Enterprise resource planning (ERP) implementations fail. I’m not saying there are a few here and there that don’t quite work out – I’m saying they fail a lot. Panorama Consulting found a failure rate of 21% last year. The same survey found that costs were on the rise, customer satisfaction was on the way down, and things look generally bad in the world of ERP implementation.
One thing you can do to increase your likelihood of success before trying to setup a new ERP system is thoroughly cleansing your data. This should help nip some of the potential problems in the proverbial bud before they become real issues.
What is data quality?
Who knew data could be dirty? With ERP, as with many things, you’re only as good as your data. To be useful, data can’t just be complete. It’s got to be formatted and organized right as well.
Clean data is another way to say high-quality data. High-quality data is able to answer the questions you ask of it. A big part of this is consistency and accuracy, two of the biggest keys to data quality.
Imagine having a billing tool that stores a customer’s address as “XYZ Corp, 123 Lane Rd, Suite 34, London, UK W1H 7EX” and a CRM that stores is as “XYZ Corporation, 123 Lane Rd, Ste 34, London, UK W1H.” If you ever ran a report trying to connect your billing to your marketing efforts, you end up with a mishmash of customers. That’s a data consistency error which can be avoided.
Now imagine the same entry as “XYZ Corp, 125 Lane Rd, Suite ??, London.” That’s an accuracy error. You might be able to use that information to walk to the company – with some help – but it’s nothing like the accuracy you would need to determine a sales tax rate.
The purpose of an ERP is to connect all of your business tools and their data to tell you the “truth” about your business. But if you have one sales figure from finance, a different one from the sales department, and a third one from marketing, the ERP can’t tell you anything you don’t already know. Making sure the same figure represents the same data regardless of its origin is a requirement to getting to the truth.
Why does data matter?
Data is the key to understanding what’s happening in your business. If you’re looking for an ERP, chances are high that no one person can see and understand your entire business. It’s going to be too big for that.
An ERP is only as good as its data, though. If you’re running a report on the average value of your customers, duplicates and other errors are going to produce more errors. Then you make a decisions based on those errors and the cycle continues.
On top of the planning costs, imagine discovering that 10% of everything you mailed out to customers never made it to them. That’s just money down the drain.
Data cleaning basics
Your data, no matter how much of it you have, is going to have some errors in it. Clients don’t update their details, people type in the wrong things, and place names change. You may also have data that, due to no one in particular, is no longer accurate enough.
If you have a legacy system that only uses a five-digit zip code, then you never had the ZIP+4 that your new system wants. If you’re looking for consistency, you’ll want to set all your data up to use the same format.
Finally, clean data is clear of duplicates. You inevitably end up with both Johns Cleaning and John’s Cleaning in your records. One of them can go.
How to clean your data
The ERP you’re putting in place is supposed to help you get your company on the right foot, but before you can do that, you’ll need to work together to make sure the data going in is worthwhile.
Data cleaning relies heavily on both manual and automated solutions. Before you can clean data, you need to understand what data you have and how accurate it is. This is usually done with an initial manual check – or by a small bit of automation in Excel, for example.
With a small set of addresses, for instance, you may be able scan a couple hundred rows to see if duplicates jump out or if fields are missing. This lets you know what kind of concerns you’re likely to face when cleaning your data.
If you have a customer dataset with hundreds of thousands of entries, you’ll need to use some automated tools to ensure cleanliness. Even with tools, though, there will be some amount of human input.
One of the keys to cleaning data is agreeing on data standards across departments. Is it going to be “North 34th Avenue” or “N 34th Ave?” Once you agree on a standard, then you’ll need to check your records against the master list. Changing “N”s to “North”s can seem like a small thing, but it ends up being a very big thing.
In practice, the ERP you choose is going to dictate a lot of your data work. You can fix duplicates and delete worthless records without knowing what ERP you’ll use, but once you pick one, you’ll be tasked with getting things into the right format.
The dataset you end up with will be tailored to your ERP, be free of any rough data, and be ready to report on, right out of the gates.
Header by Rachel Wille
Additional Images by Abby Kahler
Looking for Logistics software? Check out Capterra's list of the best Logistics software solutions.