BEST PRACTICE

Early adopters of customer relationship

management systems were often disappointed

by high costs and elusive benefits. Now some

companies are reaping strong returns on their

CRM investments.

CRM Done Rightby Darrell K. Rigby and Dianne Ledingham

THROUGH THE LATE 1990S and JlltO

2000, managers piowed miilions ofdollars into information systems meantto track and strengthen customer rela-tionships. Often built around complexsoftware packages, these customer re-lationship management (CRM) systemspromised to allow companies to re-spond efficiently, and Jt times instantly,to shifting customer desires, therebybolstering revenues and retention whilereducing marketing costs. But mostlirms failed to reap the expected bene-fits, and as executives dramatically re-duced IT expenses in subsequent years,CRM sales plummeted. After rising 28%between 1999 and 2000, CRM salesdropped by 5% in 2001,25% in 2002, and17% in 2003, according to the technologymarket research firm Gartner. Manyobservers came to believe that Cl^Mwas destined to join enterprise resourceplanning (ERP) as another overhyped

IT investment whose Initial unmetpromise nearly killed of^'the approach.

But something miexpecled has hap-pened: Senior executives have becomeconsiderably more enthusiastic aboutCRM. In 2003, Bain ik Company's an-nual Management Tools Survey of 708global executives found that firms actu-ally began to report increased satisfac-tion with their CRM investments. In2001, C KM had ranked near the bottomot"a list of 25 possible tools global exec-utives would choose. Two years later, ithad moved into the top half. In fact,82% of surveyed executives said theyplanned to employ CRM in their com-panies in 2003-a large iumpfrom the35"o who employed it in 2000. Today,CRM spending appears to be pickingup. Gartner forecasts that overall CRMsales will rise another 10% by the end of2005. So what's changed? Why has dis-appointment turned to satisfaction, pes-

simism to optimism, cutbacks to newspending?

To answer t hese questions, we studieda wide range of companies that haverecently been successful in implement-ing CRM systems, and we discoveredsome common threads in their experi-ences. Most important, they've all takena pragmatic, disciplined approach toCRM, launching highly focused proj-ects that are relatively narrow in theirscope and modest in their goals. Ratherthan use CRM to transform entire busi-nesses, they've directed their invest-ments toward solving clearly definedproblems within their customer rela-tionship cycle - the series of activitiesthat runs from the initial segmentingand targeting of customers all the waythrough to wooing them back for more.

The successful users have also exhib-ited a healthy skepticism, discountingoverblown claims that the ultimate pay-

18 IIAliVARD BUSINESS REVIEW

back from a CKM system is the creationof a "real-time enterprise." Understand-ing that liigiily accurate and timeiydata are not required everywhere intheir businesses, they've tailored theirreai-time CRM initiatives to those partsof their customer reiationships thattruiy do depend on "perfect" informa-tion. Once they've succeeded with thesmaller, more-targeted CRM project,they've used their initiai investmentsas springboards for solving additionalproblems.

We've distilled the experiences oftheCRM leaders into four questions that allcompanies should ask themselves asthey launch their CRM initiatives:

• Is it strategic?• Where does it hurt?• Do we need perfect data?• Where do we go from here?The questions reflect a new realism

about when and how to deploy CRM to

its best advantage. When Darrell Rigby,Fred Reichheld, and Phil Schefter tookstock of CRM's effectiveness in thesepages nearly three years ago (in "Avoidthe Four Perils of CRM," February 2002),lots of companies were still placing bigbets that the technology would payoff- somehow. Lacking clear customerstrategies and the organizational struc-tures to support them, many firms gotburned and grew distrustful of CRM.T he difficult lessons such organizationslearned have led them to sharpen theircustomer strategies, setting the stagefor real gains from more-focused CRMapplications.

In this article, we'll show how severalcompanies have implemented success-ful CRM efforts-aircraft parts distrib-utor Aviall, consumer product giantKimberly-Clark, diversified equipmentmaker Ingersoll-Rand, home-and office-machine company Brother Interna-

tional, and electronic connector manu-facturer Molex. We'll also lay out somebasic considerations that can help firmsdetermine which CRM proiects arelikely to yield the most value.

Is tt Strategic?There's no getting around it: A CRMprogram involves complicated businessand technology issues and requires sig-nificant investments oftime and money.CRM is nt>t a tool for buffing a com-pany's performance at the edges; itshould be applied only to processes vitalto a company's competitiveness-thosethat can distance a firm from its com-petitors or keep a function (such as callcenter response time) on par with therest ofthe industry when parity counts.If the target is not truly strategic, theorganization will be hard-pressed tosummon the vigor necessary to tackleentrenched business processes or retoolits organizational structure and gamerexpected returns. Before spending a dimeon CRM, therefore, executives need tomake sure they have the right targets intheir sights.

