................................................................................................. 12

5.6 Increased Customer Retention ................................................................................................... 13

6 Critical Success Factors ....................................................................................................................... 14

6.1 Being Customer‐Centric .............................................................................................................. 14

6.2 Ease of Use .................................................................................................................................. 14

2


6.3 Planned Implementation ............................................................................................................ 15

6.4 Workload Distribution................................................................................................................. 15

6.5 Successful Skills‐Based Routing ................................................................................................... 15

7 Conclusion ........................................................................................................................................... 18

8 References .......................................................................................................................................... 19

3


1 Introduction

Aninboundcallcenterusuallyhandlesdifferent typesofcalls intermsofcallerprofileaswellaspurposeofcall.Itisusuallynotcosteffectivetotraineveryagenttohandleeverytypeofcall.Asaresult,agentsareusuallyspecializedincertainprocessesorproducts,andtheyhandlecallsrelatedtotheseareas.Anotherskillrequirementisthatoflanguages,asmostcallcentersreceivecallsfrommultiple geographies. Most agents may be proficient in one or more languages, but not in alllanguagesinwhichcallsarereceived.InhighlytechnicalcallcenterssuchasthoseofITcompanies,agentsmayrequireextensivetrainingtobeabletohandlecallsefficiently.Withcallcentersbeingoutsourced, agents often handlemore than one client’s calls, and it is not practically possible totrainallagentswithall skillsnecessary tohandleeveryaccount.This ishowskills‐basedroutingbecameanecessityforcallcenters.

Today,automaticcalldistributorshavethecapabilitytoroutecallstoagentsbasedontheirskillsusingarulebasethatcanidentifytheskillsetrequirementsfordifferentcall typesandmapittoskilllevelsofagents.

While skill based routing techniques have evolved over the last couple of decades, there is stillplenty of room for improvement, and researchers aswell as call center practitioners are on thelookoutforbetterwaystodesignandmanageSBR.

4


2 HistoryandEvolution

The automatic call distributor (“ACD”), which is common in most call centers today, was firstintroduced in1972.Until then, incomingcallswerehandledonaone‐to‐onebasis,wherebyeachincomingcallwassenttoeachphonelineina‘hunting’patternuntilitwasanswered.Factorssuchasagentavailabilityorskilllevelswerenotfactoredinwhileroutingcalls.WiththeintroductionofACDs,itwaspossibletoconnectthecallsandqueuethemuntilanagentwasavailabletoanswer.Thisallowedproactiverulebasedqueueingofcallsbasedonthenatureofthecall,thelocationofthecalloriginandtheskilllevelsofagents.

Traditionally, ACDs routed calls based on the ‘longest idle time’ in order to have a uniformworkloaddistribution among agents.Theassumptionwas that all calls are for a similarpurposeandthatanyagentcanhandleanycalleffectively.Theobjectivewastominimizeagent idletime,andcallerwaittime.

However, as call centers evolved from a cost centric approach to a customer centric approach,personalizedservicebecameafocusarea.Inordertoprovideacustomizedservicetoeachcaller,callcentersstartedanalyzingthenatureofthecallsandattemptedtoprovidefirstcallresolution,byconnectingthemtoanagentwhohadtherightsetofskills, information,andempowermenttomakedecisions.

Someofthefactorsthatareconsideredwhileroutingcallsincludetheaccounttype(new/existing,highvalue/lowvalue), previouspurchasepatterns, language, andgeography,when the customerlast called and which agent serviced him, level of knowledge of the caller, and so on. Thisinformationisthenmappedtotheagent’sskillsandknowledgeinanattempttooffercustomizedservicetothecaller.

ACD manufacturers have been constantly improving their product offerings with severalenhancements and add on features such as overflow groups, routing calls to specific agents,allowingagentstobepartofmultiplequeues,etc.

2.1 WhatIsSkills‐BasedRouting?

Skill based routing is based on the premise that different agents will have different skill sets,abilitiesandtalentswhichwhenmatchedtotherightcallprofile,canproduceexcellentcustomerservicelevels.Whenskills‐basedroutingwasinitsinitialstages,anagentcouldbeassignedtoonlyasinglecallgroupprofileandallcallsinthatgroupwereroutedamongagentsinthatgroupalone.Akeyimprovementthatmadeskills‐basedroutingmorecommerciallyviableistheabilitytoassignmultipleskillgroups toagents.Thus,anagentcanbelong tomultiplegroupsandprioritywillbeassignedtoeachgroup.Basedonthecallvolume,anagentmayanswermultipletypesofcalls.

