A business-back approach to consumption 9

A business-back approach to technology consumption

The business of payments hinges on processing and storing large volumes of transaction data. To perform these activities well or at all, payments companies can spend as much as 10 percent of their total revenue on technology infrastructure. Reducing these costs even by a fraction of a percent can release millions of dollars that can be reinvested in growth and innovation or applied to the bottom line.

David Chubak In most payments organizations, business The way to right-size consumption is to en- demand drives technology costs. “Right-siz- gage senior business and application devel- James Kaplan ing” this demand to the technology supply opment leaders in taking a “business-back” Catharine Kelly can be a significant challenge: the economics approach to infrastructure consumption and of infrastructure investments are not always cost. This approach begins with an evalua- obvious; costs are not transparent; and it is tion of business technology needs: for in- difficult to tie the business decisions that stance, how long data must be stored, how create demand to the cost of the technology soon an application or process must be back that satisfies it. online after a failure. It concludes with the implementation of significant governance Consequently, many global payments or- and organizational changes. ganizations face a demand for technology infrastructure that outpaces the growth of In many institutions, these changes include their business. These organizations have the creation of a standardized catalogue of gone to great lengths to reduce their tech- technology products and services that nology operating costs by re- prompts the business to pay for the technol- processes, shifting to global sourcing mod- ogy it uses under the guidance of a strong els and leveraging virtualized . governance model. At the same time, new or To find still more savings, they must evalu- revised business policies set targets for con- ate how their business consumes technol- sumption and optimization and rationalize ogy and find a better way to match supply the use of infrastructure across the enter- to demand. prise. An organization’s information tech- 10 McKinsey on Payments June 2011

nology (IT) department takes on a market- • 1: Adjusting end-user entitle- making function, complete with product and ments to match technology supply to relationship managers who drive service user demand. adoption and bolster the role of IT in tech- nology investment decisions. • Lever 2: Technical optimization to ad- dress the typical inefficiencies that drive When effectively implemented, the business- up technology costs at the application back approach can reduce technology infra- level, such as the over-provisioning of structure costs by as much as 20 percent. storage, processing or network capacity; Moreover, the effort makes available addi- overstated application performance tional capacity and funding that can be ap- needs and growth forecasts; and exces- plied toward new products and services. sive disaster recovery for non-essential The business-back transformation: applications. for change For a business-back approach to take hold, • Lever 3: Business process and policy the way in which the business consumes in- changes to curb wasteful business prac- frastructure must structurally change. Four tices such as retaining data longer than distinct levers can reduce technology de- necessary, overusing hardware, mis- mand in a business-back way: aligning storage media with back-up

Exhibit 1 Excess minutes analysis (excluding pooled plans) Wireless plan usage Number of devices with unused plan minutes optimization – example analysis Unused plan Monthly plan minutes minutes Flat < 450 450 451 - 900 901 - 1350 1350+ Unlimited Total

Over 10 10 100 20 10 0 150

0 - 100 10 30 50 10 5 0 105

101 - 250 0 20 10 30 5 5 70

251 - 450 0 20 850 50 10 5 935

451 - 900 0 400 50 10 460

901 - 1,350 0 80 10 90

1,351 - 2,000 0 60 60

Over 2,000 0 20 20

Unlimited 0 15 15

Total 20 80 1,010 510 160 110 15 2055

Potential to eliminate Potential to downgrade 42% of the devices (870) have low plan minutes 31% of the devices (630) have high plan minutes with greater than 50% of those minutes unused – with greater than 50% of those minutes unused – these should be reviewed for possible elimination these should be reviewed for possible reduction Source: McKinsey analysis of sample data A business-back approach to technology consumption 11

and retrieval needs, and maintaining obsolete files can yield both capacity and sav- unused applications or unnecessary ings, while also mitigating risk. functionality In another example, some segments of wireless plan users consume far fewer min- • Lever 4: A new governance model to en- utes than their plan allows, opening the op- sure complete cost transparency and ef- portunity to reduce or eliminate those fective implementation of the minutes (Exhibit 1). business-back approach. End-user entitlements govern supply and Lever 1: Adjusting end-user entitlements thus can be allocated among various user Employees are often provided with more segments based on type and level of use (Ex- technology or capacity than they actually use. hibit 2). The new entitlement plan may as- End-user entitlements such as network stor- sign a maximum number of minutes or a age space and wireless plans can be scaled maximum reimbursement amount to certain back to correspond with real, rather than user segments. For example, only executive perceived, demand. As an example, employ- users who spend more than 75 percent of ees may consume excessive network storage their time traveling would be allowed the space by storing files that have not been used maximum 1,500 minutes or the maximum in years. Purging users’ network storage of reimbursement of $150.

