September 2014Number 20

McKinsey on Payments

Foreword

Recapturing banking’s analytics edgeBanks were pioneers in the early days of analytics innovation. However, to regaintheir position in the vanguard, they need to adjust their approach for a new era. The task begins with a redesign of organizational structure for the analytics group.

End-to-end digitization for securities servicesThe securities services industry accounts for nearly $100 billion in worldwiderevenues. But with seismic changes underway in regulation, customer expectationsand revenue structure, industry players will need to target gains in efficiency,accuracy and scalability.

A mobile path to financial inclusion: An interview with Kamal Quadir,CEO of bKashMcKinsey on Payments sits down with the leader of Bangladesh’s mobile moneyservice to discuss how the firm got its start, how the model works and the factorsbehind its success to date.

Transforming national payments systemsA number of countries are planning upgrades to their payments infrastructures.McKinsey argues that these systems need to be better, not just faster, than thosethey are replacing. The path forward starts with a comprehensive look at use casesand design options.

1

3

11

18

23

In the late 1980s and early 1990s, gettingenough data was the main challenge. In thesecond wave of analytics now breaking, datais abundant; the challenge is in handling thevolume, making sense of it and acting on it.

Virtually every bank in the world has asuite of advanced analytics initiativesunder way. However, these initiatives tendto come up short on measurable results,and often fail to even aspire to transforma-tive impact.

Roadblocks to advanced analyticsinitiatives

Advanced analytics initiatives fail for threeprimary reasons:

1. Inspired by a tool or technique, not bythe use case

Many of the datasets and techniques avail-able today are seriously “cool”—at least toprofessional analysts. Unfortunately, toomany advanced analytics initiatives are se-duced by the “wow factor” and squandercredibility and business sponsorship whenefforts fail to deliver the expected rewards.

One leading Scandinavian bank invested ina tool that harvests near-real-time datafrom transactions and Web-service plat-forms to create a virtually live 360-degreecustomer view. The bank could have usedthis view to improve its management of cus-tomer credit risk, attrition risk and recep-tiveness to upselling. However, because ofhurdles in embedding the tool into its risk,service and sales systems, it failed to delivertangible value.

3Recapturing banking’s analytics edge

Prithvi Chandrasekhar

Robert Mau

Recapturing banking’s analyticsedge

The banking industry has long been at the forefront of analytics innovation.

Long before big data became a buzzword, banking embraced the use of very

large datasets in credit bureaus, pioneered predictive modeling in risk and fraud

assessment, and introduced real-time decision engines to enhance the customer

experience. These innovations are now deeply embedded in the banking

industry, and have been widely adopted by other sectors. More recently, though,

banks have struggled to stay ahead in data and analysis, and find themselves

being overtaken by cutting-edge technology companies and retailers.

4 McKinsey on Payments September 2014

The most successful advanced analytics ini-tiatives work backward from business prob-lems, not forward from the possibilities ofdata or technologies. This fact is especiallyimportant at a time when technology is de-veloping so rapidly.

2. Too compartmentalized

When advanced analytics is done well, dataand insights from one area can have a bigimpact on others. For example, merchanttransaction patterns are now used to designand optimize spend stimulation campaigns,and can also be used to reduce fraud andrisk. However, banks often fail to capturethis cross-functional potential because spendstimulation is owned by one departmentwhile fraud protection is owned by another.

At one leading European bank, merchantdata is owned by marketing, which has itsown embedded analytics team that ishighly proficient at using the data to opti-mize campaigns. However, the data is noteven visible to the bank’s risk analysts, whosit in another building across town. It isn’tthat the marketing team is trying to hoarddata; it is just that the effort to share dataacross functions has never been given pri-ority in a demanding and resource-con-strained environment.

The most successful advanced analytics ef-forts take a broad view of data, span the

whole organization and engage all businessleaders. They set priorities for opportunitiesand data access in accordance with strategicgoals and operate across functions and busi-nesses. By adopting this approach, somecompanies have developed enhanced salestechniques capable of unlocking hundreds ofmillions of dollars in additional revenues.

3. Targeted at small improvements

Too many analytics initiatives are incremen-tal—led by support functions, serving a sin-gle departmental customer, seekingmarginal improvements over existing strate-gies. Working on such narrow initiatives canmake executives feel like Gulliver being tieddown by the tiny people of Lilliput.