Good customer service is characterized by the ability of a call center to reduce thewait time ofcallersand toprovideagoodmatchbetween theagent skill setand thenatureof the call.Often,these goals are in conflict. Wait times can be reduced by having a single large pool of agentshandlingallcalls,althoughthismodelprovidesnooptiontomatchthecaller’sneedstotheagent’s

5


skills.From thepointofviewof the caller, thechoice isbetweenvery longwait timesandbeingshifted from one agent to the other in search of a solution, often with the added frustration ofhavingtorepeatinformationmultipletimesasthecallistransferred.

Intheinitialstages,SBRsworkedonlywhenacoupleofskillswereassignedtoeachagent,resultinginseriousdifficultiestoachievecustomizedserviceinthepracticalworld.Thus,onlyafewcriteriacouldbeconsideredfortheroutingdecision–forboththeagent’sskillsaswellasthecaller’sneeds.Asthenumberofskillsconsideredincrease,themodelbecomesexponentiallycomplextodesign,implement,maintain,andtweak.

Today skills‐based routingmodels have advanced to the extent of allowing agents to be part ofmultiplepools,enablingthebenefitsofasinglequeuetobeutilizedduringpeakhourswhilestillallowingagoodmatchbetweenagentandcallerwheneverpossible. Skills‐based routing isoftenachievedwiththehelpofIVRtechnologythatcollectsinformationaboutthecaller’sneedsandthenroutesittotheagentwhoismostqualifiedtorespondtothecall.

InagrouporientedSBRmodel,itisnotpossibletodynamicallymapeachcalltoanagent.However,latestversionsofSBRmodelsallowthecreationofvirtualgroupswhereskills,skilllevelsandotherdetails are stored at an individual level and not at a group profile level. Inputs from multiplesourcessuchas theANI,DNIS,CRM,and IVRareused todefine theskillprofileofeachcall.Theagentwhomost closelymatches the skill profile is assigned to the call, thereby achieving bettercustomization.Thistechnologyalsoofferssupervisorstheflexibilitytosetprioritiesforeachskillrequirementsothatskillrequirementscanberelaxedintheorderofleastprioritywhentryingtoexpandtheagentpoolavailable(especiallyduringpeakloads,wherewaittimethresholdsarelikelytobebreached).

Thisallowsmuchmoreskillstobemappedtoanagentsuchaslanguageandproductknowledge,familiaritywithspecific customeraccountsandevensoft skills suchas success ratepercall type(collection calls vs. sales calls for example), thereby increasing customer satisfaction levels andagentproductivity.

Forexample,SiemensRolmCommunicationsInc.'sResumeRoutingsoftwarecanmatchagentswithcaller on the basis ofmore than a couple of hundred skills including the agent’s knowledge andinterestlevelsineachproductorservicethatisbeinghandledbythecallcenter.Thesystemallowsdetailedagentprofilestobecreatedbasedonproficiencyandpreferenceratingsgivenagainsteachskilltype.Thesoftwaresupportsupto250skills,eachofwhichcanberatedonascaleof0to9forproficiencyandfrom0to9oninterest.

By using the information of the incoming call number and inputs from the IVR system, the PBXdeterminesthecalltypeandroutesittoamatchingagentprofilestoredintheserver.Thissystemthusoffersmuchmore flexibilitythanarudimentarysystemwhereagentsaregroupedintorigidskillgroupsbasedonbroadparameterssuchasproductlineknowledge,languageproficiency,andabilitytohandlespecificcustomertypes.

6


However,notallcallcentersmaybereadytousesuchsophisticatedtoolsandinvestinginahigh‐endsolutionhastobedoneaftercarefulplanningandbudgeting.

2.2 ApplicationsofToday'sSkills‐BasedRouting

WhilemostACDmanufacturersoffersomeformofskills‐basedrouting,thesolutionsvaryintermsofthefeaturessupportedaswellastheireaseofuse.

As thesupport for thenumberofskills increases, thesolution’s routingcapabilityalso improves,allowingformorepreciseroutingofcalls.MostSBRsolutionstodayhaveGUIsthatsupervisorscanusetocaptureskillsets,assignagentstoskillsetsandspecifytherulebaseforroutingcalls.MostapplicationsalsoprovideMISreportstomeasurehowtheroutingmodelisaffectingkeycallcentermetrics.

Workforcemanagementtoolsalsoofferthecapabilitytoassignskillstoagentsinordertoenableschedulingand forecasting.Thiscombinedwith theACD’sskillbasedroutingmodelensures thatcustomersatisfactionlevelsaswellascallcentermetrictargetsaremetbythecallcenter.

Skills‐based routing can be classified into two categories: overflow routing, and highly targetedrouting.