Exhibit 2 Qualification Reimbursement levels Mobile phone requirements Estimated usage can be Example Maximum Maximum number of managed by User level requirements minutes reimbursement users Comments segmenting users Executives Significant Executives 1,500 $150 TBD Significant (manager and above) user >75% travel users by executive approval only

On-call personnel High-end 50-70% 900 $150 TBD (e.g., emergency user travel responders, personnel with responsibility for services which require 24-hour access)

Frequent travelers Mid-level 25-50% 600 $150 TBD (e.g., above 25% user travel travel)

Other personnel Low-level <25% 300 $150 TBD Demonstrated user travel business need (business case required) and executive sponsorship Source: McKinsey analysis 12 McKinsey on Payments June 2011

Wireless entitlement policies can then be based on a set of criteria such as data stor- amended to manage how wireless minutes age needs, data access requirements, level are used, to assign responsibility for control- of analytics, database functionality and ling wireless usage, to establish procedures mission-criticality (Exhibit 3). One finan- for managing exceptions and to mandate cial institution saved 5 percent in storage consequences for non-compliance. costs by rationalizing and reclassifying its storage supply. Lever 2: Technical optimization The utilization and unit cost of production Opportunities for technical optimization servers, meanwhile, can be improved abound in most IT shops. Storage provision- through virtualization and the removal of ing, production server utilization and disas- unneeded capacity based on peak and aver- ter recovery provisioning are all potential age utilization rates. For disaster recovery candidates for optimization. provisioning, costly real-time, synchronized Storage provisioning, for example, can be data replication should be reserved only for adjusted by reserving the most robust and mission-critical applications that cannot be therefore most expensive storage media ex- permitted to sustain any data loss or down- clusively for the most critical and analytics- time. Savings can then be realized by mov- intensive applications. Applications can be ing all other applications to cheaper assigned to various tiers of storage media replication options.

Exhibit 3 Storage Environments permitted Applying a decision service catalog Development Test/QA Performance Production tuning tree to standardize Yes storage demand Platinum SAN $8x/GB*

Mission- No critical? Gold SAN $4x/GB* Yes

Database No application? Gold NAS $3x/GB* Yes No Intensive analytic Silver NAS application? $2x/GB* Yes

Require No Static storage instantaneous data access? (Nearline) $1x/GB* Yes

New storage Is there data No Direct attached to be stored? storage demand $1x/GB* Source: McKinsey analysis A business-back approach to technology consumption 13

Lever 3: Business process and policy ther tailored to demand if a standardized, changes repeatable process for capacity forecasting To develop business processes and policies is in place, as uncertain capacity forecast- that support the business-back approach, ing and long procurement cycles often lead the payments organization should analyze to the over-provisioning of storage by 30 to technology consumption and cost to dis- 50 percent. cover the particular business rules and func- The consumption and cost analyses will help tionalities that generate end-to-end application teams understand how every infrastructure costs. Costs may have to be click, page view, transaction and data ele- manually allocated to applications using an ment retained requires processing and stor- IT must take on a age. Without that knowledge, the business will always opt to give the customer more – “market-making” function that more availability, more uptime, more real- drives service adoption and time data, more views and greater transac- tion capabilities – even when customers are bolsters the department’s role in not demanding those capabilities. A cus- the investment dialogue. tomer-facing business unit, for instance, may provide customers with access to 24 months imprecise methodology closely akin to vari- of account history with one click – even able costing, but the importance of linking though the vast majority of customers only costs to business decisions trumps precision require 2 months of history. In one such situ- in this situation. The analysis should reveal ation, a financial institution reduced its stor- that particular business decisions can lead to age costs by over 25 percent by segmenting infrastructure costs’ outpacing business its customers and designing customer-facing growth. The IT and business sides of the or- storage policies around unique client needs. ganization can then work together to de- velop policies that revise costly business Lever 4: A new governance model requirements and rationalize application Key to the success of the business-back ap- functionality, data retention and disaster re- proach is a strong governance model that covery against real value to customers and enables informed risk-cost trade-offs and as- the business. signs accountability for implementation and As an example, revising the policies and results. The payments organization must processes related to storage can yield 5 to designate roles and responsibilities for set- 10 percent in savings when coupled with ting and enforcing policies, and it must de- the tiered storage approach described vise ways to measure and report above. Since only 5 to 10 percent of users consumption to promote transparency. The typically use 80 to 90 percent of allocated organization could, for example, build and storage space, implementing usage cap and maintain an inventory of applications and retention period policies can free up signif- assets with asset-to-application mapping icant capacity. Storage supply can be fur- that would enable productivity tracking. 14 McKinsey on Payments June 2011

