It's not seamless, it's ineffective

On a conference call in the spring of 2001, Nike Chairman Phil Knight told industry analysts that he was frustrated by an underperforming supply chain management system. He blamed it in part on the company's lower-than-expected financial results. "Is this what we got for $400 million?" Knight asked.

Like his shareholders, Knight was upset, and he knew why. Because this expensive supply chain system is not linked with other key information systems of the company, the profits brought to the enterprise are greatly reduced. Without access to the vital information contained in these siloed systems, the value of supply chain projects is limited.

Achieving this free flow of business information is not just about technical connections. Let's take a look at what PHH does. In 1995, the company decided to integrate the original system and made a desperate attempt to form the "PHH Interactive System". The system gathers information from many once-separate systems to form a unified data source. As a result, tens of thousands of fleet managers at PHH client organizations have free access to a wide variety of information, including vehicle maintenance histories and trends, driver accident rates, fuel transactions, repair cost comparisons for various vehicle makes and models, and promotions information, etc. Fleet managers can use this to optimize resources and control costs, and they are very grateful to PHH instead of running to competitors.

Since the implementation of "PHH Interactive System", company expenses have been reduced by 30% and business signing has increased by 20%. It is the second largest fleet management company in the United States, after GE Capital's fleet management company. The company regards the "PHH interactive system" as a very important competitive advantage, and often sends IT personnel to demonstrate the application according to the requirements of sales calls.

Dell is also seen as one of the best examples of effectively integrated systems. Companies connect not only their internal systems, but also suppliers' systems. This integrated system, Valuechain.dell.com, brought Dell's inventory down to an incredible 4 days' inventory (compared to 30 to 50 days for competitors). Dell offers an on-demand service that provides customers with data, such as order tracking and promotions, through a website, which would not be possible without efficient integration of disparate systems. Dell became the world's No. 1 PC maker in part because its supply chain overcame inefficiencies and could interact directly with customers.

Why System Integration Fails

Examples like these are tempting, but not so easy to implement. Many companies hold back on fully integrated systems for two basic reasons: too difficult and too expensive. For example, CTO Julie Hall recently handled a major systems integration project. Her company Advance Paradigm merged with PCS Health System to create AdvancePCS. Hall inventoried hundreds of various types of system. Her integration project timeline is one year at 50% of her total IT budget.

50% of a company's IT budget is a lot, but not much for systems integration. Beth Gold-Bernstein, vice president of strategic business at integrated information portal Ebizq, estimates that a large integration project can cost between $500,000 and $1 million in software and licensing in between, while the implementation cost can be as much as twice that.

While expensive and difficult to implement, companies should prioritize system integration in order to reap the benefits of Dell's, says Bob Parker, vice president of e-commerce strategy at AMR Research. Despite this importance, many integration projects fail or are abandoned. Why?

It's tempting and easy to figure out the common causes of technology failure: hyped software, unresponsive information systems people, incompetent systems integration companies, and more. But Frances Karamouzis, a research director at Gartner, says the reason isn't just technology alone. "Technology is often ahead of integration," she said, noting that the real culprit is the failure of executive leadership. In other words, take a look at yourself. "One of the reasons many companies fail to integrate well, in addition to the technical component, is that the required changes in business processes are not done well."

Most managers mistakenly believe that systems integration is just the job of technical experts , that experts can do it in the computer room. But true integration also requires significant involvement from the business side of the company, which many companies do not do.

"Business people want the magic of integration without paying for the extra work," said Steve Morelli, CFO of StarKist Foods. This resistance is the main reason these projects fail. Once the implementation starts, the business people get scrambled and either object or give up."

"For the most part, the business people are aware of the need for integration efforts. A lot of work, but they don't understand what's really going on." Linda Reino, CIO of Universal Health Services (UHS), a hospital management company, agrees.

"Incorporating business people into integration projects is a key issue," says PHH's Talbot. "If you don't get the business side of the integration work in place, it's better to stop there."

How Systems Integration Works

Getting on the business side means managing the changes brought about by integration, which is a forward-thinking task that rarely yields immediate results, says Gartner's Karamerz . But there are other things business managers can do to give integration projects the best chance of success.

