With an estimated 300 solutions on the market that automate or facilitate some link of the supply chain, the most important assessment that can be conducted is whether to ASP, buy or make.

[From iSource Business, June 2001] Even though its industry niche is about three years old now, it's still a little hard to explain what exactly Atlanta-based Services Resource Inc. (SRI) is without resorting to corporate-speak.

So let's try this: Imagine a facility management company crossed with an energy e-procurement service provider, and you can come close.

Explaining what SRI does is a little easier. Its clients tend to be large-scale businesses or retailers with multiple sites around the country. SRI handles many aspects of building management for these companies, including HVAC and building controls, energy management, procurement or facility suppliers, and repair and maintenance. It also manages some $1.8 billion annually in energy for these clients, a service that not only includes ensuring that customers get the best rate but actually pays the utility bills for them, using the consumer bill-paying services that have sprung up on the Internet.

It was the data contained in these individual bills -- which number in the hundreds of thousands -- that got SRI staffers thinking. "We wanted to do more things with the pre-bill and post-bill information," says Leanne Wilborn, senior manager of information systems. "More sophisticated analysis for our clients, future trends analysis and aggregation of loads for energy procurement, for example. Eventually, we would like to be able to aggregate our loads so if one client wants to procure 10 megawatts of electricity at a certain price we can help arrange that."

With the current volume of information technology (IT) projects at SRI, it was impossible for them to offer these services. "We had individual sets of information, but in order to look at it collectively and get into data mining we had to get all the data -- historical and current -- into a usable format," Wilborn says. "Then we needed to have all of this data in one, central space."

Clearly, a new system was called for. The only question became, "Which one?" Ordinarily, this issue is hard enough for most companies to address  a system selection can take months, if not longer, depending on the size and complexity of the company. SRI had the additional problem of being a leader in a new whole industry niche, which is a problem that, ironically enough, quickly solved the dilemma of what system to choose: there were no systems on the market that could be tailored to fit their industry space.

Build Versus Buy

In a way, SRI may well have gotten off easy. For most companies, the problem has become the abundance of choices. There are probably some 300 or so supply chain solutions on the market that automate or facilitate some link or links in the chain. Then there has been the rise of the application service provider (ASP) model in which the solution is hosted and managed off-site by a third party. And, finally, there is a whole mishmash of B2B Web sites, which offer services ranging from procurement handling to the provision of transport management solutions. Suddenly, the possible permutations a supply chain can undergo have become endless. For companies that are short on time and research staff, this is not necessarily a good thing.

Indeed, supply chains have become very complex creatures. Consider, for example, the state of the multi-billion-dollar supply chain at Eastman Chemical Co. There are at least two enterprise resource planning (ERP) systems, one advanced planning system, electronic data interchange software (EDI), and links to the planning and order systems of Eastman's collaborative partners. "New Internet technology has allowed us to explore all sorts of new models," says Bill Graham, integrated direct program lead for Eastman.

But despite the proliferation of technology, the classic methodologies for constructing a winning supply chain do still apply. Ultimately what it boils down to, says Steve Jones, executive director of Trifolium, a Raleigh, N.C.-based enterprise application integration and e-commerce provider, are three options.

Option 1: A company can buy a product that fits its needs as closely as possible and then adapt its business around the purchased product. Products in this category can be deployed relatively quickly but tend not to integrate well with existing systems.

  • The advantages: relatively fast, cheap and predictable as to the length of time it will take to install.

  • The disadvantages: It will be more difficult to add functionality, more difficult to support separate business units, it provides little support for sharing data or transactions with other systems, and it is very dependent on the product-release decisions and cycles of the supplier to obtain new features.

Option 2: A company can buy a packaged solution that can be configured to the needs of the purchasing organization.

  • The advantages: This class of products (which includes offerings from Siebel, SAP and Agresso) allows customer-specific business rules and customer views to be implemented. Often the supplier will be able to suggest changes to the customer's business based on best practices found elsewhere in the industry or included in their product.

  • The disadvantages: Due to the product's high degree of flexibility, a consultant is usually required to install, configure and deploy the product, increasing the cost from three to 10 times the initial cost of the software. If managed poorly, the deployment can disrupt business. Even in the very best of circumstances, these solutions can tend toward the unpredictable in terms of schedule and cost. New features can be costly and difficult to obtain quickly, since they are dependent on the product plans of the supplier.

Option 3: A company can build a custom application or revamp an existing application. Usually this requires customers to apply their own resources and expertise. However, consultants can be used to guide the design and implementation without drawing resources away from the company's day-to-day business activities.

  • The advantages: There are products on the market that can provide a platform for new applications, allowing custom applications to be rapidly deployed. Capabilities specific to the customer can then be added onto the basic application framework. This type of solution can be manipulated to evolve rapidly, offering new capabilities as the business requires and providing complete control of the application to the buyer. This approach is often cheaper and more flexible than option one or two.

  • The disadvantages: This choice is only recommended if the customer is confident that he or she can build a competitive system, since they are not tapping the specific expertise of the software package suppliers. Then, there is option four: the application service provider (ASP), the infotech industry's latest sleeper offering.

The Case for ASPs

For many companies, ASPs are no-brainers. The most recent software applications are cheaper on a per-month basis than a license; and for smaller and mid-sized companies they're a dream come true. Even for larger companies, ASPs have become an attractive option. "For bigger companies laden with ERPs and legacy systems, the last thing they want to do is integrate one more application into their internal systems," says Steve Gold, managing director with KPMG Consulting.

However, the main reason ASPs are becoming so popular, says John Gonsalves, vice president of Adventis, a New York-based IT consultancy, "is that it means companies don't have to keep up with technology. Using them is like using a utility, and there is the added benefit that you don't have to jump through hoops as you might with a supplier, who sometimes can be very arrogant."

