Last month, Schindelbeck offered tips for assessing new software. Now he's sharing the top 10 things to avoid when considering a proposed software upgrade or implementation. The list, he adds, reflects "the most egregious errors we have witnessed in the industry."

• The "It's an IT Only Project" label. Dispel the perception that the IT department alone conceives, leads and delivers the upgrade. Something as momentous as an upgrade affects everyone in the organization.
• Scope creep. Making one little change in your project--and then another and another--is costly, confusing and guarantees implementation delays. Make sure everyone understands and accepts the project's objective and scope from the start.
• Using unproven products or providers. Save the trailblazing for your business, not your software upgrade.
• Biting off more than you can chew. Change is difficult, especially when it comes all at once. It's best to implement your new software in phases that will both demonstrate progress and allow people to track it.
• Underestimating the impact of the upgrade. Accept the fact that implementing new software will require changing some of your existing processes. Analyze your current processes, consider how the new software will impact them and capitalize on the change as a process improvement tool.
• Customization. A custom designed software suite may sound appealing, but added complexity and cost overruns are likely. Leading software solutions designed specifically for real estate will meet the needs of most companies.
• Keeping the project secret. Communication is an asset. The more you share about the software upgrade, the greater the likelihood it will be accepted and understood. Have a change management communication plan that incorporates all key stakeholders into the project.
• Converting bad data. Be sure to "cleanse" the data in the legacy system before converting to the new system. Otherwise, you risk giving new life to problems that frustrated you in the old system.
• Cutting corners on testing. Creating a well-considered test plan will pay big dividends. Make sure all appropriate project team members and end users are involved in testing the software.
• Cutting corners on training. This is another area where going cheap doesn't pay. Make sure you identify everyone who needs training. Develop effective role-based training for each user constituency.