2.2.1 OverflowRouting

Overflow routing aims to achieve the trade‐off between waiting time and customer service byofferingthecalltoaninitialsetofagentswhoarethe‘best‐match’.Ifalltheagentsinthatgrouparebusy,thecallisqueuedagainstoneoftheagentsinthegroup.Ifthecallisnotansweredwithinathreshold wait time, it is then offered to subsequent sets of agents based on a pre‐determinedpriority.

Overflow routing is achieved by mapping a primary and a secondary skill set to each agent.Typically, every agentwill have a primary set of skills inwhich they excel, and then a bunch ofsecondary skills inwhich they are average or just above average. For example, if a consumer iscontacting a call center for a leisure travel booking, then the routing algorithmwill attempt toconnecthimtoanagenthandlingleisurebookings.However,oncethewaitthresholdiscrossedandnoagentinthatpoolisavailabletohandlethecall,hewillberoutedtooneoftheagentsinthenextpool. This could be comprised of agents primarily handling business travel but have some priorexperiencehandlingleisuretravelers.

Thecallcentersupervisorisgivencontroloverallvariablesintheroutingprocessandhecanfine‐tunethesystemtobestmeetthecallcentermetrictargets.

2.2.2 HighlyTargetedRouting

Highly targeted routing, aims toofferahigh levelof customizationby subdividing theagents formultipleskills,thusenablingthemtobeusedformultipletypesofcalls.Inthismodel,acallcentercansavecostsbynothavingadedicatedagentforaspecifiedtypeofcall,ifthecallvolumeforthat

7


typeof calldoesnot justify thecostof adedicatedagent. Instead,agentsare trainedonmultipleskillsinordertoofferpersonalizedservicetothecaller.Highlytargetedroutingalsoensuresthatagentsarenotover‐specializedtotheextentthattheyhavetoremainidleforlongperiodsoftime,whichwoulddriveupoperationalcostsandinefficiencies.

2.3 TheFutureofSkills‐BasedRouting

Inthefuture,twostrongtrendswoulddrivethedirectionofskills‐basedroutingmodels.Thefirstistheever‐increasingdemandfromcustomers forspecializedandpersonalizedservice.Thismeansthat in an attempt to offer superior quality, call centers would capture and analyze increasingamountsofcustomerdatasuchaspersonalprofile,preferences,pasttransactionalhistoryandsoon.Thesecondtrendistheevolutionofcallcenterstocontactcenters, therebyallowingmultiplechannelsforcustomerstoreachthecontactcenter.Thiswouldgeneratetherequirementforagentswith the right kind of skills to handle areas such as internet chat, video, email etc. apart fromtelephonecalls.SkillsetsaswellaswaittimeSLAswoulddifferforeachmodeofcommunicationand the challenge to quickly adopt to these modes of communication would necessitate morecomplexmodelsofcallqueueingandrouting.

8


3 CommonMistakesinSkills‐BasedRoutingModels

Whileskillbasedroutingmayseemlikeasimpleconcept,inreality,itcanoftenprovetobedifficultto generate the benefits of this wonderful idea. Often, the routing process does not aid the callcenter in achieving the expected results as call centers fail to define what a ‘successful’implementationis.Effectiveplanning,monitoring,andmeasurementofresultsandtweakingofthemodelareessentialinensuringthatthecallcenterderivestheexpectedbenefitsfromaskills‐basedroutingmodel.

A successful SBR model is one that can constantly evaluate and improve the ideal match ratebetweenthecallerandanagentwhocansuccessfullyhandlethecaller’sneeds.Thismodeldeliversthe results that the company expects, without overshooting anymetrics that the call center hasdefined.

Oftencall centersdonot identify thecoreobjectivesandgoalsresulting in ‘success’beingpoorlydefinedorwronglydefined.Forexample,somecallcentersmayhaveupselltargetsastheirnumberonepriority,andinsuchacase,tryingtominimizethecalltimewillbecounter‐productive.Ontheotherhand,ifacallcenterisprimarilyresolvingclientproblems,thenspeedofanswerandfirstcallresolutionetc.couldbegoodmetricstofocuson.ConfiguringtheSBRmodelshouldbebasedonthecallcenterobjectives.BeforeimplementingtheSBRmodel,thecallcenterneedstomeasurewheretheycurrentlyarewithrespecttotheirobjectivesandhowtheyplantoachievetheirtargetswithregardtotheobjectives identified.SBRmodelsshouldthenbeusedasafacilitatortoachievethestatedbusinessobjectives. In thissection,wewill lookatsomeof thecommonmistakes thatcallcentersmakewhileplanningandimplementingskills‐basedroutingstrategies,andhowtheycanbeovercometoachievebetterresults.