Paul Fulchino knew the stakes in-volved when he brought CKM into Aviallafter being appointed CEO in 2000.Fulchino had ambitious plans to trans-form the Dallas-based distributor of air-craft parts into the premier vendor ofsupply chain management services tothe aviation industry. By becoming thepreferred partner of both the big orig-inal equipment manufacturers (OEMs)and the commercial and military fleetowners, Aviall cotild consolidate cus-tomer demand and extend its reachworldwide, which would reinvigorateits sales and strengthen its margins.

But f-ulchino faced a daunting ob-stacle to realizing his vision: Poor in-formation and cumbersome processeshampered the company's sales and ser-vice operation. Difficulties with an exist-ing IT system had increased sales reps'workloads, sometimes keeping sales-people trapped in local branches, help-ing managers input order informationinstead of making sales calls. What'smore, the company hadn't trained thesales reps in proper time and territory

2004 119

B E S T P R A C T I C E • CRM Done Right

management, which led to inefficientphone call routing and haphazard call-ing schedules. Customer inquiries wereoften routed to distant call centers thatlacked up-to-date data on orders, prod-ucts, and prices.

The weak customer service left keyaccounts vulnerable to competitors'at-tacks and undermined the company'sability to charge the premium pricestypically associated with fiawless prod-uct delivery. A better-trained and moreproactive sales force was a strategicnecessity. Without one, Fulchino's ag-gressive plans for the mmpany wouldgo unfulfilled. So the new CEO, workingclosely with his sales and marketinghead, Jim Quinn, and his technologychief, Joe Lacik, dedicated Aviall's ini-tial CRM outlays to this critical chal-lenge. Rather than attempt a full-scaleimplementation ot a broad CRM pro-gram, the executives took a more fo-cused approach, installing only the salesforce, order entry, and call center appli-cations to begin with. Their goal was tocoordinate customer information seam-lessly from the outside sales agents, firstto the inside sales support staff, then tothe customer service representativeswho were manning the company's ^bregional call centers. The relatively nar-row focus allowed the sales force to be-come familiar with the system withoutbeing overwhelmed and delivered quickvictories that helped win broader man-agement and line support and gathermomentum behind the proiect.

The gains were striking. Before hav-ing the CRM system, the sales force re-lied on an outmoded database for man-aging client information. The system'sinflexibility made it difficult for salesand service staft'ers to get even basicinformation on a customer's order his-tory and credit status. "There's nothingmore frustrating than having a cus-tomer spend 15 minutes on an order andthen realizing at the very end thatthere's a credit issue,"says Lacik."In the

old system, credit problems didn't getflagged until you tried to piace theorder. Then the credit group wotiid becalled In, and you either had to have thecustomer on hold for a long time or callthem back. In our business, there's amoment ot'truth: You have to have theright product, the right information,and the right price. If you don't havethose three things put together,you losethe call-and if you lose the call, 90% of

It's possible to useCRM systems to managethe entire customerrelationship cycle, butthat's usually a bad idea.

the time you lose the sale." With the newsystem, a customer's credit history in-stantly popped up on the order screen.

The rich information the new systemprovided allowed Jim Quinn to flip aswitch in the sales force. It helped theagents get organized and spurred themto make more customer calls, knowingthey could immediately deliver firmquotes on tailored sets of products orservices. Placing an order had once re-quired them to go through 11 screensand nearly 50 steps; now they could doit with one screen and ten steps. Justfour months into rolling out the CRMsystem, the number of daily sales callstripled, and the customer base grew by33%. In fact, the productivity ofthe en-tire sales and service operation skyrock-eted, helping Aviall recapture marketshare and win large orders for new prod-uct lines. The number of orders handledper day iumped from 1,000 to 2,300,even as error rates declined, with no in-crease in staff.

The e.Kpanded capacity, together withthe improvements in service, have builtthe platform the company needed to re-shape itself as a fuli-service provider at

Darrelt K. Rigby (diirrcll.rii;b\'Ca'^bain.com) is a partner with Bain S; Canipany and di-

rects the firm '.s Ghbai Ketail practice. Dianne Ledingham tdianne.ledint;hani(a>bain.com)

is a partner and leader in the company's Technology and Ihrlbrmance Improvement

practices. Both are based in Boston.

aviation logistics support. Aviall's salesand profits have grown rapidly, and ithas steadily stolen market share fromcompetitors. In a testament to Aviall'ssuccess, engine maker Rolls-Royce re-cently awarded the firm a ten-year sup-ply contract worth $3 billion-the largestdeal ever struck by any company in theindustry. Says CIO Lacik: "We showedRolls-Rt>yce the levei of visibility we hadinto our customer base - visibility thatwe couid share with them to give thema deeper understanding of customerbuying trends and behavior.... A simpleanalysis showed Rolls-Royce that it hadseveral years' worth of supply in someproducts while being understocked inothers because it was not matching man-ufacturing adequately with customerdemand. That was a pivotal moment inwinning the contract." Tightly focusedon a single area of critical strategic im-portance, CRM has become a linchpinof Aviall's reinvention.

