Driving out dirty data

Spend Matters recently launched its procurement digital transformation survival guide for practitioners. So far, you can visit our analyst Bertrand Maltaverne’s thoughts on why change is hard and how to work through it. Other topics will include ‘how to build a digital roadmap for procurement’ and ‘how to navigate?

the vendor landscape.’

?

For my part, I will be sharing some thoughts on how data — specifically ‘dirty data' — hurts the digital transformation process. More importantly, I will explain how you can dig yourself out of the ‘dirty data’ hole.

?

For example, one common problem that leads to poor data quality is inconsistent data formatting. Let’s take the example of supplier records — perhaps Walmart is listed in your records as Walmart, Walmart.com, Walmart LLC and Wallmart. Due to this inconsistent and inaccurate formatting, the supplier record for Walmart is fragmented, which could lead to problems such as spending off contract.?

?

While the problem itself is fairly easy to identify, fixing it is tedious and ever-evolving; for every inconsistency you fix, a new one pops up in its place. ML and AI can help here. A machine learning model can assist with data normalization and cleansing by completing repetitive tasks, such as removing “inc.” or “LLC” from supplier names for both historical and real-time data.

?

This is just one example of the helpful information to be covered in our procurement digital transformation survival guide for practitioners. For more on this topic and others relevant to surviving a digital transformation, keep an eye on Spend Matters research hub.?

— Abigail Ommen, Director of Analyst Production and Research Analyst