Promises, promises. Computerization will help your business run more smoothly, productively, and profitably, right? Ultimately, the answer is yes. Unfortunately, the path you take to get there is often littered with "ifs" and "buts."
I've been working in information technology for 24 years, and have helped many companies, large and small, implement new hardware or software into their businesses. Based on my experience, I've developed a process for selecting and implementing computer systems. The entire process is complex, but I've boiled it down to 10 "learning objectives" you should know. Pay close attention to these objectives, and you'll improve your odds of buying the proper computer tools for your company, and of implementing them painlessly.
#1. Know What You Want, Know What You Need, Know the Difference
This is most important. Remember the adage that if you don't know where you're going you'll end up somewhere else? It applies to computer systems as well. You must know what you want to achieve, or you'll get something else. You must have clear business objectives. That may seem obvious, but I've been approached many times by business people who believe they need a new computer system, but can't tell me why.
Follow these steps:
- Write down your goals. Make them S.M.A.R.T (Specific, Measurable, Achievable, Realistic, and Time-oriented).
- Survey your team for their goals.
- Rank the goals.
The top goals are needs. You must have these or there's no reason to do this project. The second tier are wants. You'd like to have these things, but the project would be successful if these goals weren't achieved. The bottom tier should probably be ignored (see Figure 1)
What do you do if you think you need a computer system, or new software, but you're not sure what it could do for you? Ask! Join an ACCA MIX® group. Get on the Service Roundtable and ask your peers. Learn from their experiences. Look at the back issues of Contracting Business. Contact a vendor, or a consultant.
However, do yourself a favor: Don't ask your wife's sister's husband's nephew, the reputed "computer genius." They think he's a computer genius because he plays computer games all day. He doesn't know your business, your industry, or business systems. Avoid him.
#2. Focus on Project Management
Would you drive from Texas to Alaska without a map? Probably not. Could you get there without one? Possibly, but it would be easier, faster, and less expensive if you had a map.
Computer implementations are similar endeavors. You need a plan. The project plan will make the implementation easier, faster, and less expensive. The plan needs to be shared with the team and documented. There are some very sophisticated project planning tools (Microsoft Project is my current favorite), but I've seen very good project plans documented with spreadsheets, and word processing. One of the first projects I was ever involved in had a hand-drawn Gant chart that was stuck up on our project team's wall, and it worked very well.
The essence of a project plan is really simple: Each task in the project plan must have three items: WHAT needs to be done, WHO will do it, and WHEN it must be finished. Each team member needs to know what he/she needs to do, and when it must be done (see Figure 2).
Some things to look out for in the project plan:
- Add time for the unexpected. Everything takes longer than you think.
- It's easy to "overbook" people. Make sure you don't have one of the team members doing three full-time tasks at the same time.
- Get buy-in from the team. Each team member needs to sign off that he or she can complete his or her tasks by the specified dates.
- Remember the pre-requisites and schedule accordingly. It may be that task B can't be completed until task A is complete.
- Hold people's feet to the fire. Their job is to complete their tasks on time.
- Schedule lots of time for testing, and time to fix the problems you discover in testing.
#3. Don't Use Too Many Vendors
Some businesses are tempted to buy the best system in each category, and then try to integrate those different systems. For example, they buy the best accounting system from one vendor, the best inventory management system from another, and a dispatching system from a third. Then they find a system development team to integrate them all. This is bad for several reasons:
- It's the most expensive solution
- It's very complex
- The user interface may be substantially different in each category
- If something goes wrong, the vendors will blame each other and you'll be left holding the bag.
Instead, buy one good integrated system. Each component may not be the best in its category, but that will be overcome by the value of having each component integrate smoothly with all the other components. And you'll have a common user interface.
This strategy also allows you to follow the "one throat to choke" theory. Make the vendor of the software responsible for implementing the software for you and with you. The vendor should do any customizations. One vendor is responsible for the whole system; they have no one to blame but themselves if there are problems.
#4. Prepare Demo Scenarios
Everyone knows to demo before you buy. But what most people don't know is that you need to tell the vendor (preferably in advance) exactly what you want to demo.
