How Freddie Mac IT Used Data to Change the Dynamic with the Business

“Since implementing Apptio, communication between IT and the business units is steadily improving. More and more, everyone is on the same page, communicating with the same language regarding IT usage, rates, and levels of performance and reliability.”

Executive Summary

To communicate cost and value to its business units, Freddie Mac uses TBM to translate IT component and resource costs into service costs the
business values and choices they can make. The business was so pleased to be brought “out of the darkness and into the light” with the consumption based showbacks that they led the charge to replace the company’s fixed IT cost allocations with a consumption-based cost model.

Freddie Mac Corporate Overview

The Federal Home Loan Mortgage Corporation (FHLMC), known as Freddie Mac, buys mortgages on the secondary mortgage market and sells them as mortgage-backed securities on the open market to fulfill their mission of helping to create a stable and affordable U.S. housing market.

The Challenge

Freddie Mac has always invested in technology innovation. For example, it was the first to automate the underwriting process, allowing lenders and borrowers to find out in seconds if they are pre-approved for a loan. Within the business, IT costs were seen as unavoidable. IT costs were charged back to BUs (business units) based on a complex collection of fixed percentage allocations and cross-charges accumulated over time in spreadsheets. Costs appeared on BUs’ Profit and Loss statements (P&Ls) in terms of accounts and cost centers, in technology terms that were difficult to decipher within the business. The business knew IT for its support of their laptops, phones and printers. Back-end services like disaster recovery were “out of sight, out of mind — a magical free thing they get from IT,” says Kevin Brown, IT Assets Manager at Freddie Mac. “IT was a black box. We needed transparency to lift the veil,” explains Brown.

Building a Service-Based Showback Model

Several years ago Freddie Mac documented their services in an IT service catalog. The services fell into three categories: workplace services (e.g. workstation, mobile phone), application services (the business applications recognized by the business) and application hosting services (e.g. infrastructure and support services).

Brown wanted to publish standard rates (pricing) for each service together with how many units the business was consuming so the business could see a clear, predictable relationship between their consumption and IT costs. “We wanted to pull back the veil on the black box,” Brown says.

His team set about modeling its IT service costs to determine what rates to charge. “Our first cost model was all done with spreadsheets. We had seven to ten people working on it.”

Looking for a more flexible and scalable approach, Brown discovered that technology organizations from across different industries had developed a new methodology called technology business management (TBM) with standards and software for managing technology costs and other business aspects of IT.

“I’d been working at this for 25 years, figuring out most of it for myself. Finally, I found a group of people like me from a lot of companies and industries, building methodology and standards and sharing best practices,” Brown remembers. Among those best practices is the ATUM taxonomy for cost and KPI modeling. “I show people the poster of the ATUM model to show them that this isn’t some arbitrary thing we’re making up, we’re following IT best practices. I have an industry behind me.”

Freddie’s IT cost transparency team began configuring its own TBM system to automate the monthly process of turning raw operational data into IT consumption and cost reports for each BU.

Each BU’s workplace service costs are based on a published fixed rate (price) times the volume of units shown by the data that month (e.g. number of workstations).

Each BU’s application and application hosting services are a function of three factors:

  • Applications the BU uses (shared applications are apportioned by a previously agreed fixed percentage)
  • Price-times-quantity for each hosting service consumed by its applications, such as the number of servers, GB of storage or hours of support
  • An added “complexity” cost, which is a function of the total number of all “instances” attached to an application per the CMDB including servers, operating systems, databases

Determining BU costs by multiplying fixed prices times monthly consumption, rather than actual costs from the general ledger, Brown’s team is nevertheless able to land within 1-2% of actual annual costs. This residual is allocated to BUs based on the number of applications they own. Prices are adjusted every year and as services are added and changed.

With an automated approach, the cost transparency team is able to makes frequent tweaks to data and rules “We’re able to make changes almost on the fly. So when we go out and do demos and we’re talking to business lines, and they provide us simple little changes, I can go back to my team and introduce those changes almost instantaneously. For example, we just got access to better data around our disaster recovery service, so we configured the data rules, model and reports and rolled it out to the business inside two weeks.”

Working Through Data Quality

When it came time to show the business what they had been working on, the cost transparency team was apprehensive about the completeness, relevance and quality of operational data they had to work with. “Getting available data to support our model had always been something of a struggle,” he says. But rather than wait for data quality to be perfect before rolling out reports to the business, he showed them the progress he had made used reports to enlist help from both the business and IT to improve the data.

“Data improves as people begin using it,” says Stacy Shifflett, director of business insights at Freddie. “You get to see the synergy when you start showing it and people at first are skeptical, and then they start thinking it’s useful and they say, ‘Oh, well. If you can capture this, can you capture that?’”

