Validate Assumptions, Validate Your VendorsBy Anna Maria Virzi | Posted 2004-11-01 Email Print
Thinking more like a consultant than an employee can help you make some tough decisions.
Before Hilton Hotels, I had no experience in the hospitality industryexcept that I had stayed in lots of hotels. I had been a consultant at Deloitte, with a background in information technology.
I was hired just after Hilton Hotels purchased Promus Hotels in December 1999. That more than doubled the size of the company. Hilton had about 38,000 employees; Promus had 45,000. My job was to integrate the systems for the two.
In effect, that meant, again, being a consultant to company executives. Indeed, I'm at my best when I can manage without having an agenda, alliances or turf to protect. I review the way things are done and challenge assumptionsand see whether the facts behind the assumptions need to be revalidated.
Hilton had always been a Unix shop, and there was an assumption that it would continue using Unix. To prepare for the Year 2000 changeover, Hilton had purchased PeopleSoft software in 1997 to handle payroll for Hilton employees and had invested in new, larger Hewlett-Packard Unix servers.
When I arrived, I didn't want to assume that keeping the Unix servers, like anything else, was a given. We didn't toss out the Unix servers immediately, though. We first switched Promus' payroll and other financial applications to the Hilton systems. Once the systems were integrated, however, we looked to further improve efficiency.
The main assumption to be challenged was: Should we change the platform? I.T. staffs often identify closely with particular vendors and their technologies. As a result, staffs may avoid big changes because they fear the unknown.
When we looked at our options, we realized that it didn't make sense to continue to be a proprietary Unix shop. Or to continue using Sybase's database with PeopleSoft. The hardware, operating system, database and maintenance on the Unix platform were all more expensive than alternatives that offered better price/performance.
We first weighed switching to Oracle, but then we asked: "Why not Microsoft SQL?" The costs for Windows/SQL Server were significantly less than keeping the Unix/Sybase platform.
I knew other companies used PeopleSoft with SQL, but none of those deployments were as large as ours. We took the risk and selected SQL on Dell computers for our PeopleSoft payroll application. This not only offered the opportunity to save money while meeting performance objectives, but we could also standardize our systems in the future.
Before we proceeded, we had promises from Dell chairman Michael Dell, PeopleSoft CEO Craig Conway and Gordon Mangione, product manager of Microsoft's SQL server team, that they would do anything within their power to make the project work. In return, I promised to be a reference customerif the project succeeded.
At one point, we almost had a disaster when we fired up the payroll system and it hit a bug on the SQL server. Hilton operates the ultimate real-time environment; almost all data is available at all times, with virtually no batch processing of new information. Employees process work and run reports at will. That means there is little predictability of demand, which makes it difficult to run stress tests. When we encountered an untested condition, the system stalled.
When that occurred, Microsoft provided extraordinary support. Six of the company's best developers came to our Beverly Hills office to work with our PeopleSoft developers and fixed the problem.
In the end, the new enterprise resource planning system reaped major savings. Hilton spends about $8 million a year to operate the system, a savings of $5 million or 38%.
My advice: If you challenge assumptions and try to change the way you do business or run your technology shop, you can't do it philosophically. You have to do it practically. And that means making your technology partners stand behind their products and make their success depend on your success.