ERP Failure Avoidance Guide: Part 3 – Data Cleansing
Data cleansing is a unique combination of high risk and low contingencies, which is what makes it a top reason for ERP failure. It is high risk because one consistent error or one missing field can cause a transaction to fail, and you don’t need many transactions to fail before you have chaos. Another contributor to high risk is that it is not generally possible to test all of the master data prior to ERP implementation; the best you can do is test the biggest random sample you can manage. The reason data cleansing has few mitigating contingencies is because the responsibility of data cleansing is normally restricted to a very few people in the ERP team. This has the intended benefit of greater consistency and accountability, but it also means that it is very difficult to add short-term resources in the event of problems.
When you first start an ERP implementation, it is easy to assume that data cleansing means reformatting and rearranging all of your existing legacy data. It does encompass that, but it is much, much more, and it is difficult to grasp that until you start building out the ERP system.
The “legacy cleansing” work is understandable. Zip codes don’t match cities in customer addresses; the same customer is listed four different times with four different spellings; an sku is listed as obsolete in the manufacturing system and available to promise in order entry. These are just mistakes that need correcting.
Old Sins Revisited
The truly complicating aspects of data cleansing occur when no legacy data ever existed. This is where both sins of omission and data inaccuracies can trip you up. Sins of omission will generally prevent business from occurring, while data inaccuracies result in execution problems. Suppose that in your new ERP system, specifics about how a material is shipped is part of the material master data record, and that product ships in boxes stacked on a pallet. ERP wants to know how much product per box, how many boxes per pallet, what the dimensions of the pallet are, what the weight of the product is, and what the weight of the boxes and pallets are. In legacy, someone in the warehouse mysteriously dealt with all that stuff. If you do not enter the information at all (a sin of omission) the material cannot be activated and orders taken. If the data is incorrect, you will generate bad business information relating to transportation requirements, transportation costs, packing costs, and packing material needs. And you won’t discover any errors until there is either a financial variance, or a crisis at the shipping dock. Hey presto – ERP failure.
Be relentless when it comes to data cleansing. In the example above, the wrong number of pallets is a hassle, but what if the error was an incorrect process recipe and a large amount of product was no good? To avoid an ERP failure, put really good people on data cleansing, and think constantly about ways that you might find potential errors.
Featured white papers
Manufacturing ERP Failure: 6 Common Causes
Get your comprehensive guide to the causes of manufacturing ERP failureDownload
The mistake first-time ERP buyers make (and how to avoid it)
Avoid making first timer ERP selection errors with this guide
Two failed ERP selection scenarios… and how to deal with them
This unfortunate event can happen with any ERP selection.
Three ERP failure case studies and what you can learn from them
Actionable takeaways from three high profile ERP failure case studies