3.1 Mistake1:Everyskillsetshouldhaveadedicatedpoolofagents

Severalcallcenters,eventoday,believethathavingadedicatedpoolofagentsforeveryskillareaistheeasywayouttoprovideenhancedcustomerservice.However,thiswouldresultinasmallerpoolofagentsavailabletohandleagiventypeofcall,consequentlyincreasingthewaittimeforskillswherecallvolumeishigh.Onthecontrary,ifcallvolumedrops,agentidletimewouldgoupresultinginincreasedoperationalcosts.Therefore,itisimportanttooptforadynamicroutingmechanismthatcanevaluatethetrafficloadineachcustomergrouponanongoingbasissoastobalancetheworkloadacrosstheentirecallcenteroperations.

3.2 Mistake2:Iftheforecastingwindowisshorter,theaccuracyisbetter

Callcentersupervisorsoftenmakethemistakeofusingshorterforecastingwindowsforworkforceschedulinginanattempttoimproveforecastingaccuracy.Whileitisimportanttoevaluatetheperformanceoftheworkloadallocationandschedulesonanearrealtimebasisandmakeminortweaks,itiscounter‐productivetoforecastworkloadsbasedonshorttimeframes.Statisticaltheoryprovesthatwhenthesamplesize(thedatapointsusedfortheforecast)becomestoosmall,theforecastaccuracydropsdrastically.Assamplesizesreduce,itfailstoincludeavarietyofanomalies

9


thatappearonlylessfrequentlythanthesamplesizewindow.Forexample,itiseasiertoaccommodatebreaktimesintoagents’schedulesifthedayisbrokendownintohalfhourslots;however,ifaforecastisgeneratedeveryday,thenchancesarehighthattheactualcallvolumeandhandletimewoulddeviatesignificantlyfromtheforecastednumbers.Itisalwaysbettertousenearrealtimepredictivealgorithmsalongwithalongerwindowforecastoutput,topredictthepotentialnatureofcallarrivalpatternsbasedonthecallvolumesforecasted.Thiswillallowthecallcentertohandletherandomarrivalofcallsandavailabilityofagents.

3.3 Mistake3:Alowercallhandletimeindicatesbettercustomerservice

Although,alowercallhandletimeistypicallyanindicationoffastandefficientservice,itmaynotalwaysbethecase.Callcenterswithupsell/crosssell/revenuetargetsmayfindthatshortercalltimesareoftentheresultoffrustratedcustomersabandoningthecall.Thisisalsotrueofinboundcallcenters.Thus,thecallhandletimemetrichastobereadalongwiththenumbersforcallabandonment.Sometimes,awell‐matchedagentmaybeabletoconvincethecustomertospendmoremoney,althoughthecallitselfmaytakelonger.Therefore,everymetricshouldbetrackedinconjunctionwithothermetricsaswellastheendobjectives,inordertoobtaintherealpicture.

3.4 Mistake4:Agentscanacquirenewskillswithtraining

Whiletrainingcanbehelpfulinincreasingtheproficiencylevelofanagent,itmaynotalwaysbecosteffectivetoinvestintraininganagentinacompletelynewskill.Ajudiciouscallhastobetakenbythecallcenter’sHRandoperationsteamsinordertoarriveattheoptimummixbetweentrainingandnewhires.Recordingandreviewingcallsisagoodwaytoidentifyhiddensoftskillsofagentssuchastheirabilitytohandletoughcustomers,theirabilitytocrossselletc.,whichcanthenformabasisofroutingsuchcallstotheseagentsinordertoimprovethetoplineandthebottomlineofthecallcenter.

3.5 Mistake5:Skills‐basedroutingcanbedoneawaywithifagentdesktopshavesufficientbackgroundinformation

Today,highlysophisticatedCTI(computer‐telephonyintegration),andagentdesktoptoolscanensurethatallinformationpertainingtothecaller,suchasnatureofcall,priortransactions,currentstatusoftheaccountetc.canbemadeavailableontheagent’sdesktopassoonasthecallistransferredtotheagent.Thismakestheagentequippedwithalltheskillsnecessarytosuccessfullyhandlethecall–notnecessary.Informationavailabilityandagentskilllevelsaretwodifferentaspectsofofferingasuccessfulcustomerserviceexperience.Eveniftherightinformationismadeavailabletotheagent,itisoflittleuseifhedoesnothavetheexpertisetouseiteffectively.Thus,skills‐basedroutingandagentdesktoptoolsarecomplimentarytoeachother–oneisnotareplacementoftheother.

