CASE
General Electric's supply chain is not simply enormous. It's a Byzantine web of sourcing partners, touching all corners of the globe: 500,000 suppliers in more than 100 countries that cut across 14 different languages. Each year, GE spends some $55 billion among its vast supplier base. Long-time GE CIO Gary Reiner knows this problem all too well, since, among his other duties, he is responsible for how the $173 billion conglomerate spends that $55 billion, utilizing GE's Six Sigma practices and taking advantage of its hefty purchasing power. GE, for instance, buys $150 million in desktops and laptops each year from a single supplier, Dell-"at a very low price," says Reiner. For years, GE's Global Procurement Group faced a challenging reality:
trying to accurately track and make sense of all of the supply chain interactions with half a million suppliers- contracts, compliance initiatives, certifications, and other critical data, which needed to be centrally stored, managed, and made accessible to thousands across the globe. GE was using what it called a Global Supplier Library, a homegrown system that, Reiner says, had a "rudimentary capability." Reiner and his staff knew that GE needed something better, but they didn't want to build it. They wanted a supplier information system that was easy to use and install, could unite GE's sourcing empire into one central repository, had multi language capabilities, and also offered self-service functionality so that each of its suppliers could manage its own data.
The destination was obvious: To achieve one common view of its supplier base, and one version of the truth in all that data, a goal that torments nearly every company today. But to get there, Reiner and his IT and procurement teams took a different route. In 2008, GE bought the application of a little-known Software-as-a-Service (SaaS) vendor that would ultimately become the largest SaaS deployment to date. "When we judge a solution, we are indifferent to whether it's hosted by a supplier or by us," Reiner says. "We look for the functionality of the solution and at the price." And that, he claims, has been the way they've always operated. Reiner says that his group doesn't see a big difference in cost and in capabilities between on-premise and SaaS products. "And let me emphasize," he adds, "we don't see a big difference in cost either from the point of view of the ongoing operating costs, or the transition costs."
Furthermore, when looking at implementation costs, "they're largely around interfacing with existing systems, process changes and data cleansing," he says. "Those three costs exist regardless of whether GE hosts that application or whether the supplier hosts that application." The Aravo technology platform was untested at GE's level of requirements and, with just 20 or so customers, coupled with the sheer scale of GE's needs, did not really concern Reiner. "We could have been concerned about that," he concedes. "But that would have also been a concern if we had hosted the software on our own servers.
We knew Aravo could handle it." Plus, Reiner says that no other supply chain vendor offered the type of functionality that Aravo's Supplier Information Management (SIM) product offered, and Reiner and his team reasoned that it was much cheaper to buy than build. "We'd much rather work with them," he says, "than build it on our own." One GE sourcing manager told Aravo that GE's ROI on the project is not just positive, "it's massively positive." "They're using SaaS for 100,000 users and 500,000 suppliers in six languages: that's a major technology deployment shift," says Mickey North Rizza, research director at AMR Research. She says that the sheer volume of transactions, combined with the fact that GE supply chain and procurement employees around the world can now access the same sourcing partner information, all from the same central spot, is significant not only for the supply chain management space but also for the SaaS and cloud computing world. "Finally we have a very large company tackling the data transparency issue by using a SaaS product," North Rizza says. "It's a huge deal."
So far, the thorny issue of data quality in GE's supplier data has been improved, because suppliers now use the self-service capabilities in the SaaS system to manage their own data. GE has 327,000 employees worldwide, and its sourcing systems have more than 100,000 users. There is still more work to do to the SIM platform-for example, GE sourcing employees will add more workflows and new queries to the system; more languages might be added as well (six are operational now). Reiner says that GE is committed to working with Aravo for the long term and that the system has performed well so far. And SaaS, as an application delivery mechanism, appears to have a bright future at GE. When Steven John took over as CIO at specialty chemical manufacturer H.B. Fuller Co., he inherited a North American payroll system implementation that was expensive and going nowhere.
The business units hadn't participated in the technology decision, and the project was bogged down with customization issues and other concerns. John chose to relinquish control of payroll software and switched to SaaS. "I wanted to do an implementation that was simple and straightforward-to configure but not customize-and see the benefits of a standard, global platform," John says. "This was a way to teach, save money and outsource a noncore system." Giving up control was an easy trade-off compared with the headaches he would face trying to fix the existing software. "You're getting a lot more innovation," says Ray Wang, an analyst at Forrester Research Inc. "The products are a lot more configurable than what most people have in their own applications. You can change fields, rename things, and move attributes and workflows. So there's a good level of control there."
What's more, the configuration choices are more refined and well thought out, giving users a few good choices instead of myriad options. John found that configuration rather than customization allows H.B. Fuller to maintain its "lean core." "I believe that more standardization leads to more agility," John says. "SaaS allows us to say, ‘This is good enough . . . for what we need.' So you don't end up with these horrible situations where you have these highly customized systems. We go with configuration option A, B or C. If one of those three doesn't meet our need, we can try to influence the next release. But in most cases, A, B or C is going to meet the need." At H.B. Fuller, the move to SaaS for human resources tools allowed the company to empower its people. "I can do a reorganization and have it reflected within minutes, and I don't have to call someone in HR to update everything," John says. "I can also pull up other people's organization charts and see where they are and what they're doing and better understand the organization."
When it comes to managing SaaS, neither the IT department nor the business unit using the software should be eager to relinquish control. "The buying decisions are shifting from IT to the business leaders," who often opt to charge the software as an expense rather than wait for approval through the capital budget committee, Wang says. Still, he adds, "it's very important to engage IT in these SaaS decisions because there are overall IT architectures and blueprints to consider." It becomes very costly when applications don't integrate or interoperate well with one another. "It's good to at least have some parameters and policies in place so that people understand what type of apps will work better within the environment, what will be cheaper to share information and data with," says Wang. One of the problems with SaaS is that if your vendor were to go bankrupt, everything would shut down. You don't own the software. It's on lease. The question is, what do you own? If the vendor doesn't have a separate on-premises deployment option, "you need the ability to take out transactional data, master file information, any kind of migration programs, just in case, so you can convert it to an on premises alternative if they were to go down," Wang says. In the long term, Wang envisions an IT culture where software as a service is commonplace. "We may live in a world where everything is provisioned. All our applications don't stay on premises, and business leaders are out procuring applications," he says. "IT teams are testing them to make sure they work well in the environment and there are no bugs or viruses and things integrate well, and basically the IT staff will spend a lot of time provisioning services and implementing, integrating, doing installs. That's where we envision the market in 2020."
CASE STUDY QUESTIONS
1. What factors should companies take into consideration when making the decision between developing their own applications, purchasing them from a vendor, or taking the SaaS route, as discussed here? Make a list of factors and discuss their importance to this decision.
2. What risks did GE take on when they contracted with a small and less experienced vendor? What contingencies could have been put in place to prevent any problems from arising? Provide several examples.
3. What should companies do if none of the "configuration options" perfectly fits with their needs? Should they attempt to customize, or select the least-worst alternative? When would they do each?