How to handle the datasets that can not be combined into your in-house spend analytics solution.
When companies contemplate implementing their own in-house spend analytics solution, projects can be derailed by not addressing how to manage the multiple sources of spend data available.
While many ETL and reporting systems can handle combining multiple datasets from multiple sources, there is still the issue of how to deal with datasets that can not be simply combined. How do we address invoice data versus PO data? How do we account for P-Card data?
These issues can cause quite a few headaches and can not be solved by a simple technical solution. We need to understand the characteristics of the data sources and why we would want them included in our spend analysis and how to avoid double counting.
Invoice data is the most common source of spend data and the most accurate. However, invoice data is often lacking in detailed descriptions of what was purchased.
Purchase order data provides better descriptions, but the spend associated isn’t always accurate. Additionally, not all spend is covered by a PO.
As with PO data, purchaseing card data obtained from a provider can be more descriptive of what was purchased. Invoice lines for P-Card spend are often monthly lines to the provider, as opposed to detailing what was actually purchased and where from.
So how do we handle incorporating these differing data sources into our spend analytics? If you are lucky enough to use the same ERP system for Invoices and PO’s, there is often a means to connect the two sources. This would entail beginning with the invoice data, and then appending the PO content onto the invoice lines. This is obviously the best case scenario, but not often the situation most companies find themselves in.
When companies are faced with differing invoice and PO systems, there is often the desire to connect the two data sources to bring in one file with elements from each system combined. Based on experience, this is a huge mistake.
Due to the very nature of the data found within an Invoice and PO system, it is almost impossible to create one unified file from two differing systems. Instead of trying to force our data into a best-case scenario that isn’t applicable, the reality is that you need to work with the data you have.
The best option is to bring in both sets of data and handle the issue of double counting through the reporting interface. If you are able to connect some PO details to the invoice data, then that’s great. You should supplement the invoice data where it is available/possible. However, you still need to bring in all of your PO data intact. By bringing in both datasets, you are able to take advantage of the all the information you have in your system and utilize both when it comes to classifying your data. By having both datasets within your system, you are able to view all spend information for a given supplier when looking to classify your data.
Let’s say you have a large amount of spend with BASF Corporation. BASF Corporation has an extensive catalog of products, so simply looking at invoice level spend may not provide you with much information about what you are purchasing. By pulling in both sets of data, you are able to take a look at the PO details to see what the products are.
You may find that you are only purchasing resins from the corporation based on the item descriptions. Based on your visibility into the PO information, you are now able to classify all of the supplier’s spend to resin. You also get the benefit of being able to classify invoice spend at a higher level, while providing the more detailed classification for the PO level where it is available. This gives you the ability to dive deeper into the spend without losing the accuracy of the invoice information. PCard data can be treated in a similar fashion, but there is also an opportunity to replace the invoice level payment to the provider with the more detailed information. However, complications can arise when you have additional payments to the provider outside of your PCard spend.
At the end of the day, you are better off keeping two things in mind when dealing with your spend sources:
SpendView can help you handle multiple datasets. SpendView is a comprehensive Spend Analytics solution that consolidates, cleanses, and classifies all spend-related data in one place, giving organizations full visibility across all of their spend. With SpendView, take control of your spend, make better purchasing decisions, and identify savings opportunities that directly impact the bottom line.
To thrive with your data, your people, processes, and technology must all be data-focused. This may sound daunting, but we can help you get there. Sign up to meet with one of our analytics experts who will review your data struggles and help map out steps to achieve data-driven decision making.