"It all depends on how companies think of their technology portfolio," Gold says.

Ron Shabon, vice president of management information systems (MIS) at Dress Barn, was drawn to the ASP model because he didn't know how quickly or how much the New York retailer's e-commerce operations would grow. "With 700 locations, we are reasonably big, but we have a small MIS staff with a lot of projects on our plate," he says. "We decided we wouldn't have time to build and manage our own system."

Eventually, Shabon honed in on NaviSite, a provider of outsourced Web hosting and application services, because it was economical and flexible, he says. "Initially, we rented equipment from them, but eventually we bought our own."

Of course, there are drawbacks as well. Pricing models vary, making it hard to compare one ASP with another. Companies often fail to calculate the total cost of ownership, overlooking some small, but expensive, detail. Security was a concern, at first, for some companies, but that seems to have been alleviated. Ownership of the technology, not to mention data, has also been a worry. Then there is the issue of catalog content: some suppliers make the user manage that while others include it in the service.

Another tradeoff that users have to make with ASPs "is the surrender of the ability to perform certain customization," says Kit Manning, vice president of online services for OfficeTempo, a Boston-based dot-com that aggregates office supplies and services for small businesses. And actually, it's this issue, not price or security, which tends to drive would-be ASP users to their next option: customizing off-the-shelf software.

When to Buy?

This is by far the most popular option, for a number of reasons. "If you go with an off-the-shelf solution, you are leveraging many years of experience that a company has built up," says James Towsend, president of Information Strategies, a Washington, D.C.-based Microsoft Solution Provider partner. "If you do it on your own, you have to take into account the extra costs and the probability that you are going outside your core competency."

Here, too, drawbacks must be considered.

"It's pretty rare that something is a 100 percent fit," admits Robert Schoenthaler, director of KPMG Consulting's supply chain solution. If a company reaches 80 percent of their system requirements, especially for functions like advanced planning or B2B integration, then it's doing well, he says. Still, "try to exhaust what you can do out of the box first," he advises.

Also, "customizing is harder than one would expect," says Robert Novick, principal with North Highland Inc., a national business- and technology-consulting firm headquartered in Atlanta. "A lot of people make the mistake of trying to force changes into the system. We recommend that you live with the system for six to 12 months before you make alterations."

It's only at this point, he says "that you're able to 'think outside the box.'"

Nonetheless, Schoenthaler says, for most companies "there is so much room for improvement, they shouldn't waste their time on such minute details. Look at the big picture instead of looking at the business progress you reap from being able to talk to your suppliers using these products."

Indeed, the majority of analysts are fairly adamant in their condemnation of a custom-built system, especially when compared with the off-the-shelf option.

"Very rarely do I see people wanting to write new systems these days," Schoenthaler says. "Most companies start with the assumption they will use an off-the-shelf product. And, with the selection that is out there, chances are you will find a pretty good fit."

But, as the SRI example illustrates, there are reasons to build.

Building As a Last Resort

Following conventional wisdom, SRI's Wilborn scoured the landscape looking for a suitable system, with the help of North Highland. "We eventually realized that the solution we were looking for doesn't exist," she says. And that, say analysts, is the main, if not only, reason to build your own system these days.

But on the other hand, companies shouldn't assume they will never have to build a system in-house. In fact, says Novick, "it's likely a company will have to build something for themselves at some point. The Net marketplaces were the most recent examples, as early adopters two years ago either had to build the platforms themselves or build the integration into the marketplaces."

Novick doesn't know for sure in which product category the next wave of innovation will occur, but wireless is a likely possibility. "What are the products out there that will integrate wireless applications into a supply chain operation? We don't know. But whoever is thinking about that next product or app is building it for themselves, because they can't buy it."

Indeed, Forrester Research recently found that some 26 percent of companies surveyed developed their own tools for internal integration and 10 percent produced homegrown tools for B2B integration. A more common scenario, says Jay Grieves, lead developer for e.magination network, a Baltimore, Md.-based integrator, is "writing custom software to tie disparate packages together. It almost never makes sense to build very small pieces. It does make sense, however, to build large systems out of small pieces, then custom write the glue that holds it all together."

A final reason a company might consider building a system is to stand out in the crowd.

"If you are using an application off the shelf and doing things exactly the same as other businesses, you are not competing at all," Grieves says. Depending on the particular industry within which your company works and who your competitors are, a specially built system makes sense.

How Do You Know?

Eventually, SRI, with North Highland's help, embarked on the formidable task of building a new system from scratch. Currently, the company is running some prototypes and raising funds for a nationwide roll-out.

The process has not been as hard as some analysts have said, but it hasn't been easy either. "Ten years ago such a project would have been more feasible," explains Erik Engstrom, director of e-business for Panja, a developer of Internet integration technology based in Richardson, Texas.

"But business systems have become much more complex and full-featured. Keeping up with the pace of packaged systems is too large a task for most companies."

For his part, Grieves advises companies considering this route to ask themselves, "Is there anything already on the market that fulfills this need? Is there similar software out there and, if so, why doesn't it fit your needs? If you want to build custom software, first you should defend it."

Surprisingly, given the complexity of the proposed system, Wilborn found justifying the project to be rather easy. "There is no question about the ROI [return on investment]," she says. Conservative estimates showed it would take less than a year for SRI to break even. Aggressive estimates project a break even point of a few months. This is partly because the price tag is not nearly as high as some would have expected. "It will cost a little over a million dollars," says Wilborn, a "soup to nuts" figure that includes not only consulting and programming, but also equipment, business intelligence software and project management. Despite the promising numbers, it's a large undertaking and Wilborn doesn't recommend it lightly. "Make sure you aren't reinventing the wheel before you make this kind of investment," she advises.
Companies in this article