How to Avoid ERP Failure by Tracking Data

ERP failure due to data cleansing or master data problems is tricky to assess during the course of a project for two reasons: first, it’s difficult to quantify how big the total work load is, and how much tangible progress has been made against that work load, and second, unless you are highly outside the normal distribution, you will be cleansing legacy data and entering master data the night before ERP go-live. So this is one of those categories which will never be done early enough or completely enough to make you sleep well. The relevant question to this article is when is data cleansing/master data loading in such jeopardy that you should lose sleep because the ERP implementation is at intolerable risk? This is not a hard science – prudent judgment needs to be applied – but there are warning signals that should flag you to exercise that judgment.

The best indicator of master data readiness is whether or not your integration tests achieve all of their stated objectives. If, during every test, there are unplanned gaps because of missing or incorrect master data, then ERP failure is likely. If your final integration test – which should be an after-the-fact volume parallel of the real world – stumbles significantly because of master data, then go-live will stumble as well. Do not proceed with a go-live without a high degree of confidence obtained from successful volume testing of master data.

Warning Signals

ERP testing will also tell you a lot about data cleansing. If during testing, you are missing customers, or inventory, or purchase orders, you will likely track back the culprit to conflicting or erroneous data in legacy. Data cleansing normally boils down to a matter of trust, since most data cleansing problems must be fixed in the legacy systems, and for that you depend on the business users. When the night before go-live arrives, and the data has not been cleansed – and sometimes it won’t, because people just don’t get it – you have a tough decision to make, on whether to go ahead or pull the plug at the eleventh hour. Pulling the plug at the eleventh hour is obviously a highly visible thing to do, and you need to be prepared to answer the question “who didn’t do their job?”

Another warning signal of ERP failure is if data cleansing/master data milestones are missed by an ever widening-margin throughout the life of the project plan. This says you are behind and falling farther behind over time.

This is the area that will embarrass you, the ERP project manager, because generally speaking, you – and everyone else on the team – have delegated, communicated, followed-up, in general, done everything you were supposed to. When you find out that it didn’t get done as promised, it is still your fault, and is almost always too late to take effective corrective action. But if your data isn’t ready, postpone the implementation; with bad or incomplete master data, you will fail.

author image
Tom Feltham

About the author…

Tom Feltham was Editor at ERP Focus for three years from 2013-2016. Tom still keeps tabs on the ERP software market and occasionally puts pen to paper when inspiration strikes.

author image
Tom Feltham