Be specific. Document your most important processes (most will come from the goals in item #1) and give them to the vendor. Tell the vendor you want to see these processes demonstrated completely. Be fair. Don't waste their time and yours demonstrating unimportant processes. Focus on the important processes. Save these scenarios for testing (see item #9).
#5. Expect and Manage Change
The software you purchase and implement won't do everything exactly the way you do it now. That's a given. When there are conflicts, there are two possible changes: you can change to match the system, or the system can change to match you.
Most of the time, you should change to match the system, because it's less expensive and because good software vendors implement best practice processes. Use this opportunity to clean up bad processes you may be using.
However, there are instances when you should insist on the software changing. The only time you should insist on this is when it is so vitally important to your business that to not change the software would cost you a substantial amount of money, force you to change the way you go to market, or force you to change your interactions with your customers. Otherwise, change your processes to match the system.
#6. Know the Total Cost of Ownership
Most people focus on the purchase price for computer hardware and software. However, that's only part of the total cost. If you only compare the purchase price of two software packages, you may miss something significant. It's not uncommon, for example, for implementation costs to be greater than the purchase price. You should consider all the cost components to develop a total cost. Computer analysts refer to this as the total cost of ownership (TCO). The main cost elements are:
- Purchase price or lease
- Implementation costs (consultants, internal team)
- Annual maintenance fees
- Custom programming
- Software upgrades (probably one per year)
- Initial training
- Ongoing training (for new hires or for upgrades)
- Internet fees
- Hosting fees
- Hardware purchase
- Hardware upgrades (every three years).
The average life of business software is three to seven years. Estimate your costs over the expected life of your system, and put these in a spreadsheet.
#7. Know that Everything's Negotiable
If you always pay full retail list price for your automobiles, you can skip this section. Once you have all the TCO cost elements in a spread-sheet (from item #6), play around with them a little bit. Does anything look unreasonable? Decide how much you want to pay. Develop a negotiating plan.
Don't be shy. Some software vendors have great flexibility in the purchase price, because the marginal costs of an additional sale are small. Conversely, they may not have as much flexibility in the consulting, custom programming, or training because they have to cover their employee costs.
#8. Manage Expectations
System implementations are difficult, complex, expensive, disruptive to the business, and risky. Some people want to believe a new system will solve all problems. It won't. For a while it will likely cause more work while unanticipated issues are resolved.
Some people want to ignore the new system and hope it will go away so they don't have to learn anything new and can stay in their comfort zone.
You need to provide vision, motivation, financial backing, and encouragement. You will need to clearly articulate the reason for doing this several times a day. Remind them that it will happen. Be positive. Keep the team focused on the goal.
#9. Test Carefully and Thoroughly
You must test carefully and thoroughly before you "go live" with the new system. It's much cheaper to find and fix a problem during the test phase than to find and fix it after the system is "live."
Create a test plan. This is a miniature version of the implementation plan. Involve your whole team. The test plan should be a list of tasks — who, what, when. Reuse the "demo scenarios" as "testing scenarios."
The elements of your test plan should include:
- Complete system test. Some consultants call this "A day in the life test." The idea is that you process everything you do in a normal day. You should also test the "abnormal" days, such as month-end and year-end processes.
- Volume test. This is to ensure the system can handle all your people using the system at the same time. Your vendor may have automated tools for this. If not, you can get everyone on your team to sign on, and test at the same time.
- Backup and recovery test. Your new system should have a back up and recovery plan. Test it. Enter transactions. View the transaction results. Backup the system. Delete the database. Try to view the transaction results (this should fail). Re-cover the database. View the transaction results (this should work) After you go live, remember to take your backups offsite to a safe place.
If you find problems during your tests, document them. Keep testing and retesting until all the problems and glitches are fixed.
#10. Plan and Celebrate Your Success!
Information systems implementations are tough. Put at least one party in the plan. I recommend waiting a week after implementation for the dust to settle. It can be a happy hour, a formal or informal dinner, a ball game, etc. Hand out awards, kudos, and pats on the back. Your people have all worked hard, the project was a success, and you need to tell them so. Celebrate!
David Heimer is chief operating officer for the Service Roundtable. The focus of his career is strategic information technologies for business. You can reach him at [email protected] Roundtable.com. To learn more about using technology to improve your business, visit www.ServiceRound table.com