Fixing an ERP Implementation

It's not news that not all ERP implementations don't go smoothly. Sometimes you can catch the problems while it is still in the implementation phase. Sometimes the problems don't become obvious until after the system goes live.

If that's the case you've got a serious problem. Given all the time and money invested in the system, you need to do something to salvage it, if possible. Fortunately there are a number of fixes you can apply which can help make a failed ERP system functional.

Understand, you can't always fix everything. If the implementation went off the rails early enough the mess may be unfixable. However if it went that far wrong in the beginning it's more likely you realized something was wrong early in the process.

More likely, things just aren't working right in the failed implementation. Generally this is, if not fixable, at least patchable to get the system to more-or-less work.

The first step is to go through and identify what isn't working on a department by department and function by function basis. Make a complete list of what is working poorly or not working at all. Identify your pain points in the system as it is now.

You'll probably want input from all the affected departments on this to make the list as complete as you possibly can. Also make it specific and determine what is an acceptable level of performance for the fix. For example if one department is having trouble getting screens to display in a timely fashion, it's not enough to note “speed up screens.” You need to quantify that. How fast do you need the screens to be? (Or to put it another way, what's the slowest you can put up with?

The next step is triage. It is seldom a problem will have only one fix. Examine the possible fixes.

Now, estimate the time and the cost for the classes of fixes. Some of your proposed solutions will take too long or cost too much, no matter how elegant they are. These are the ones that you will apply the quick-and-dirty fixes to. There will be others where you have the time and effort to do the job right. Finally there will be a class of problems you can patch with the quick-and-dirty fixes, but you will have to go back quickly and implement the costlier solution.

Now look at the major pain points. Not all these problems are going to be causing the same amount of difficulty. Prioritize according to which ones are hurting the business the most.

Next, go looking for low-hanging fruit. These are problems which are causing the most disruption but which are easiest to fix. Prioritize your list so these obvious and easy to fix problems are at the top.

Don't be afraid to change the business processes to get immediate fixes. Likewise, consider the possibility of new training for users to help solve the problem. The result of all this probably won't be pretty, and you may have to re-do a lot of it once you get the project stabilized.

The reason for doing this is to produce some results fast. Note that the low hanging fruit won't necessarily be the worst problems, merely the ones that will show fast results. This is mainly psychological since the organization is going to need some quick, obvious fixes to build confidence in what you're doing.

After you get some quick wins in place you can start working on the harder, more time-consuming parts of the problem.

What comes out of this won't be as efficient or as streamlined as the original design, but it should be functional. It should be good enough to keep you going until you can go back and spend the time and resources to do everything right.