While the recession has probably forced some IT managers to stretch the shelf-life of their ERP systems longer than originally anticipated, the time may finally be right to start thinking about an upgrade.
Most ERP vendors have been hit hard by the slumping economy, which will translate to better deals for buyers of ERP applications.
But leaving these monetary concerns aside, for IT managers that are actually going to upgrade, consolidate, or expand their ERP systems, a handful of tips on how to avoid common mistakes in the implementation process might be useful.
Actually, seeing as the headaches associated with ERP are the stuff of legend, reading up on where most companies go wrong might be necessary — especially if you weren’t around when your company first dabbled into ERP however many years ago.
The “we’ve always done it that way” approach is a foolish way of handling an ERP project (or really any project you’ll ever undertake in IT or life in general). A decision you made five years ago might not make sense today, so when you’re making an upgrade to a key IT system, the status quo isn’t going to be good enough.
“A lot of people don’t invest enough effort in identifying their future processes,” said Ray Wang, a partner with the San Mateo, Calif.-based research firm Altimeter Group. “If you’ve got this way of taking orders and you’ve done it this way for the last 20 years, chances are you’re probably going to automate that and not take the time to look at the business process and redesign it while you can.”
One of Wang’s unnamed clients recently decided to automate the way they take orders. The only problem was that the process was first developed in the early 1980s and have remained largely unchanged for over a quarter century.
“When they started enabling Web orders, they found that they ran out of paper,” he said.
It turns out every Web order coming into the system was being printed out, entered by hand and faxed over to the warehouse.
This example might be an extreme one, Wang said, but it makes a strong case for IT managers in the middle of an upgrade to their ERP landscape.
“Figure out how to streamline what you have,” he said. That means stepping back and really looking at how people are doing things differently in the workplace today, compared to processes during the initial ERP implementation.
Wang advised IT shops to hold design sessions where they actually ask employees how their systems can work more efficiently. While this may be a novel idea for some IT departments, it definitely shouldn’t be overlooked, he added.
Take your time, do it right
“Everybody’s trying to do ERP on the cheap, but to be quite honest, you get what you pay for at the end of the day,” Wang said. “This is especially true with implementations. You can’t substitute for upfront planning time.”
For Wang, his use of the word “cheap” has nothing to do with the money you’re paying the vendor, system integrator, or consultant to help you with the project. The old adage that “time is money” is certainly something that many business leaders take to heart.
As an IT leader, you will have to avoid being “cheap” in this regard, and really develop a clear focus on what you are trying to accomplish. This, of course, will take time.
“The most important first step is to assess your own needs and limitations,” said Warren Wilson, a research director covering dynamic application ecosystems with Ovum Summit. “Exactly what do you need this to do, what problem are you trying to solve, and what resources do you have or can (you) hire?”
If you don’t have the problem you want to solve mapped out properly, you will not have any success targeting the solution, he added. While this seems simple enough, Wilson said that he is amazed at how many companies tackle this step either at the end of the process or not at all.
Another often overlooked best practice, according to Albert Pang, a research director covering enterprise applications with IDC Corp., is to make sure you avoid starting the project at the end of your fiscal year.
“Close your books with the existing system and then get prepared at the start of the new fiscal year to make the purchase,” he said.
The modular model
While all the bells and whistles that are available from some ERP vendors can be appealing, sticking to the basics is probably the smartest move for any implementation or upgrade.
While a “plain vanilla” ERP system implementation will probably not win you any awards or get you on the front cover of a future issue of ComputerWorld Canada as a model case study, it will almost certainly not get you fired either.
“Implement your financial modules first, which is usually the core of your systems, and when that’s said and done, move onto other modules,” Pang said.
Paul Hammerman, a principal analyst with Forrester Research Inc., agreed, saying that most ERP systems revolve around a core transactional hub. “Starting with that core and building out to add other applications in stages is the way to go,” he added.
Implementing your system in one business unit as a pilot or proof-of-concept, then expanding on that implementation to other units, can be an effective way to ensure this happens.
