Postmortem on a Troubled Implementation

In response to a slackening market--or at least to the potential for one--most upper-tier ERP vendors have adopted two strategies in recent years. First, they've tried offering solutions that benefit mid-tier and smaller companies. Second, they've delved into the CRM arena by building/buying software to address this high-growth market. As they've moved forward with both strategies, they have sought showcase clients to help put some meat behind their message.

Being a showcase client can have its benefits--a great price on the technology solution, please-you-no-matter-what service, and an extra strong determination on the part of the vendor to make the project a success. But there's also risk involved, as Bruce M.*, an executive with a privately held publishing firm, learned the hard way. (Because of potential litigation with the third-party consulting firm, Bruce did not wish to reveal his full name or the identity of his company.) When Oracle approached his company a couple of years back about being a showcase "smaller" client for its ERP offerings, he agreed. The project was slated to take nine months and cost around $1 million.

Two years, $2 million and a few thousand Tums later, the system finally works, though temp employees are still on site taking final steps to ensure that things won't fall apart again. Bruce doesn't place all the blame on Oracle. The near disaster faced by his company resulted from a string of mistakes and poor decisions, as he explained when I recently spoke with him.

GKC: With less than $50 million in annual revenues, why did you even consider purchasing a big ERP system like Oracle?
Bruce M.: We didn't initially consider Oracle. Back in 1998, we needed to upgrade our technology infrastructure, so we brought in a Big 5 firm to do an assessment. The Big 5 firm recommended that we go with a tier-2 ERP system. Looking back, I think that was an erroneous assumption. I don't think they understood some of the publishing-related nuances that would later create problems for us. Because of those nuances, ERP really didn't make sense for us. But at the time, we didn't know that and we went forward with the recommendation.

GKC: Given your company's size, why did you turn to a Big 5 firm for help rather than to a local consultant?
BM: Senior management had an entrepreneurial vision and perhaps felt that if they were going to be big someday, they ought to think big now. So they brought in a Big 5 firm. The assessment was expensive and the company gulped hard at the price tag, but the 'let's do this big' attitude still prevailed.

On my first day on the job, the Big 5 firm was giving its final proposal. Already things were getting more expensive than we'd projected. We ended up widening the project scope during the assessment and we were suffering from sticker shock. I recommended that we use a smaller regional firm to help us with selection and implementation, since it would run about half the cost of the Big 5 firm. My suggestion was taken.

GKC: The Big 5 firm recommended a mid-tier system. What made you go with Oracle instead?
BM: I engaged a regional consulting firm to assist us. They represented themselves as a general ERP consulting company, so they had experience with Oracle as well as various tier-2 packages. After looking at several mid-tier options, they explained why these wouldn't fit our business. In fact, of the two vendors we'd narrowed down to, one pulled itself out of the running when it got a better grasp of our needs. The other system was fairly weak, so our options were zero.

At that point, the local firm said we should talk to Oracle. Because it was trying to penetrate our end of the market, they told us, Oracle was looking for smaller businesses and we'd be able to get a good deal.

GKC: What was your initial reaction to this suggestion?
BM: My general impression of Oracle was that if you bought from them, you'd better be ready to hire a lot of consultants, so I did a bit of a double take. But Oracle came in and said they wanted to make us a showcase client. 'Tell us what you can afford and we'll make it happen,' they said. We had selected 25 users to go through the evaluation and demo, and Oracle did a tremendous presentation for them. They said they'd changed their direction and knew the expensive consultant route wasn't feasible for small companies.

We finally agreed that we could afford the system providing that we could roll it without customization and that it wouldn't require us to employ high-cost consultants. Oracle came within our price range, and the consulting firm did also. There was very much a 'we'll do whatever it takes to make this work' attitude.

I went back to the company and said that Oracle wanted to make this work because they needed a small company reference on NT. They'd committed to our price and were determined not to let this fail. For those reasons, I felt like we couldn't go wrong.

GKC: What happened next?
BM: While Oracle said they ran on NT, I don't think they considered it a Tier 1 platform at the time. Saying they ran on NT was a bit of a misrepresentation.