Some of the other issues faced by call centers during an SBR implementation are caused by thecreation ofmultiple virtual teams that are a fraction of the size of the entire call center. This isfurthercomplicatedbythefactthatthenumberandcomplexityofthevirtualteamsisdependent

10


onagentavailabilityandchangeseverytimeanagentlogsinoroutofthepool.Someoftheissuescausedinclude:

Efficiencyofcallhandlingmaydropwithmultipleskillsbeinga factor thataffectshandletimeandwaittime.Consequently,moreandmorecallsmaybequeuedforalongerperiodoftime.

Skills‐based routing can result in priorities being assigned to calls based on skill typerequirements.Thus,newercallsmaybehandledwhileearlier calls continue to remain inqueue.UnlessforcedoverridesaredefinedintheSBRrulebasebasedonathresholdwaittime,callscanenterintoaspiraloninfinitewaittimes.

Unavailability of forecasting and scheduling tools that effectively support the skills‐basedroutingmodelisanothermajorchallengefacedbycallcenters.

11


4 PracticalApplications

Someofthepracticalapplicationsofskills‐basedroutingarediscussedinthissection.

4.1 BlendedCallCenters

Blendedcallcentershandlebothinboundandoutboundcallsusingthesamegroupofagents.Inthiscase,inboundcallactivitiesareassignedasoneskillandoutboundcallhandlingisassignedasasecondskill,withagentsoftenhavingbothskills.AlongwithCTIandanadvancedACD,IVRisalsousedasaninputforroutingcallsinthisenvironment.

4.2 MultipleQueues

Multiplequeuesenableagentstobepartofmultiplequeuessimultaneously,thustakingawaytheissuesassociatedwithsmallervirtualteams.Agentsorsupervisorscontrolthequeuesfromwhichtheytakecallstoformandprioritizethequeues.Inaddition,theremayalsobemultiplesupervisorshandlingdifferentqueues,andagentscanselectwhichsupervisorwillmonitorthemduringdifferenttimesoftheday.

4.3 ApplicationProcessingServers

Skills based routing has given rise to the possibility of having two different servers working inparallel–oneforapplicationprocessingactivitiesandanotherforcallhandling.CTIlinksbetweenthetwoserversenablecallstoberoutedtoagentsbasedonavailabilityaswellasskillsets,ratherthanusingonlyoneoftheseastheroutingcriteria.

4.4 VirtualCallCenters

Skills‐based routing can be a good aid to implement a virtual call center – where agents frommultiplephysicalgeographieslogontoasinglevirtualACDenvironment.Hereroutingcallsbasedonagent’sskillsisanetworkrequirementandhencetheroutingisdoneoverthepublicswitchednetworkratherthanbytheACD.

4.5 Skills‐BasedSimulators

Skills‐basedroutingsimulatorscanbeusedtoanalyzewhat‐ifscenarioanalysisbeforegoingaheadwiththeactualimplementation.Itwillhelptoidentifywhethertheskillsidentifiedareaccurate.Aninaccurategroupingofskillsmaynecessitatehugeamountsofadditionalexpensesintermsoftheheadcount requirements to handle specific skills. Simulators help to identify such situationsupfront,givingcallcentermanagementtheopportunitytotryadifferentscenariothatworkswell.

5 AdvantagesofSkillBasedRouting

12


Apartfromtheincreasedcustomersatisfactionlevelsthatcustomizedcallroutingcanoffer,thereareseveraladdedadvantagesofimplementingskills‐basedroutinginyourcallcenter:

5.1 IncreasedFirstCallResolution

FirstCallResolutionRatesareconsideredthebestmeasureofcustomersatisfactioninacallcenter.AnincreaseinFCRratesleadstoaproportionateincreaseincustomersatisfactionlevels.OftenFCRratesdropbecauseofpotentialskillmismatchesbetweentheagentandthecallrequirements.Oneof thebiggestadvantagesofawell‐configuredSBRsystemis theability to increaseFCRrates.AnincreaseinFCRratesalsotranslatestoloweroperatingcosts,improvedrevenue,andhigheragentsatisfaction levels, as satisfied customerswill not shift to competitors and a call with a positiveoutcomereducesagentstresslevels.

5.2 ReducedAgentWorkTime

Whentherightagenthandlesacustomercall,thequeryislikelytogetresolvedfaster,reducingthecalltimeaswellaspost‐callworktimeoftheagent.Therightagentisalsolikelytospendlesstimeonnon‐valueaddingactivitiessuchasdiscussingirrelevantissues.Thefocusonreducedworktimeoftenresultsinagentsattemptingtofinishthecall,regardlessofwhetherthecallerwassatisfiedornot.WithSBR,thecallcenterwillbeabletoreducecallhandletimewithoutcompromisingoncallquality.