Analyzing company types starts by analyzing your company, suggests Lisa Reisman, director of Andersen Business Consulting's digital strategy and research group. According to a recent study by Reisman and fellow analyst Kim Collins, different types of companies have different integration needs. If your company doesn't covet the latest technology, it makes more sense to buy software from a big software company, such as SAP or PeopleSoft. That's why technology lags can wait for such software companies to mass-produce additional software modules for various business functions. The advantage of this integration is that various modules of the same software company can easily be combined with each other.

On the contrary, for those companies whose competitive advantage depends on the latest and most advanced technology, it is more necessary to obtain relevant capabilities from the best software companies in a field, such as supply chain management or CRM, which generally Will bring products to market earlier than big software companies.

The next step is a cost-benefit analysis, suggests David Root, CFO of cost-benefit analysis training firm Eagle's Flight. "Integration is a decision based on return on investment," he said. "You need to ask, 'How much will it cost, what will it cost in terms of time and labor, and what will be the associated benefits?'" It can help managers understand exactly what business processes are beneficial for integration. Numbers are the most telling. "A lot of integration work is going to cost a lot more than one would expect for a lot less benefit," Morelli said. "It doesn't mean it's wrong, it just needs to be thoroughly analyzed."

PHH's Talbert says his IT group works closely with business managers to determine where integration is ripe. They compare the current cost of managing the business process with the projected cost of the integration.

"For example, we asked a senior assembler to customize a light truck with a bucket elevator," Talbot said. "This process is performed by phone or fax and takes about two hours from PHH staff. .But this kind of work can be done only once a year. So if we integrate our system with the advanced assembler system, even if it costs only five thousand or ten thousand dollars, it is not practical. If this kind of work is done every month It takes two or three times, and that makes a big difference."

Talbert further suggested that a cost-benefit analysis should drive the actual level of integration. There are many ways to integrate systems, each with its own pros and cons. For example, batch integration processing, where updated batches of information are sent between the two systems at predetermined intervals (perhaps 24 hours). While this integration doesn't provide continuous updates 24 hours a day, it's really cheap and easy to implement. On the other hand, real-time integration can tie the systems together very tightly, and information can be continuously updated. While this has obvious benefits to the business, it can be very complex and expensive. The trick here is to keep in mind that business needs guide the integration strategy.

Balancing technical issues with business needs

Business managers should also consider technical issues and business needs in the integration project. For example, the VP of sales pushes the chief information officer (CIO) to implement a CRM project as quickly as possible. The technical manager was faced with a choice between meeting the VP of sales deadline or completing the various parts of the integration project as comprehensively as possible. This is a difficult choice that has to be made.

"CIOs are trying their best to lag behind," argues StarKist's Morelli, "but sometimes he finds it's an agency's drive. He's either catching up with it or being overtaken by it." There's a lot of pressure, Morelli says: "Most business people need this integration because they're exhausted and their bosses are pushing them to do more."

Once the integration is in place, the manager sponsors Business users should be helped to understand how various projects will change their work process. Because the systems of different departments are linked together through integration, so will the business process. As a result, the new tasks that emerge after linking require unprecedented collaboration between two and more departments.

Morelli presents us with a hypothetical business process for a purchasing department. Before the integration, employees were required to perform 6 different tasks to complete a purchase order. But after the stand-alone procurement system is integrated with the rest of the company's systems, the combined procurement process may require 60 tasks. There are good reasons for these new tasks to show that companies can reduce order errors caused by re-entering data, or keep inventory levels low.

However, when users face new tasks, it is sometimes difficult to see the whole picture. "Business users don't realize that integration will bring them new rules and workloads that they don't currently have," Morelli said. "No matter how the system is integrated, you're not going to make it work unless the information is also available. Integration. Relevant aspects of business work must be kept up to date."

Over the years, integration has been a huge challenge. As technology environments become more complex and business pressures mount, integration isn't going to be easy anytime soon. Despite the difficulties, the potential rewards are irresistible. For Jim Trotman, president of corporate strategic development at PHH, the results already show that. "Over the years, I've considered integration and construction to be a huge pain for me many times over the years, but now it pays off month after month and I'm almost an evangelist for it," he said.

This text is reproduced with permission from Putting Two and Two Together by Carol Hildebrand, published in Darwin Magazine, January 2002, copyright registered by CXO Media Inc. Translated by Zhai Shu. Carol Hildebrand is a senior editor at Darwin Magazine.