As perhaps the world's first project manager, Noah had a tough challenge on his hands, what with all the intangibles -- bad weather, a mercurial boss -- and the need to manage an extremely diverse set of stakeholders. Yet for all the difficulties, 40 days and nights was nothing compared to a typical enterprise resource planning (ERP) software installation.
Try the equivalent of 14,000 days -- that's how much time Andersen Consulting said it would need to install SAP's powerful, complex R/3 software at Bay Networks Inc. Andersen could make such a prediction because, unlike Noah, it had done that sort of thing before.
Maynard Webb, vice-president and CIO of the computer networking hardware company Bay Networks, in Santa Clara, California, had done that sort of thing before too, which is why the 14,000-day prediction didn't come as a shock. Webb understood that consulting costs would be the biggest line item in his budget -- in some ERP projects, they add up to 10 times the cost of the hardware and software combined.
Good consulting advice has become so important to the success of really big ERP projects that the fees that consultants charge for those services are getting out of control, Webb says. And while the function of typical IT consultants has changed dramatically from the days when they basically came in, installed a mainframe computer and left, the way they charge companies for their services -- by the hour -- has not.
Getting paid by the hour doesn't create much incentive for consultants to watch their billing waistline, according to Webb. It also runs counter to the two most important goals in managing ERP projects: controlling the scope and the schedule. Let either of those slide and the project is doomed. How, asks Webb, do you keep a lid on ERP project expenditures when the consultants have different incentives from your own staff? "When the system integrators are being paid on a time-and-materials basis, they don't push back if you want to add things to the project," he says. "Why should they? It means more money for them."That conflict led Webb to ban the a la carte pricing system entirely when he put the Bay Networks SAP proposal out to bid in March 1995 -- it would be prix fixe or nothing. Webb wanted no more of the constant price bickering that had plagued some of his earlier consultant-heavy projects at other companies. "I got tired of constantly checking on who was working on what and whether they were worth their hourly rate," says Webb. "It got to the point where I was spending almost as much time doing time-and-materials sheets as I was solving real business problems."More importantly, Webb wanted his consultants to be as motivated to finish the project on time and under budget as his own staff. Webb's demand that consultants share his risk is a growing trend in the ERP market, according to Karl Newkirk, partner in Andersen Consulting's Americas SAP Business Organisation in Cleveland. "Clients are saying: 'If I'm going to bet the ranch like this, I want some insurance'," he says.
That's what Webb was looking for in April 1995 when Andersen bid on the job.
Together, Webb and Kevin Reedy, an Andersen partner in the electronics and high-tech practice in Palo Alto, California, came up with a package deal in which Bay Networks would pay Andersen the cash equivalent of 14,000 days' worth of consultants' time for the project in return for getting the R/3 system up and running by July 1996. The first 5,000 days' worth were spent planning the project and finding the right Bay Networks people for the team. Webb and Reedy also spent a lot of time prepping the top executives at Bay about the ERP options so that they understood exactly what they were getting into.
That left 9000 days' worth -- packed into five months -- for Reedy to get his work done. Any more than that, and Andersen was paying for his people's time, not Bay Networks'. "They were going to lose money if we slipped," says Webb.
"They were focused on the same thing we were -- getting it done." Bay Networks staff bonuses were tied to finishing the project on time and on budget. "The metrics were the same for them as they were for us," says Reedy. "Nobody talked about raising the days to 15,000."To spice up the bets on the magic number, Reedy developed what he calls a "gainshare" strategy, which called for both sides to assume some cash risks.
Reedy agreed that if the project was not completed on time with a level of quality that was acceptable to Bay Networks' board, Andersen would refund 10 per cent of the package price to Bay. If, on the other hand, the project was completed ahead of schedule with a favourable grade from the board, Bay would write a cheque to Andersen for the same amount.
For Reedy to come away with the prize, his 14,000-day estimate had to be correct. That was a gamble in itself. Although the hardware and software costs are easy to predict, people are what make or break an ERP project. A lot depended on the real (rather than promised) support for the project from the top executives at Bay and the quality of Bay's project team staff. Perhaps most important was the balance of power in the project team: would Bay try to control Reedy and his staff, or would the team share the decision making?Reedy relied on Andersen's SAP "estimating model" -- a big Excel database that measures 50 different project variables based on Andersen's collective ERP experience -- to reduce uncertainty in the estimating process. More importantly, the estimating model became a handy way to manage scope creep during the course of the project. "We went through some tough times early on," says Webb. "Kevin had a fight within Andersen to let him do the gainshare because we were being vague about the [project estimates]. At the beginning we didn't know how many reports [custom applications that sift SAP data into printed reports] we'd need."Webb and Reedy did battle early and often over the Excel printouts. Each week Reedy would run the numbers, and if they added up to more than 14,000 days he would call a meeting with Webb and the team. "We originally planned to have 70 reports, and as we got into it the users came up with 120," says Reedy. The team negotiated with the users to get the number of reports down to 80, and Reedy automated a few applications that he'd planned to write and fed the numbers into the database until it spat out 14,000 again.
Though the deal seems to place a bigger burden on Reedy's shoulders than on Webb's, he claims that he prefers the gainshare arrangement to a time-and-materials project. "The risks were more than compensated by having control over my staff," says Reedy. "When we're working on a time-and-materials basis the client has to approve all the résumés and the hourly rates of my people. When you're on a tight schedule that can be frustrating. On the Bay Networks project, I had the freedom to manage my staff and my costs any way I wanted to. When we were late in the game and I needed to bring in an expert to do something, I just did it. I didn't need to send his résumé past five different people for their approval. Webb didn't care, because his price wasn't going to change, no matter whom I brought in."After the system had been up and running for a quarter, Webb wrote a cheque in the low six figures to Andersen for beating the deadline and meeting the quality standards. But if Webb had run a credit check on Reedy's team, he would have discovered that the bonus money had already been spent.
"I don't know if I really ended up with a 10 per cent bonus," Reedy admits, "because I did go in myself and put in the extra time and brought in the high-priced experts to get the project done. It was more important to me to avoid the penalty."SIDEBARKeep Your Chequebook OutDon't underestimate the real costs of installing client/server packaged applications in a big company, say the experts. The biggest cost is tapping the brains of consultants. In a recent survey of 40 Fortune 1000 senior IS executives, Massachusetts-based Forrester Research found that 54 per cent of respondents were spending more than $US5 million on sage advice alone, and 26 per cent were spending more than $10 million. Sobering numbers that won't get any lower in the next few years, as companies continue to use client/server packaged applications like SAP's R/3 to drive business process re-engineering (BPR). Change, it turns out, is not only difficult, but expensive.
Forrester estimates that in a typical SAP R/3 installation more than 50 per cent of the implementation budget is spent on BPR. "When the essence of the project is change, then companies must be ready for pain," says Karl Newkirk, partner in Andersen Consulting's Americas SAP Business Organisation. "I think the idea of enterprise-wide software solutions is here to stay, and, unfortunately that means there's plenty of pain left to go."
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.