Where Does It Hurt?It's possible to use CRM systems to man-age the entire customer relationshipcycle all at once-initial purchase, after-sales service, subsequent purchases, rec-ommendations to other customers (forthe full range of functions a CRM sys-tem can automate, see the exhibit,"TheCustomer Relationship Cycle"). But asthe most aggressive early adoptersfound,that's usually a bad idea. Such anapproach ends up creating unused tech-nology capacity, causes unnecessarybusiness disruptions, and ultimatelyfails the payback test. When companiescarefully examine their customer rela-tionship cycles, they usually find somedeep-seated, pernicious problems in afew areas that undermine overall per-formance. It is these pain points thatshould be the focus ofthe CRM effort.

For Kimberly-Ciark,one ofthe world'sleading consumer packaged-goods com-panies, the pain point lay in its vast re-tailer promotions operation. The man-ufacturer was running thousands ofpromotions every year, usually offeringa discount on a particular product to aparticular retailer, but it was unable toaccurately gauge the success of any of

120 HARVARD BUSINFSS REVIEW

CRM Done Right • BEST PRACTICE

The Customer Relationship Cycle

A comprehensive CRM system can, in theory, automate every aspect of a com-

pany's relationship with its customers, from ail the activities needed to target cus-

tomers through those for product development, sales, service, and retention. But

smart companies sharply focus their CRM implementations, carefully choosing

which segment ofthe cycle, and which functions within that segment, are likely

to deliver the greatest return on an initial CRM investment. Success with this first

effort often lights the way to subsequent projects-automating additional func-

tions in the same segment (as Kimberly-Clark did), steadily moving from segment

to segment (as Brother did), or even moving to critical business processes beyond

CRM (as Molex did).

• Concepi develcpmeni

• Feature and function prtontizaiJor

• Customei panel management

• CosI-position requitements management

• Competitive intelljqerceand research

I Segmentaiion

I Behavior modeling

I Scoring and targeting

I Campaign management

I Pricing

I Promotion

1 Win/loss analysis

I Sdles lorecasiing

I Lead managemeni

Bidandquole

management

' Pipeline managemeni

• Cross selling

I Personalization

I Order management

1 Share-of-wallei analysis

I Loyalty program managemeni

I Retention managemeni

I Win-hack campaign management

I Collaborative innovation

I Presalese>;perien(e

i Offering delivery

I Interaction managementand help desk

I Queue managementand escalation

I Service broadcasting

Kimberly-Clark's CRM Path

Campaignmanagemeni

Promotion

Segmentation

Pricing

Brother's CRM Path

interaction managemeniand help desk

Loyally programmanagemeni

Campaignmanagement

Feature and functionpiioritiration

Molex's CRM Path

1 Pipeline1 rnanagement11 i

1 lead 1 Cross1 managemeni | seiling

1i1 Sale^1 forecasting

Budgeting Sjppiycfiainwf̂ • management

thetn.The firm had aggregate numberson its trade promotions, but it couldn'tbreak them down by individual cus-tomer, product, or shipment. As a result,Kimberly-Clark found itself spendinghuge quantities of marketing dollars,uncertain which promotions were prcvducing retailer loyalty, shelf space, andsales, and which were going to waste.Cotnpany executives therefore reasonedthat startingwith a modest,customizedCRM system to collect and analyze prt>motion data could substantially im-prove the effectiveness of its overall cus-tomer relationship cycle.

Kimberly-Clark started by buildingonto an existing software program foraccount management, called Profit Cal-culator, which its sales department haddeveloped to track investments in indi-vidual promotion efforts. By integratingthat with shipment data, the enhancedsystem could go beyond just providinggeneral infonnation about whether ROIwas positive or negative. It could moreprecisely measure the impact of a par-ticular promotion on sales and profitsfor both Kitiiberly-Clark and its retailercustomers. Says Bruce Paynter, Kimberly-Clark's vice president for custotner de-velopment; "Now we can see what thereal-time impact on our sales and profitis when rtmning promotions. Moreover,we can integrate this information intoour sales and planning process with ourcustomer." Renamed Business Planner,the software became the heart of thecotnpany's sales and marketing efforts:salespeople used the tool in the field todesign promotional packages for spe-cific retailers, while the company's mar-keting staff used it to plot broader pro-motion plans.

Rolled out to all of Kimberly-Clark'sbusinesses in 2000, and supported byintensive training programs led by theorganization's top executives, BusinessPlanner rapidly proved a success. In itsfirst year, the system was used to man-age more than 2,300 protnotional eventsinvolving all of the company's U.S. con-sumer product lines. "We applied real-time promotional-lift models [modelsof just how much a given promotioncan lift sales I at the tnarket, customer.