One leading pan-European personal loansprovider is investing in an analytics platformto integrate data from the credit bureau,customers’ Web-browsing history, serviceoperations, and fraud in close to real time.The business case for the new platform isbuilt on improving risk prediction. Impor-tant though this is, such a focus is undulylimiting. Designing the new platform to sup-port multiple applications from the outsetwould be cheaper and more effective thanbroadening later on.

The most successful initiatives address mul-tiple business problems and have a clearview of the implementation path. Now thatdata and analytics have the power to changethe game, banks need to raise their aspira-tions for the scope of their analytics efforts.They have access to so much rich and power-ful data; it’s time they made better use of it.

To recapture its historic leadership in analyt-ics, the banking industry needs to set up ini-tiatives that are cross-functional, tied to

Banks need to raise theiraspirations for the scope of theiranalytics efforts. They have accessto so much rich and powerful data;it’s time they made better use of it.

specific use cases and ambitious in scope.That means organizing analytics appropri-ately—a process that involves more art thanscience—and pointing it at the right chal-lenges. This article looks at some of the is-sues that arise in organizing for analytics,and then explores two areas where efforts arestarting to have a big impact on both cus-tomer experience and bottom-line results.

Organizing for analytics innovation

At the moment, the possibilities opened up bytechnology are running far ahead of the busi-ness applications. Business leadership and or-ganization are the main limits to innovation.

Business leaders who understand data andanalytics, have a vision of the real-worldpossibilities they can unlock, and know howto marshal the organization to deliver onthat vision are rare. And even a great busi-ness leader needs a strong support structure

before analytics can become central to acompany’s culture (Exhibit 1).

The support structure for analytic innova-tion starts with internal consultants whocan apply advanced techniques to prob-lems and opportunities and teach as theywork. Data strategists are needed to struc-ture existing data and ensure that futuredata is captured in a usable and meaningfulway. Finally, successful innovation requiresdata scientists who are skilled at buildingmodels and creating tools that anyone inthe organization can use to access the in-sights generated.

Hiring people with these skills has neverbeen easy, and will get steadily harder. Ana-lytical talent is in short supply across theboard. The McKinsey Global Institute fore-casts a shortfall of 140,000 to 190,000 ana-lysts by 2018 in the U.S. alone.

5Recapturing banking’s analytics edge

Analytics

Head of analytics

Analyticsconsultants

Datascientists

IT Datastrategists

Businessdecision-makers

Source: McKinsey analysis

Exhibit 1

To derive value from data analytics, banks need talent in a number of specific roles

Provide solid understanding of statistics and analytics to inform business decisions

Leads strategy design and execution for data and analytics; provides links between IT, analytics and business

Ensure data is captured and stored in usable and scalable ways

Ensure internal customers are supported by best-in-class models and algorithms

6 McKinsey on Payments September 2014

Assuming that it manages to hire the righttalent, a company then needs to organize itinto a structure. There are three broad ap-proaches (Exhibit 2), with a variety of hybridoptions:

The business-driven approach, where ana-lytics is embedded in individual businessunits and functions, works well for aligninganalytics efforts with specific business prob-lems, and is often preferred by companieswith a relatively long and successful tradi-tion of analytics. However, the model tendsto suffer from incrementalism.

A centralized approach with analytics ex-pertise concentrated in an independent unitis often favored by companies that are put-ting a stake in the ground and declaring that

analytics will be a big part of their future. Itgives the group a chance to learn, experi-ment and grow analytics knowledge in allbusiness units. It also tends to be the bestsolution for controlling investments and en-suring data standards are maintained acrossthe organization. Unfortunately, this ap-proach can take years to get off the groundbecause it involves not only building a teambut also earning the respect and trust of thewhole company.