Since we wanted the system up and running before Y2K, the consulting firm laid out a game plan to roll out on fast-forward. They projected nine months to implement 14 modules of Oracle software. That should have been a red flag right there. The consulting firm had only implemented sites with four or five modules. They didn't have the breadth and depth for something this extensive, as we soon discovered.

GKC: Had you done you due diligence on the consulting firm?
BM: Yes. We talked with four reference sites and they were all happy with their Oracle implementations. The one red flag we missed was that the references weren't using NT. Shortly after we signed the contract, the consulting firm lost some of its key behind-the-scenes people, and I think that's where the problems really began. Those people would have supported the on-site team, which ended up struggling with the modules of Oracle they hadn't dealt with before.

Also, as the consulting firm came to better understand our needs, they told us that Oracle as is couldn't do certain things after all and that we would need nine customizations. We had to bring in Oracle Consulting--one of the two things we'd really wanted to avoid--and they came back with a price tag almost as steep as the cost of the software. I took Oracle to the table on this--saying the product had been misrepresented during the demo--so Oracle agreed to share the expense of using their consultants. We pared down the customization to a bare minimum, but it still added to our overall cost.

GKC: After Oracle Consulting stepped in, were you able to roll out?
BM: Yes, we rolled out in the early test phase, but the system was still unstable. I was fairly uncomfortable by that point, but the consulting firm assured us everything was fine. Up to a week before the go-live date, they still hadn't done a full test run, and that defied everything I knew. But all the individual pieces were working, and the president of the local consulting firm told us they felt as good about the implementation as anything they'd ever done.

GKC: When did you go live, and what happened?
BM: We went live on December 7, 1998. By mid-January, things were getting way out of control. The release wasn't running properly and was taking huge amounts of time to run. Orders weren't getting processed all the way through. Data was getting hung up and messed up in the system. By the end of January, we couldn't even release a single order at a time. We were dead in the water.

The third-party consultant blamed all the problems on NT because all of their other implementations were on Unix. They felt abandoned by Oracle in terms of the NT environment.

At that point, I leaned on Oracle. 'This is a partnership,' I insisted, 'so get your people in here!' To its credit, Oracle sent people who literally lived in our computer room. They pointed out a number of problems and patched them, but the system still didn't work. The local consulting firm also put four people on-site without cost to us. But after two months of both groups working together, we still had a crippled system. Knowing we were in real trouble, I moved the company back to its old legacy system for order taking and shipping. The AR and GL pieces, however, had been transferred to Oracle and there was no way to roll that back. As a crutch, we manually entered AR information to handle invoicing.

By March, we concluded that the local consulting firm lacked the necessary skill set and the links with Oracle, and we wanted them out of the picture. We wrote a letter to Ray Lane of Oracle and said, 'We're supposed to be a showcase client. Well, we're ready to show!' We got an immediate response from Ray Lane's office. Oracle sent out one of its top field support teams, which looked at the system and told us we would have to re-implement. They said we had hundreds of thousands of rows of stuck, corrupt data and that the system simply hadn't been installed properly. They blamed it on the fast forward approach.

GKC: Did you agree to re-implement?
BM: We had no choice. We got rid of the consulting firm and told Oracle, 'This is your product. It's messed up, and it was set up based on what you told us to do. Help us.' Oracle came in and spent months here. A top Oracle support person did nothing but clean up our data for three months. Another team totally reinstalled the software. We went live on December 21, 1999--just days before Y2K. Today, the system is pretty clean and things are working better, though we still have temps on-site.

GKC: If you had to do this again, would you still go with Oracle?
BM: I would have looked harder and probably determined that we shouldn't have put in an ERP system at all. A company this size has no business looking at a Tier 1 ERP, but the big ERP guys are looking down to smaller businesses in fear that their high-end market will eventually run out.

GKC: Were you satisfied with Oracle's level of effort, and will you consider them for CRM?
BM: Yes on both counts. Because we're on Oracle now, we'd be foolish not to consider them for CRM. They've dedicated huge resources to that area. They mean business, and I don't think they can afford to lose the CRM battle.