Exhibit 4 Step 1 Step 2 Step 3 Step 4 Step 5 Step 6 Business-back Assemble costs Allocate the IT Organize Based upon a set Develop the Develop approach to by type of IT costs to business applications and of criteria, business cases business reducing IT spend applications their associated identify hot spots for each hot spot policies for costs by business of consumption, to drive material governing the consumption domain and including current impact in consumption of product spend and reducing IT run infrastructure growth rate costs Use to compare IT compared to tier run costs with 2 drivers business metrics such as revenues, accounts and transactions

Costs by Costs by Costs by List of hot Implementable Business IT domain IT application business spots for business cases policies for domain business change app teams

Mainframe Products Hotspot Driver Storage App 1 # Business Case 1 1. Target utilization Distributed 1 Business rules 2. Tiering 2 3. Disaster recovery Network 2 Member data 4. Back-ups End User 3 3 Data updates Processing Business functions etc. 1. Virtualization 2. Target utilization etc.

Source: McKinsey analysis

An essential in the new governance department’s role in the investment dialogue. model is a catalogue of standardized tech- IT now claims ownership of the complete nology products and services. The business product lifecycle and manages relationships can use the catalogue to select among stan- just like any other line of business. IT’s prod- dard applications and packages at a fixed uct managers design and build reusable serv- price, while customized solutions would ices that align with business needs, and IT’s come at a premium. Over time, use of the relationship managers translate business de- catalogue would foster uniformity in the mand into specific services to be ordered. business’s technology environment and re- The payments organization must thus estab- duce the number and level of resources lish a relationship management structure needed to manage that environment. The that enables and incentive process of buying infrastructure would be- alignment among infrastructure compo- come more consistent, with appropriate nents, application architects and business technology specifications assigned to specific owners. IT will have to equip itself with new business requirements. skills in areas such as expense and product As part of the new governance model, IT management, which will necessitate building must take on a “market-making” function current employees’ skills as well as recruiting that drives service adoption and bolsters the new talent. A business-back approach to technology consumption 15

Business-back in six steps and to guide the business in assigning in- The business-back approach to reducing IT frastructure to support business needs. consumption can be implemented in six se- * * * quential steps (Exhibit 4). Implementing the business-back approach • Step 1: Gather and categorize costs by to IT consumption is organizationally chal- type of technology expense: mainframe lenging and time-consuming, but the poten- expense, distributed expense, network ex- tial cost savings and business benefits make pense, etc. it worth the effort. Organizations that suc- • Step 2: Allocate these costs to the various cessfully manage technology consumption business applications. using a business-back approach can expect to reduce their IT run costs by 10 to 25 per- • Step 3: Organize these applications and cent, with potential to realize a savings of 5 their associated costs by line of business or to 10 percent in the first year. product, and then evaluate the cost of run- ning the applications against business per- Moreover, the business-back approach can formance metrics (e.g., revenues, number inspire product innovation. Through the of accounts or number of transactions). business-back exercise, product developers may discover data and capabilities in their • Step 4: Based on a predetermined set of infrastructure that they never knew existed. criteria, pinpoint “hot spots” where Such discoveries, along with the funds saved spending substantially exceeds return. Ex- and capacity gained from eliminating super- amine the drivers behind those hot spots, fluous functionality, may give rise to a host such as business rules and data updates. of new customer services. • Step 5: Develop a business case for reduc- ing consumption and run-costs in each hot spot. David Chubak is a principal and James Kaplan is an expert principal, both in the New York office. • Step 6: Formulate a set of business poli- Catharine Kelly is an associate principal in the cies to govern technology consumption Atlanta office.