2004 121

B E S T P R A C T I C E • CRM Done Right

and category level to did ourefforts with customers," Paynter says."Using the knowledge gained throughthe Business Planner, we have hecn ableto redirect $30 million in marketingspending across all our U.S. consumerbusinesses to drive incremental salesand profit and further build brand forour customers and Kimberly-Ciark."

Equally important, managers say,Business Planner armed customer rep-resentatives with consistent data andbusiness rules, which has broadenedtheir perspective. Rather than thinkpurely of managing sales, they think interms of managing the business. Today,key-account reps can assess likely finan-cial results and engage in scenario plan-ning jointly with retailers.

And their effectiveness in reducingpain in trade promotions has revealednew opportunities. Building on the suc-cess of its Business Planner software,Kimberly-Clark is now implementinga more ambitious system designed toreach beyond its retailer customers intoa wide array of consumer-advertising andpromotional activities. The enhancedsuite, coined Brand Builder, helps thecompany plan and evaluate the successof individual activities-a freestandingcoupon inserted intothe Sunday papers,for instance - and measure the com-bined effect of a number of integratedactivities.

The Brand Builder suite comprisesthree related components: It includesa state-ot-the-art collaborative tool thatlets sales agents, designers, vendors, andretailers plan promotions online. It putsmarketing research and informationleamed about consumers online in realtime. And by integrating promotional-spending data with scanner and finan-cial information, it provides a powerfulanalytical tool. In fact, with the newanalysis tool, Kimberly-Clark has movedfrom relieving a pain point for its re-tailer customers to making a science ofmarketing. The company now knows,for example, that the payback for someconsumer promotion programs is twiceas high as for others intended to pro-duce the same results. With that kind ofinformation, the firm can identify which

elements of marketing-coupon valueor creative impact, for instance - resultin higher returns.

Focusing on pain points can not onlybe an effective way to build a successfulCRM program but can also get an un-successful CRM initiative back on track.That was true for Ingersoll-Rand, the$10 billion diversified manufacturer.

ln 2001, Club Car, the Ingersoll-Randdivision that makes motorized golf carts,or"golf cars" as the company calls them,was showing signs of trouble, with reve-nues beginning to drift downward as aneconomic downturn hit the golf indus-try. But management lacked the infor-mation needed to diagnose the reasonsfor the slowing sales. Individual repsand order managers used their ownidiosyncratic processes for dealing withcustomers. Sales forecasts were madeinformally using guesswork and rudi-mentary spreadsheets, and the salesforce had little infiuence over productcustomization.

Realizing it needed much better in-formation, Ingersoll-Rand rushed to rollout a hroad CRM system that was sup-posed to incorporate everything fromlead evaluation to proposal generationand from product configuration to orderentry. But the effort proved too muchfbr the organization to digest. Club Car'smanagers weren't convinced ofthe ul-timate benefits. After spending morethan $2 million and completing a firstround of user testing, the company dis-covered that the system wasn't deliver-ing the anticipated productivity gainsand reporting capabilities. In fact, thesystem would dramatically increase theadministrative workload of the fieldsales reps instead of freeing them tospend more time with customers. Theunit's president had the foresight tohalt the effort and made the organiza-tion back up and refine its goals. ClubCar's management team took a freshlook at the key processes in its cus-tomer relationship cycle and refocusedits CRM initiative on the two deepestpain points: forecasting sales and tak-ing orders.

Today, just two years after the CRMeffort was relaunched. Club Car has

successfully automated its sales opera-tion, significantly improving both cus-tomer service and business decisionmalting. By more direcfly involving thesales force in the redesign of the sys-tem, carefully paring down the data andprocesses it encompassed, and improv-ing the underlying technology, the com-pany eliminated many ofthe CRM sys-tem's original drawbacks. Sales reps usethe new system at customers' sites tomodify the cars with them, and for thefirst time, the reps can see the financiaiimplications of different configurationsbefore setting prices and delivery dates.The order information the reps collectis automatically combined with generalindustry data on golf cart demand andequipment replacement cycles to gen-erate reliable sales forecasts. That, inturn, has led to smoother, more predict-able manufacturing schedules.

Do We Need Perfect Data?Part ofthe early attraction of CRM sys-tems lay in their ability to deliver real-time information - t o give marketers,salespeople, and managers a clear pic-ture of what's happening in the marketat any particular moment. But perfectinformation comes at a high cost. Thesystems required to collect and dissem-inate it are expensive; so are the finelytuned processes needed to react quicklyto it. Despite the hype surrounding real-time enterprises, the fact is that fewcompanies need perfect informationthroughout their customer relationshipcycles.