5.3 DecreasedCustomer‐HoldTime

OneofthebiggestproblemswithSBRintheinitialdayswasanincreaseincustomerholdtimeduetodedicatedagentshandlingeachqueue.Withtheadventofmultiplequeues,SBRhasenabledcustomerholdtimetobedrasticallyreduced,asanycallsthatcrossathresholdwaittime,areautomaticallyroutedtoanewgroupofagents.

5.4 ReducedOperationalCosts

Asthecall‐handlingtimeandcustomerholdtimereduces,theper‐minutechargesforthetelephonenetworkalsocomedown.Thus,SBRcontributespositivelytowardsreducingtheoperationalcostsofthecallcenter.

5.5 IncreasedCallCenterRevenues

Skills‐basedroutingmodelscanhelptoincreasecallcenterrevenueinthefollowingways:

1 Increasingthenumberofcallshandledbyagentsinagivenperiodoftime:Ascall‐handlingtimereduces,agentsareabletohandlemorecallswithinagivenperiodoftime.Thisautomaticallymeansthatyouneedalessernumberofagentstohandlethesamecallvolume.Thus,thecallcenterisabletohandlemoreaccountswiththesameheadcount,improvingrevenues.

13


2 Increasingrevenuesources:Ascustomers’satisfactionrisesduetocommunicatingwiththerightagent,theybecomemoreopentocrosssellandupselloffers.Thiscombinedwithusingagentswiththerightskillsforsaleshelpscallcentersinoptimizingtheirrevenuegenerationpotential.

5.6 IncreasedCustomerRetention

Ascustomersarehandledbyskilledagents,thecustomerchurnduetopoorcallcenterexperiencedrasticallydrops.Thus,asuccessfulSBRmodelcanhelpincreasecustomerretentionlevelsinthecallcenter.

14


6 CriticalSuccessFactors

The concept of skills‐based routing is deceptively simple – all you need to do is to map eachcustomer contact request to the right agent. It should ensure that the agents are neither over‐qualifiednorunderqualifiedtohandlethecallandthattheyhavetheappropriateskills.Fromthecallcenter’sperspectiveitisalsoimportantthateveryagentbeutilizedfullyandthattheSLAsonproductivityandefficiencymetricsaremet.WhileSBRseemslikeagreattooltoachievethebestofbothworlds–costefficiencyandcustomersatisfaction–inthepracticalworld,ithasahighfailurerate.Thisisbecausetheoftenbotched‐upimplementationcanaddalayerofcomplexitytoalreadycomplexcallcenterarchitecture,withoutofferinganycorrespondingrewards.

Inthissection,wewilllookatsomeofthekeycriticalsuccessfactorsforasuccessfulSBRmodel–beingcustomer‐centric,easeofuse,properworkloaddistributionandplannedimplementation.

6.1 BeingCustomer‐Centric

OneofthekeyelementsofsuccessfullydesigninganSBRmodelistostartwiththecallers’needs(calltypes)ratherthantheagents’skillsetsEffectivelysortingandcategorizingcalltypesgivesamapofthecustomers’requirements,profiles,andpurposeofcalls.Basedonthis,skilllevelsneededfor each category of calls can be identified. This can then be used to create agent profiles thatcloselymatcheachcallcategory.Theagentscanthenbemappedtoeachprofilebasedontheirskillset.Sometimes,thisexercisealsohelpstounearthtrainingneedsforyouragents,enablingtherightnumberofagentstobeavailableineachprofiletype.

Youwillalsoneedtoinvestintherighttechnologytocapturetherelevantcallinformation,whichensuressuccessfulcallrouting.ThismayincludeastateoftheartANIsystem,DNIS,IVRtechnologyandmaybeevenhumanintervention.Itisimportanttoinvestinthesupportingtechnologyaswell,ifyouwanttomakesurethatthemodelissuccessful.

6.2 EaseofUse

Ifthemodelistoocomplicated,itoftenfails.Forexample,theSBRdesignhastobealignedtotheforecastingandschedulingmodelsintheworkforcemanagementsystem.DefiningtoomanyskillsintheSBRsystemmeansthattheworkforcemanagementsystemhastobeabletoforecastagentrequirements foreachskillgroupandthencreateschedules foreachagentgroup. Ifcallvolumesforaparticularcalltypearelow,thenchancesoferrorsintheforecastingandschedulingmodulearelikelytobehigh.Inaddition,anydeviationfromagentschedulecanwreakhavocinthesystem.Forexample,ifthereisonlyoneagentscheduledtohandleacalltype,andthatagentisabsent,youwouldrealizeitonlywhensuchcallersallgointoaninfinitequeue.