For example, in one of their first demos to the CFO of the Multi-family business and his IT business technology officer (BTO), Brown explained that he didn’t have the data to show his actual consumption so for now this was just based on a static formula. The BTO said that this wouldn’t do him any good, he needed to know actual consumption. The BTO applied pressure behind the scenes. Within three days, Brown had a call from his infrastructure department explaining that they were beginning a project to map applications, storage and compute.

Focusing all this attention on data has been incredibly successful at spotlighting the inefficiencies in the application portfolio data. Using TBM’s data quality analysis, the cost transparency team revealed the impact of unassigned servers to their cost model. This helped the service management team implement process to more reliably capture server-toapplication relationships in the CMDB.

Changing the Conversation

While the old chargebacks continued, the cost transparency team rolled out the new reporting to give the business a clear picture of how much of each service they were consuming and at what total cost. Brown says they started rollout meetings by making it clear they intended to partner with the business, saying “We need your help in controlling costs,” sometimes framing it around where Freddie’s average unit costs were higher than industry average benchmarks.

Using reports from the TBM system, the cost transparency team engaged the business in conversations about opportunities to reduce their consumption.

There were many seemingly small victories that added up to a perceptible shift in business attitude, from victims of IT taxes to masters of their own destiny. Brown recalls, “We showed a BU the cost for services they didn’t even know they were using. They determine they didn’t really need them, so they stopped using them. The next cycle of reporting showed an overall decrease in their costs.”

Seeing cost and utilization together, one business unit relinquished a significant amount of storage allocation after seeing the amount that was unused by its applications. Another BU decided to remove more than half of the development and test environments for its accounting application, saying, “Well, I guess they’re not free anymore!”

Up-Leveling to Business Capabilities

In addition to managing business demand for IT, the cost transparency team has been better able to communicate value and shape investment by mapping application costs to business capabilities. Some are industry-specific specific parts of the business such as onboarding mortgages, issuing securities and quality control. Others are shared such as accounting and HR.

With help from the architecture team the cost transparency team used fixed percentages to map applications to business capabilities. Some applications are dedicated to a capability, some are evenly shared while others are split, say 90% / 10%, based on input from architects and the business unit leaders themselves.

To engage the business in improving the capabilities model there is also a category called “unallocated.” “When we started reporting that, our business partners in architecture saw that and said, ‘Well, wait a minute. Why do we have to apply your data? It’s missing some information.’” This awareness made it easier to go back and get the data they wanted.

Today, each business unit can click through a dashboard starting with how much it costs for IT to support the business capabilities they use. Within any business capability they can drill down to see their portion of costs for each application that supports that capability, as well as workplace services. A click on an application takes them into their cost and consumption of each hosting service.

The mapping to business capabilities is already helping to shift the conversation to value. “One BU executive is excited because he can use the TBM data as a defensible, documented part of business case for innovating how they are delivering a business service to the market,” Brown recounts.

“The success that we’ve had introducing business capabilities is truly a translation of IT speak into business speak,” he adds.

From Showback to Chargeback

During the first two years of its TBM journey, Freddie provided the business with “showbacks” on service consumption costs while still using the old headcount-based model to allocate expenses to BU P&Ls. Now it was the business leading the charge to make the switch.

After seeing IT costs translated to business capabilities some began to see the potential in TBM to help make their businesses more competitive.

Chargebacks on monthly service consumption calculated by the TBM system began in August. In addition to continuous refinement of that model, TBM roadmap includes using the TBM system’s benchmarking and planning capabilities.

Brown is quick to point out that Freddie is still at the relative beginning of its TBM journey. “We’re probably at the awkward toddler walking stage,” he admits. Even those awkward steps have their rewards.

With the TBM model calculating costs that can ultimately affect BU bonuses, Brown was prepared for pushback, especially from those that would pay more than they did under the old model.

He was pleasantly surprised, recounting “One business unit leader recently told me, ‘You’ve taken us from darkness and into the light!’ And that was from a BU that will wind up paying significantly more under the new approach.”

Advice for Those Starting Their TBM Journeys

The Freddie Mac team offer this advice for those beginning or considering their own TBM journey.

Shape conversations: “Don’t provide data just because you can get it,” Shifflett advises. “Shape reports around topics that generate the right conversations and decisions, such as ‘Can we retire this application?’ and ‘Should we grow the business in this direction?’”

Start small: You can’t take it on all at once,” Shifflett adds, “so don’t be so hard on yourself in the beginning. Start with whatever data you have, and chart a course for where you want to go. You get more and better data once people start seeing what you have in reports.”

Data, data, data: “It’s all about the data,” says Brown emphatically. “It’s okay that your data’s not perfect upfront. Prioritize what data elements you want to improve. Make that into a roadmap.”

Additional Resources