Why pay for real-time informationon business processes that customersdon't really value or that managerscan't rapidly adjust? A hotel managercertainly needs real-time data on theavailability of rooms but not on thecustomer's opinion ofthe carpets anddrapes. A cable company needs real-timefigures on service outages that demandimmediate repairs but not on the prof-itability of its pay-per-view programs.Real-time information priorities aredriven by real-time business opportuni-ties and must be customized to each in-dividual business. (See the sidebar "Cal-culating the Cost of CRM.")

122 HARVARD BUSINESS REVIEW

B E S T P R A C T I C E • C R M D o n e R i g h t

Routine Aches Versus Strategic Pain Points

Doctors commonly distinguish between routine aches ("Take two aspirin

and call me In the morning") and perilous pains ("Meet me at the hospitai

in five minutes!"}. Likewise, successful CRM practitioners have learned

to distinguish between routine aches in the business ("Perhaps we might

address that issue in our next five-year plan") and strategic pain points

("Fixing this problem wiil double our profits") before prescribing CRM

solutions. Addressing strategic pain points typically promises superior

financial rewards and the opportunity to build vital momentum for CRM

programs. Here's how to identify them:

Routine Aches Strategic Pain Points

The problem is well known but

minor, even though it affects some

vociferous customers.

Solutions are quickly and easily

copied by any competitor.

The problem could have been

fixed long ago without a CRM

system.

Solving the problem is not vital to

the organization or its culture.

Solving the problem would fix one

immediate problem.

Solving the problem would deliver

soft, unguantifiable benefits.

Solving the problem would not

make much of a splash in the

organization.

The problem is sometimes hiddenbut has a critical impact on the

satisfaction and loyalty of the most

valuable customers.

Solving the problem creates

a substantial and sustainable

competitive advantage.

Solving the problem cost effectively

requires the speed, accuracy, and

effectiveness of CRM technologies.

The solution will become a rallying

point for the organization.

Solving the problem would create

important new capabilities that would

open up additional opportunities.

Solving the problem would deliver

tangible financial returns that would

justify further investment-even in

difficulttimes.

Solving the problem would represent

a highly marketable success, both

inside and outside the company.

Companies need to clearly distin-guish between activities that truly de-mand perfect data and those that canget by with "good enough" informa-tion. The requirements for each arequite different. The approach BrotherInternational took to its CRM imple-mentation is a good case in point. TheU.S.-based distribution arm ofthe Jap-

anese maker of typewriters, printers, fax-printer-copiers, and sewing machines,Brother International faced a persistentproblem: a high rate of product returns.A leading cause ofthe returns, the com-pany believed, was dissatisfaction withservice from its call center. In the late1990s, as office products became moresophisticated, end users began to re-

quire more assistance. But Brother'scall centers were answering only 46%ofthe queries coming in from new pur-chasers, and the quality of the help prc>-vided varied widely. Service represen-tatives were failing to address recentbuyers' questions and complaints. Inparticular, call center staffers lacked ac-curate customer information and quickaccess to solutions for callers' problems.To help customers troubleshoot techni-cal issues, staffers often had to searchthrough binders of product information.Frustrated consumers were returningtheir products to retailers.

FHere, Brother's executives saw, wasa pain point that could be remediedonly through the provision of "perfect"information; they therefore looked toCRM to bolster their call centers. Thecompany rolled out the new system instages, starting with the printer call cen-ter in September 2001 and then addingcenters at two- to four-week intervals.This staggered approach allowed Brotherto refine the system as it was imple-mented and adjust the training programas circumstances warranted.

The results have been impressive.Thesystem can identify customers as theycall in, quickly locate their purchaserecords, and supply call center workerswith standard responses to commonquestions. That's reduced Individualcall times by 43 seconds on average, re-sulting in substantial savings. Brotherestimates, for example, that this yearthe total savings could reach $635,000.What's more, Brother is now answeringan average of 140,000 calls a month, andthe typical customer is left on hold forless than five minutes. The newly auto-mated process has also cut the time re-quired to train new call center opera-tors, saving even more money. Productreturns fell by a third, from S-0% in fiscal2000 to 3-4"̂ ' the following year.

And the benefits reach beyond thecall centers. Because the system can cap-ture data on the nattire of incoming calls,it has given the company importantnew insights into customers'needs andbehavior. That has improved Brother'sability to tailor outreach campaigns,

continued on page 12S

124 HARVARD BUSINESS REVIF.W

B E S T P R A C T I C E • CRM Done Right

Calculating the Cost of CRMWith clear thinking and a basic grasp

of decision diagrams, any manager

can estimate the true value of infor-

mation. Let's take a look at a dis-

Todd Green, Ace's owner, was deeply

concerned when a new competitor en-

tered his area. A newspaper poll indi-

cated that 20% of his customers planned

guised example we'll call Ace Grocery. to switch. But a consultant said that 50%

1 Do nothingRewardto prevenl

Forecast defectionscorrectly (ilOOeach)

Retain [50'fc)

Expectedvalue

vulnerabiliTy

[20% = 200)^ • V 100% = 200

No reward

No reward

Defect (50%)

