Part 2: Why is data quality so important for good procurement analytics?

Once the raw data is in our systems, there are a myriad of additional checks we need to perform. For price benchmarking, we need to ensure we are benchmarking apples with apples. AdviseInc was the first company to crack the logic to enable this, using a statistical process to capture outliers and add information to tag lines that could not previously be benchmarked – for example, £1 or 1p unit prices.  

We also look at uniqueness and depth, asking questions such as: how unique is the contents of a field? How many blank cells are there? Should all cells be populated? And, has a cell been populated with a value that makes no sense? For example, when manufacturer product codes have been entered incorrectly with a catalogue ID for the product instead – this happens more often than you’d think. 

Clearly then, there is a lot we can do to help and give feedback on data quality. The question is, so what? 

At AdviseInc, we’re all about driving a positive change, providing tools to local teams that help them to improve. As such, once we have identified the gaps, errors and mistakes in a customer’s data, we pull together a plan to help them resolve the situation. Most challenges start right at the beginning: we see lots of teams spend hours trying to ascertain what certain order lines are supposed to mean, deciphering pages of content in order to make sense of it. Our golden rule to solve all this manual intervention is to get as much information as sensibly possible on a catalogue, and to invest in a good catalogue management solution. 

NHS trusts that have inventory management software know the importance of good quality data – systems fall over when they don’t have clean data – but those without an inventory system get away with it to some degree.  The issue remains unseen. All the problems happen downstream, and in accounts payable it shows, with teams of people often getting involved to understand the why, what, when and how of a purchase so they can account for it correctly. One supplier told us that they employ more than 40 people just to chase money and answer queries related to POs and invoices that have gone bad. That cost is paid for by the NHS.

From our perspective, the problem the NHS has is that organisations within it don’t always think beyond their own actions, and the consequences they can have on other parts of the system. This is where “system thinking” comes in, but we will bore you with that another time. A few more seconds spent upfront could alleviate a lot of wasted, non-value effort downstream.

So, to summarise what we have discussed so far: 

1. 1>Get the calibre right. Aim for high quality data to be input at source
2. 1>Don’t accept poor quality data.  Send requests back to requisitioners, or better still, own and maintain a good quality catalogue (from the likes of GHX for example).
3. 1>Manage your catalogue. Your catalogue manager should be the rock star in the team. Stay on top of products and pricing, and ensure only good quality information is entered. Period.
4. 1>Police free text requests that should have been on catalogue, and educate users on the importance. See (2) above. 
5. 1>Think about what could cause issues downstream. Don’t create blanket orders for all your orthopaedic purchases. We worked with a trust that did this and had absolutely no visibility of what had been purchased. 
6. 1>Automate as much as you can. Investment in automation is never regretted – it frees up precious resources to do value add work. Do not spend that resource manually pulling data out of systems and manipulating it each month.