It ismuchbetter toplan for that rarelyused foreign language skill byusinganagentwithotherskills,andtraininghimonthelanguageratherthanhaveadedicatedagentforthespecificlanguage.Thiswould help to keep costs low and enable the agent to bepart ofmultiple skill groups, thusmaking it easier to plan and schedule. Another problemwith SBR is unevenworkloads, and call

15


centersupervisorshavetoplanandmanagethesysteminsuchawaythatcallvolumesaremoreorlessuniformlydistributedtoavoidexcessiveagentturnover.

6.3 PlannedImplementation

Evenifyouhavesuccessfullyidentifiedthecalltypesandagentskillprofiles,itisstillchallengingtomake the SBR implementation a success. SBR implementation involves changes to be made atmultiplelevels–people,process,andtechnology.Atthepeoplelevel,itinvolvesretrainingagentstoensuretherightskillmixandsometimesevenhiringnewagentstofillskillgaps.Italsoinvolvestraining supervisors onhow touse the tool for effective configuration, real timemonitoring andtweaking.At theprocess level, itmay involveachange in theway initial information is capturedduringthecall,soastoroutethecalleffectively.SBRalsoinvolvestechnologychanges–bothatthehardwareandsoftwarelayers.Itwouldinvolvebuildinginterfacesbetweentheexistingworkforcemanagementsystemand theSBRsystem.Allof thismeans that thechangemanagementprocesshas to be handled in a plannedmanner for the implementation to be successful. This involves,implementingapilot,testingitandcorrectingthebottleneckspriortorollingitouttotheentirecallcenter.Ifthecallcentermetricsshowadropinperformance,itisanindicatorthatthedesignmayrequire furtherchange. It isalsobetter togoaboutSBR inaphasedmanner, so that it iseasy toreverttotheearliermodewithrelativeeaseandminimalloss.

6.4 WorkloadDistribution

OneofthekeychallengesofanSBRsystemisworkloaddistribution.Itisimportanttoensurethatthehigherskilledagentsarenotover‐worked. It isalso important tostructure thecompensationpackagetobecommensuratewithskilllevels.Forthis,theHRdepartmentofthecallcenterhastoworkcloselywith theoperations teamtodevise therightcareergrowthpath, trainingcalendars,andcompensationpackages.

Itisagoodideatouseaworkforcemanagementtoolforcreatingstaffschedulesforagentsbasedonhistoricalcallvolumes.Thiscanhelpinensuringthatstaffingisdonebasedoncallvolumesforeachskillset,andanearuniformworkloadisassignedtoagentsregardlessoftheirskills.

Callcentersupervisorsmustalsofocusonthesoftaspectsofagentmotivation.Highlyskilledagentsmaynotalwaysbemotivatedbymoremoney.Otherways tomotivateyouragents includemorevacationtime,rewards,andrecognitionamongteammates.

ThereshouldalsobeadetailedHRplanthatisalignedtotheoperationalplanforthecallcenter,sothat there is a constantprocessofhiring for specific skills and trainingagentswith rudimentaryproficiencyinaskilltoimproveit.

6.5 SuccessfulSkills‐BasedRouting

Asuccessfulskills‐basedroutingprojectshallbroadlyconsistofthefollowingsteps:

1. Identifycallerneeds

16


2. Identifyskill‐setsneededtomeetthecallerneeds

3. Categorizeagentsintoskillgroupsbasedonskill‐setsidentifiedinStep2.Ideally,therewillbe multiple skill groups with preferences assigned to handle each call type. Thus, if allagentsinthemostpreferredgrouparebusy,thenthecallcanbedynamicallyroutedtothenextgroup.

4. Definequeues forcallhandlingbasedonthetechnologyavailable inthecallcenter.Usingbasic ACD technology, separate queues can be created for each skill group. Thismethodpresentsdifficultiesasneitheragentsorcallers candynamicallymove fromonegroup toanother.Thiswouldcreateproblemssuchasunevenworkloaddistributionamongagentsandlongwaittimesforcallers.AdvancedACDtechnologyenablesroutinglogictobeusedtoseamlesslyroutecallstothesecondaryagentswhentheprimaryagentsareunavailable.

5. Define the call routingmodel. If the call routing isdone throughDNIS,differentnumberscan be given to different caller profiles so that calls received through one number areroutedtoaspecificsetofagents.Abetteroptionistouseasinglecallcenternumberandidentify the call type using an IVR menu. While there is a possibility of callers gettingconfusedwithacomplicatedIVRmenu,itisstillbetterthannocustomizationatall.Anotheroption is to use an advanced IVR system. A major portion of the call profiling is doneautomaticallybyidentifyingthecallerusingtheirANI,callerID,accountnumberetc.,whichinturnisusedtogetfurtherinformationaboutthecustomerfromtheCRMsystemsuchasthe caller priority, products used, and even the agent who handled the last call. Thiscomprehensiveprofileisthenusedtoroutethecalltothebestagentforthecall.