A top-down or CEO-sponsored approach candeliver the quickest results and cut throughbusiness silos. On the other hand, a high-powered central team runs a bigger riskthan other approaches of producing whiteelephants that undermine the broader

Most targeted Best individual results

Tight controls and data structuresMore cost-effectiveCan drive impact across the business

Strategic and high impactQuickly gains traction

Difficult to scaleBusiness leaders must be comfortable with analyticsReinforces silos

Pros

Cons Difficult to allocate resourcesSeparated from individual business problems

Resistance from other business leadersRisks being a flash in the pan

Business-driven Centralized (third-party or independent unit)

Top-down

IT/CFOIT CEO

Integrated group (chief analytics

officer)

Independent analytics group

Business functions

Analytical groups

Data quality &

manage-ment

Source: McKinsey analysis

Exhibit 2

When selecting an organizational approach to analytics, banks must weigh the pros and cons of three models

organization’s support for analytics. Toavoid this trap, the central team must takecare to ensure that its efforts support andempower business leaders.

In practice, elements of two or all three ap-proaches are often blended. The most effec-tive approach for any company will dependon its growth aspirations, cost constraints,leadership skills, talent, and a host of otherconsiderations.

Getting the most out of scarce analytics tal-ent is a defining challenge for any bankCEO. People need to be inspired with a com-pelling consumer-centered vision, empow-ered to deliver, and supported with the righttools and technology.

Fraud: Supercharging preventionefforts

Fraud control is one of the business areaswhere analytics is most deeply embedded forbanks. Widely used industry tools such asFICO’s Falcon are based on advanced neuralnetwork algorithms that give lenders an up-to-date view of fraud risks. Most of thesetools are managed within departmental silosthat are accountable for fraud losses and op-erations costs. Because these groups are so

used to dealing with data and incorporatingit into their operations, they are often a goodplace to kick off an effort that graduallywidens in scope.

So how should banks glean insights fromtheir fraud control? It’s likely that most arealready attuned to finding links betweendata and fraud events. The real impact kicksin when they find links between fraud eventsand other factors such as account renewal,satisfaction scores and cross-sell results—oreven better, between these factors and pur-chasing behavior. Joining the dots can helpto reduce fraud and increase customer satis-faction, card usage and other metrics.

One European bank found that the realvalue of fraud management was in market-ing. Market research showed that the biggestdrivers of customer satisfaction (and dissat-isfaction) were fraud-related experiences,ranging from irritation at having transac-tions declined while travelling to gratitudefor protection when a card was lost. Furtheranalysis showed that phone calls from thefraud department to verify transactions sig-nificantly increased customer satisfaction, asthey were seen as evidence that the bankcared about its customers.

With this in mind, the bank redesigned itsentire fraud-control process to serve a newpurpose: maximizing customer satisfaction.Fraud losses and operational costs were nolonger goals, but constraints. The companyestablished new data streams to enable cus-tomers to show it their cellphone location,and built new analytical models to segmentcustomers according to their sensitivity tofraud. It redefined contact frequency norms,increased staffing levels and retrained fraudinvestigators.

7Recapturing banking’s analytics edge

Getting the most out of scarceanalytics talent is a definingchallenge for any bank CEO. People need to be inspired with a compelling consumer-centered vision, empowered to

deliver and supported with the righttools and technology.

8 McKinsey on Payments September 2014

It took a year-long cross-functional effort toembed this new approach, but it paid off.Customer satisfaction scores increased frombelow average to the top quartile, and thenew strategy covered its costs in less thanthree months.

To make an initiative like this work, compa-nies need business leaders who are willing towork with fraud and other departments for apositive outcome. And some initiatives re-quire close collaboration with a third party(Exhibit 3). In such cases, addressing issuesof data governance and privacy can be aschallenging as the analytics themselves.

Digitization: Beefing up onlineservicing and marketing

Web service, like fraud control, is a well-es-tablished part of the banking ecosystem. It istypically managed as part of operations—asa low-cost way of getting statements to cus-tomers and collecting payments from them.

Some banks have made big efforts to use theirWeb service platform to increase sales. Today’spowerful algorithms support targeted adver-tising that uses both browsing history andstatement data to show the right ad to theright prospect at the right moment. Even so,person-to-person channels still account for

A joint venture with a telecom company enables one bank to use location information to reduce fraud . . .

The telecom company knows you are in Norway

A large European bank fights fraud by combining its own card transaction data with location data from a mobile telecom company

“Does the telecom company have the right to sell the

customer’s location to the bank?”

“Is the bank obliged to take the data from the telecom company

to prevent that fraud and protect other banks and customers?”