Defect (100%)

Defect (100%)edly

ptedicB loydlty

Loyal

Farecas! Rewdrctincorrectly to preveotprediiis defection;vulnerability — Retain (100%)

i-orecaslcorrecilypredictsloyalty

Noreward ^ $800K $800K-•

Retain (100%)$800 K

Number of customers = 800

2 Give $100 to everyone

Forecastcorrectlypredictsviiinerabililv

Rewardto preventdefectionsiiOOeachi

Expectedvalue

-a $80K

Forecastcorrecliypredictsioyaiiy

$800K

Number of customers = 900

ofthe potential defectors could be re-

tained if Ace offered each $100 in reten-

tion rewards. Unfortunately, spending

the $100 on loyal customers would proba-

bly not make them more loyal, and Ace

didn't know which customers were which,

Todd began to sketch out his options.

1 Do nothingAce had i,ooo customers, each worth an

average of $1,000. If Todd did nothing,

he'd retain only 80% of them, and Ace's

value would fall from $i million to

$800,000. Todd tried another option.

2 Give $100 to everyoneTodd would lose fewer customers but

would spend so much that the store's ex-

pected value would still fall to $800,000.

Todd began to wonder if he could use

CRM data to predict which customers

would defect. FHe estimated how much

that would be worth,

3 Reward just the right customersIf Todd could manage to offer $100 to

all (and only) the potential defectors,

Ace's expected value would only fall to

$880,000. So $80,000 became the most

Todd would pay for perfect information.

4 Purchase a CRM systemTodd called CRM Systems, a software

vendor with the best package for predict-

ing defectors. The system cost $20,000,

but Todd estimated that additional im-

plementation costs would add another

$30,000, That $50,000 was still far below

the $80,000 hurdle rate, so Todd asked

for a meeting. He learned that the CRM

system could correctly identify 60% of po-

tential defectors. Unfortunately, 30% of

loyal customers would also be identified

as potential defectors. Todd calculated

the value of this "imperfect" information.

Since the implementation of this

software would drop Ace's value to

$774,000, Todd rejected the proposal,

Todd decided that the best approach was

to offer ali customers $100. Ace would re-

tain 900 customers instead of 800, and

his most loyal customers would not feel

shortchanged.

126 HARVARD BUStNE55 REVtEW

CRM Done Right • BEST PRACTICE

3 Reward just the right customers

Expectedvalue

Rewardto prevent Retain (SO%1 ^ tonu

I -$20K ^$20K

Defect (S0%)

p) . nNo reward ^ ^ Defect (100%)

1 L ,- f \ r-|

Forecast Noreward ^ - ^ Defect(100%)

incorrectly

predicts loyalty

Coit Q!

information

Loyal

Forecast Reward

incorrealy to prevent

predicts defections

vulnerability , — , (SlOOeach)

100% = 800

Forecast

correctly

predicts

Retain(100%)

$800K

Retain (100%)•a $800K

$880K

Number of customers = 900

4 Purchase a CRM system

Vulnerable

(20% = 200)

-$50K

Co;lof

information

Loyal

Forecast

correnly

predicts

viilnerabrlity

>=120

Forecast

incortectly

predi(t5 loyalty

Forecast

incorrectly

predicts

vjinerabiiity

30%-240

70% -- 560

Forecast

cotrectly

predicts

loyalty

Rewardto prevent Retain (50%)

defections i $ 6 0 K

-$12K 6

Expectedvalue

J $48 K

Defect (50%)

No reward o Defect (1DO%)

Defect (100%)

-$50K

Reward

to preveni

defections

(5100 each)

-$24K

No reward

Retail) (lOOt-i)

$240K

$560K

•a $216K

•a $560K

$774K

Number of customers = 860

What Todd LearnedThrough this process, Todd learned

to address four critical questions.

How good is the information?

Todd knew that the value of information

getierally rises with its immediacy (mak-

ing timely actions possible) and accuracy

(making actions more productive). But

Todd learned that reported accuracy often

tells only half the story. Correctly predict-

ing 6o% of defectors sounded good until

Todd's analysis showed that 30% of the

loyal 80% would aiso be wrongly labeled

as potential defectors. That would mean

that a customer identified as a potential

defector would be a true defector only

33% ofthe time. Forecasting is not always

more valuable than guessing.

What is it good for?

Information that helps satisfy customers

is far more valuable than information

that merely satisfies curiosity. Todd calcu-

lated that the imperfect CRM informa-

tion combined with a $ioo retention pro-

gram that saved only 50% of vulnerable

customers created an expected value of

just $774,000. However, ifthe $ioo reten-

tion program could save 100% of vulnera-

ble customers, the same CRM system

would create an expected value of

$834,000. Of course, offering loyalty

rewards to every customer would now

create an even greater value-$9OO,ooo.

What are the costs?