6. Train/re‐skillyouragents.Insomecases,agentsmayneedtobere‐skilledinordertomatchthe requirements of the SBR system. In other cases, agentswhoalreadypossess the skillmayneedtobetrainedtoimprovetheirproficiencylevelsinordertoimprovecallquality.Thisisespeciallytrueforrevenuegeneratingcallcenters,aseachagentwhocansellbetterand identifycross‐sellandup‐sellopportunitiesdirectlycontributes to the top‐lineof thecallcenter.

7. ConstantImprovement:InordertokeeptheSBRworkingwell, it isimportanttocriticallyanalyzecallcentermetricsreportsandcustomerfeedback.Usetheseasinputstofinetunethesystem,sothatthebestcombinationofskillprofilesandcalltypesisachieved.

The success of a skills‐based routing implementation is measured by an evaluation of theimprovements generated. However, often call centers fail to measure the current state prior toimplementationagainstwhichimprovementscanbemeasured.Forexample, ifyouarerunningacontactcenterwithmultiplecontactchannelsand if theobjective is toachievea90%matchrateregardless of the channel of contact, then it is important tomeasure thematch rate per channelpriortoimplementingSBR.Thiswillhelptoidentifythebottlenecksatachannellevelincasethetargetedmatchrateisnotobtained.Itisimportanttoobtainthecurrentstatefiguresatagranularlevelpriortoswitchingovertothenewsystem,sothattheSBRmodelcanbetweakedintheright

17


mannerforimprovingresults.Itisalsoimportanttosetinterimbenchmarkstomeasureprogresstowardssuccessaschangeswillnotcomeovernightanditisnecessaryforthesupervisorstoknowwhethertheimplementationisgoingintherightdirection.Knowingwhereyouareatthestartoftheprojectandatspecificintervalswillhelptomeasure,evaluate,andreportsuccessinaneasiermanner.

18


7 Conclusion

Withthediverserangeofsolutionsavailablefromvendors,itisimportantthatcallcenterschoosetherightonefortheirneeds.Forexample,acallcenterhandlingmultipleproductlinesmayoptforamultiplequeueapproachwhereasacustomerservicecallcenterhandlingdifferentqueriesaboutasingleproductmaywanttoassignmultipleskillstoagents,andselectthembasedonavailability.Therequirementsofallstakeholders–thecustomers,thebusiness,andthecallcenter‐havetobeevaluated before choosing the best option to implement.If you are looking for ways to increaserevenueandcustomerretentionthroughyourcontactcenter,startwithskills‐basedrouting.Theremaybenobetterprojectforadramaticimpactinyourcontactcenter.Skills‐basedroutingisoneofthemostvaluableandimpactfulchangesyoucanimplement.Hereareconsiderationsofyourskills‐basedrouting:1.Identifytheskillsthatyourcustomersneed.2.Identifytheskill‐setsofyouragentpoolthatmatchyourcustomerneeds.3.Segmentyouragentsintoskillgroupsbasedontheirskills.4.Defineyourqueues.5.Definehowcallswillshouldberoutedtothoseagents.6.Trainyouragents.7.Monitor,measureandadjust.

19


8 References

1. Rohde,D. (1995). Siemens rolm enhances skill‐based call routing.NetworkWorld,12(35),24‐24.

2. Rodney, B. W., &Whitt, W. (2005). A staffing algorithm for call centers with skill‐basedrouting.Manufacturing&ServiceOperationsManagement,7(4),276‐294.

3. Fried, J. (1997). Skills‐based routing leverages agents' strengths. Telemarketing & CallCenterSolutions,15(8),26‐26,132+.

4. Fiszer, M. (1997). Customizing an inbound call center with skills‐based routing.Telemarketing & Call Center Solutions, 15(7), 24‐24. Retrieved fromhttp://search.proquest.com/docview/208160459?accountid=10559

5. Jordan, B. (1997). Skills‐based routing: An industry survey.Telemarketing&CallCenterSolutions, 15(8), 134‐136. Retrieved fromhttp://search.proquest.com/docview/208148580?accountid=10559

6. Passios, T. (2009). Ask the EXPERTS. Customer Inter@ction Solutions, 28(6), 20‐21.Retrievedfromhttp://search.proquest.com/docview/208158451?accountid=10559

7. Klenke, M. (2001). Skill‐based routing: Boon or bust? CC News. Retrieved fromhttp://search.proquest.com/docview/225206815?accountid=10559

Skills Based Routing - Call Center Service Provider · PDF filefamiliarity with specific...

Documents

Transcript of Skills Based Routing - Call Center Service Provider · PDF filefamiliarity with specific...