“Should the customer opt in and instruct

the bank to get the data from the telecom company?”

“What does it cost? Who should pay?”Fraud?The bank sees

a transaction using your card in Spain

. . . but the partners are still trying to forge a workable governance approach

Source: Interviews with industry experts; McKinsey analysis

Exhibit 3

A credit card issuer is using telecom data to fight fraud

the lion’s share of sales volumes. Finding waysto tip the balance and generate more salesfrom Web and mobile channels could enablebanks to capture significant untapped value.

One Latin American bank found that a lackof basic banking knowledge among its cus-tomers was the biggest barrier to widespreadtake-up of its online services. Customers feltthere was no one to answer basic questions,such as: Which product should I apply for?What do I do if I forget my PIN? How do Imake a payment?

The bank responded by launching a series ofshort online films to guide customersthrough every aspect of the online bankingprocess. Instead of using analytics to identifywhich products to sell to which customers,the bank used it to decide which films toshow them. It also offered to set up instantmessaging chats with agents to clear up anyquestions not covered in the films.

Skeptics dismissed the new approach asnaïve, complaining that screen space, agentsand creative agency resources were being di-verted away from “hard” sales messages to-ward “soft” educational content. Early salesresults were disappointing, but the bankstayed the course. A year after launch—whenthe creative teams had improved the educa-tional films and analytics had revealedwhich sequences of films to show to whichcustomers—the strategy had tripled salesrates on the Web channel.

In more mature markets, analytics havebeen used to develop new service offerings.Knowing that automated bill payment is adriver of account longevity, one U.S. bank isbuilding a mobile platform that takes advan-tage of the cameras in smartphones. All cus-tomers have to do is take a photo of almostany recurring bill using the bank’s mobileapp, and the app will set up an automated

9Recapturing banking’s analytics edge

Tailoring the online customer experience

How can analytics innovation translate into a more compelling—and

profitable—customer experience?

Take Joe, a long-term user of online banking. He regularly logs in

to his bank’s Web site to make payments, check balances or trans-

fer funds between accounts. Over a few months, Joe notices new

content on his bank’s site, but not wanting any new financial prod-

ucts, he ignores most of them. Eventually a message catches his

eye, prompting him to book a holiday using his air miles before

they expire.

A couple of weeks later, he clicks on a banner offering a report on his

credit history. Not long after that, he clicks on another that takes him

to an analysis of his credit-card spending over the past 15 months.

He gradually gets used to lingering online a little longer to scan new

messages from the bank.

Later that year, when he logs in to pay for his daughter’s music les-

sons, he sees a banner suggesting he put his Christmas bonus into a

higher-yield money-market fund rather than leave it in his savings ac-

count as he did last year. It takes him only a couple of clicks to do so.

Those two clicks for Joe translate into a doubling of the value of this

customer relationship for the bank. As part of a systematic program

to make the most of its Web-servicing platform, the bank has de-

signed unobtrusive tailored journeys for every customer, based on the

preferences each user reveals in clicking on or ignoring different

messages at different times. The algorithm that determines which set

of messages each customer sees runs into hundreds of thousands of

business rules, and would have been too complex to manage even

10 years ago. The program has enabled the bank to turn around its

reputation as a digital laggard and generate a 10-fold increase in on-

line sales conversions.

payment. Using this type of seamless inte-gration of data and technology to capitalizeon analytics insights across organizationalboundaries is what will drive banks’ growthand competitiveness over the next few years.

* * *

The banking industry is capable of creatingoutstanding customer experiences while de-livering step-change improvements in prof-its. However, this is no easy journey, and itcan easily be thrown off track by the questfor a secret sauce—that elusive but all-pow-erful transformative algorithm. To capturepragmatic real-world improvements, banks

need to work out how to use analytics inno-vation to enhance the customer experienceand mobilize functions and units across theorganization.

Although expertise in this growing field isscarce, leadership and organization arescarcer still. Leading and organizing to cap-ture the promise of advanced analytics needsto be front and center of the agenda for bankCEOs.

Prithvi Chandrasekhar is a senior expert in the

Bangalore office, and Robert Mau is a knowledge

expert in the New York office.

10 McKinsey on Payments September 2014