Todd's analysis demonstrated that the

$20,000 CRM costs were swamped by the

additional expenses of training, data col-

lection, data analysis, information dis-

semination, and implementation pro-

grams. If Ace had failed to include these

costs, it would have incorrectly calculated

an expected value of $804,000, chosen to

implement the system, and actually de-

stroyed value.

Which results matter most?

Although expected values were neces-

sary, they weren't sufficient for the final

decision. Several other considerations, in-

cluding the number of retained custom-

ers and fairness to loyal patrons, proved

crucial as Todd weighed Ace's options.

NOVEMBER 2004 127

B E S T P R A C T I C E • C R M D o n e R i g h t

McDonald's Tech Turnaround

It was nothing if not ambitious. In January 2001, fast-food giant McDonalcJ's

launched a five-year, $1 billion initiative to overhaul its information systems

to enhance customer service at its more than 30,000 restaurants around the

world.The vast effort, dubbed "Innovate," involved installing a new, Internet-

based data management infrastructure for the company that would cover

everything from electronic links with suppliers to software for tracking cus-

tomer purchasing patterns to sensors for remotely monitoring the tempera-

ture of fat in french-fry vats. The company believed that once the new sys-

tems were in place, it would be able to scrutinize every detail of its business

in real time, ensuring that each outlet fine-tuned its operation to serve cus-

tomers flawlessly.

But just two years after Innovate's launch, a new CEO, Jim Cantalupo,

pulled the plug on the project, announcing that McDonald's would take

a $170 million write-off related to the termination ofthe initiative. Facing

financial pressures and an eroding stock price, Cantalupo (who died sud-

denly a year later in April 2004) determined that, to revitalize itself, McDon-

ald's needed to focus on two priorities: customers and existing restaurants.

The massive IT program simply would not deliver enough improvements in

those areas to justify the enormous expense and operational disruptions it

presented.

McDonald's might be an extreme example-billion-doilar IT initiatives

remain rare-but it also offers encouraging news. After disbanding the

Innovate initiative, the company launched a series of more modest CRM

efforts-such as installing touch screen-ordering systems and improving a

customer service hotline-aimed at solving particular customer relationship

probtems. These highly targeted efforts have contributed to a resurgence in

sales and profits.

which include surveys and newsletters,to well-defined buyer segments. Nowthat it has better information on thequestions most frequently asked aboutits products, the company can use itscampaigns to disseminate answers in ad-vance. That, in tum, has reduced the vol-ume of inquiries coming into the callcenters. Each day, Brother sends data ontypes of inquiries and customer prob-lems from the call centers to its corpo-rate parent in Japan, where the infor-mation collected helps the company'sproduct-planning, design, and customer-satisfaction groups evaluate both prod-uct performance and customers' prefer-ences and experiences. Brother expectsthe exchange will lead to improved cus-tomer satisfaction and enhanced prod-uct performance over time.

Used the right way, real-time infor-mation can help companies cope withhigh levels of complexity in their cus-

tomer relationship cycle, making prior-ities clear. Molex, an Illinois-based globalmanufacturer of electronic and fiber-optic interconnection systems, has alarge customer hase and a vast pipelineof potential orders. At any given mo-ment, the company is pursuing closeto 15,000 different sales opportunitiesworldwide. For years, Molex used e-mailand spreadsheets to keep track of itspipeline, but the resulting data wereoften weeks out of date. That made itdifficult to consistently set sales priori-ties so the company could pursue theleads with the highest potential. Thelack of updated information aiso madeit hard to synchronize Molex's global ef-forts. Because the firm had so manymaior customers with operations in dif-ferent parts ofthe world, several Molexlocations could be working on similaror related programs for the same cus-tomer without knowing it.

In 2002, therefore, the company in-stalled a CRM system to manage itsorder pipeline. F'or the first time, execu-tives from the CEO on down could seethe full range of sales opportunities inreal time. That made it possible to mea-sure the real value of those opportuni-ties and get updated information aboutthem 24 hours a day, rather than just afew times a year.

The gains were immediate: improvedorder management, more precise salestargets, and better global coordinationof inventory and pricing between re-gions. Since implementation, both thenumber and the value of potential salesin the pipeline have climbed signifi-cantly, as the sales staff has used theshared information to identify oppor-tunities earlier. Molex's managementbelieves it is meeting the original goalofthe project, which was a 5^ increasein revenues.

Once the initial implementation wascomplete, moreover, management real-ized that the data being captured couldalso be used to improve budget plan-ning. The pipeline data now form thefoundation of the revenue portion ofthe budget process, and the companyplans to use the information for partsforecasting and supply chain manage-ment as well.