Instead of forcing you to buy the “whole enchilada,” Wilson said, more vendors have been “SOA-enabling” their products to make them easier to adapt and more flexible to roll out.
“You can start with financials and then add supply chain, but the idea is to keep it focused and contained,” he said. “Get it in, up, and running and see whatever else you might need as you go.”
Of course, you want to have an overall plan, but only a plan that covers the core functionality that you will need. All the bells and whistles can come later, but only if the opportunity presents itself.
While it would be really difficult to find a company that doesn’t test their ERP system before rolling it out to their users, many IT shops simply treat this step as an afterthought, Pang said.
“You just can’t do enough testing,” he said. “After testing, you should do more testing and then after that, continue testing again.”
Even if you think a system is fairly stable, you just never know how the ERP system will react to the rest of your environment, Pang added.
Testing can especially come in handy for companies that are implementing a brand new version of ERP software that hasn’t matured yet.
Wilson put it bluntly when he said that anybody that doesn’t test their systems at the scale they’ll ultimately need is being foolish.
Pat Phelan, a research director at Gartner Inc., said that along with not having full-time team members dedicated to project management, lacklustre testing is one of the most common mistakes she encounters.
“In the places where I have seen testing be successful, they’ll do a bit of scenario or risk assessment upfront and spend more time testing the things that are either likely to break, highly complex or business-critical, and less time focusing on testing the more mundane areas,” she said.
On the other hand, she added, for some areas such as payroll, the risk-based approach to testing doesn’t work.
“In that case, I’ve seen organizations be very successful if they create what I’ll call a ‘controlled test environment,’ where they don’t necessarily parallel test every employee or every transaction,” Phelan said.
While it’s important to never get complacent and stay the course, there’s something to be said about consistency as well.
For Wang, it’s important to go with a vendor that matches the technology you already have in play. You can’t literally start from scratch in your data centre, so if you’re a .NET shop, you might want to stay that way by going with a vendor that builds on Microsoft’s .NET technology, he said.
Pang said the last thing that any IT leader wants to do is “rip and replace” existing infrastructure to accommodate an ERP system. “If you’ve already invested in infrastructure on the Java side, then Oracle would be a good alternative versus a Microsoft product,” he said.
Project management also plays into this issue, according to Phelan.
“You need people who’ve actually been through a project like this before, understanding your business, understanding the technology and understanding the fundamentals of project management,” she said.
As for other vendor selection tips, Phelan said that companies should not hesitate choosing a Tier 2 product, assuming the scope of the ERP project is limited.
“But you can’t stretch that product beyond what it’s capable of doing,” she said. If you choose not to heed that kind of advice, she added, expect to have a system that poorly fits your needs and leaves you with a significant amount of future customization work.
Find the A-Team
It’s important to remember that as soon as your company engages with an integrator or a consultant, the meter starts to run.
For Wilson, a key piece to the ERP upgrade puzzle is making sure that you vet your implementation partners carefully.
“You have to make sure that they’re not only experienced in your vendor’s software, but also experienced in the particulars of your industry and its geography,” he said. Different industries have much different processes that ERP has to be configured to support, Wilson added.
“Pick somebody that’s done for somebody else exactly what you need to be done, and hopefully more than once,” he said. “Probably closer to a dozen times.”
If you can’t find a company that fits the bill, look for the closest match and spend a lot of time checking with references and learning from their customers’ mistakes, Wilson added.
According to Phelan, this step cannot be stressed enough.
“You can have a mediocre product be successful if you get a systems integrator that is excellent in making do in a difficult spot,” she said. “Likewise, you can have an excellent product fail miserably if your systems integrator is not on the ball.”
In addition to a strong systems integrator, you must ensure you have competent, full-time project team members in core roles.
“Make sure you have a deep bench of IT skills to go with the implementation,” Pang said.
While a reliable systems integrator or consulting firm is extremely important, he added, at the end of the day, ERP systems should be considered a core asset.
“You really don’t want to put it in the closet, lock the door and walk away from it,” Pang said. Developing an internal competency among IT staff should not be overlooked.