Where Do We Go From Here?As Kimberly-Clark, Brother, and Molexfound,the data produced by a narrowlyfocused CRM system often reveal ad-ditional opportunities for importantbusiness refinements. And those refine-ments, taken together, can amount toa broad CRM application that extendsacross the company. The difference be-tween this sort of wide-ranging CRMimplementation and the prerecessionCRM applications is that each step inbuilding the system represents a care-fully planned, well-defined advance-ment in strategic thinking. Kimberly-Clark started with trade promotionmanagement then extended its tool setto include total retailer customer man-agement and, more recently, consumermanagement. Brother's cali centers haveenhanced its U.S. marketing and out-

128 HARVARIJ HUSINLSS Kl VITW

CRM Doru' Right • BEST PRACTICE

reach campaigns and even fed its prod-uct development and quality controlprocesses on the other side ofthe world.Molex's ck\ir view into its order pipe-line has led to improvements in bud-geting, procurement, and supply chainmanagement.

Smart CKM adopters don't rest ontheir laurels. They rigorously analyzethe data their systems produce to iden-tify new, well-defined opportunities toextend the technology's power. In mostcases, these opportunities lie in activi-ties adjacent to the customer relation-ship cycle, as the natural path of thesecompanies'CRM expansions show.

Aviall, for its part, plans to extend itsCRM system in two directions-addingupstream links to its suppliers anddownstream connections to its custom-ers. These will give the firm an end-to-end view ofthe aviation supply chain,starting with the status of suppliers' in-ventory and extending all the way tocustomers' requirements for parts andmaintenance. Because it will enable thefirm to better match supply and de-mand, Aviall expects the expanded sys-tem to become an important source ofcompetitive advantage.

It's also often possible to extend thebenefits of CRM to related businessunits. Ingersoll-Rand, for example, rec-ognized that the customers of its ClubCar division - golf courses - were alsopotential buyers of its other divisions'products, such as Bobcat miniexcavatorsand loaders. Lxtcnding its CRM systemto include those divisions could createnew opportunities for cross sel I ing. Thecompany began to do just that in late2002, and already the number of newleads generated has been substantial -an additional $6.2 million worth of leadsfor other Ingersoll-Rand products in thefirst year. That success has led to evenbroader plans.

Today, Ingersoll-Rand wants to useCRM as the glue to bind together allfour of its operating sectors (which repvresent more than 100 worldwide manu-facturing facilities) so that the companycan operate as one integrated enterprisein the eyes of its customers. Like Aviall,Ineersoll-Rand mav utterlv transform

its business through its investments inCRM, but again the changes will comein carefully measured stages, with suc-cess building on success.

Business Before TechnologyYou'll have noticed that we haven't spenta lot of time describing the details ofthe technology in this article. That's in-tentional. In cvalutiting and designingCRM systems, business needs shouldtake precedence over technological ca-pabilities. Managers should not be dis-tracted by what CRM software can do;they should concentrate instead on whatit !,hintli1 do- both for their companiesand for their customers (see the sidebar,"McDonald's Tech Turnaround"). For-tunately, as competition among CRMvendors is increasing the software israpidly becoming more tlexible. It's notyet simple to install a CRM system, butthe technology is getting more depend-able, the implementation process is be-coming more streamlined, and the fail-ure rate is going down.

That gives companies the freedom toapply CRM with greater precision, tar-geting critical gaps in their customer re-lationship cycles where performancesuffers. By setting priorities for their in-formation requirements carefully, mak-ing sure they're guided by overall cus-tomer strategy, companies can launchhighly disciplined CRM efforts that willhave a greater impact with lower in-vestment and less risk. CRM, in otherwords, is coming to resemble any othervaluable management tool,and the keysto successful implementation are alsobecoming familiar: strong executiveand business-unit leadership, carefulstrategic planning, clear pertbrmancemeasures, and a coordinated programthat combines organizational and pro-ccss changes with the application ofnew technology. No longer a black hole,CRM is becoming a basic bLiilding blockof corporate success. ^

Reprint kO4nH; HBR OnPoint 8355To order, see page 151.

We comfortably fit 123 millionpeople on a small island.

Just imagine what we can dowith our business class.

AIMAA STAR ALLIANCE MEMBER » ' '

t BOO 2 FiV-ANi WWW.FLY-ANA.COM

NOVEMBER 2004Daily Non-Stop Flights to Japan from NYC, LA, DC and San Francisco.

Harvard Business Review Notice of Use Restrictions, May 2009

Harvard Business Review and Harvard Business Publishing Newsletter content on EBSCOhost is licensed for

the private individual use of authorized EBSCOhost users. It is not intended for use as assigned course material

in academic institutions nor as corporate learning or training materials in businesses. Academic licensees may

not use this content in electronic reserves, electronic course packs, persistent linking from syllabi or by any

other means of incorporating the content into course resources. Business licensees may not host this content on

learning management systems or use persistent linking or other means to incorporate the content into learning

management systems. Harvard Business Publishing will be pleased to grant permission to make this content

available through such means. For rates and permission, contact permissions@harvardbusiness.org.