Another flaw in the human character is that everybody wants to build and nobody wants to do maintenance. - Kurt Vonnegut
Imagine sitting at your desk at corporate headquarters and getting a notification that a buyer at your plant in Boise, Idaho is about to buy a replacement part for a machine that is still under warranty.  You were notified because the buyer assigned a requisition for the part using a company that was not part of the warranty program.  Since you were flagged before the order was placed, you are able to inform the buyer to go through the proper channels and avoid the additional cost. 

Scenario 2: A new admin places an order for office supplies with a company that is not a preferred vendor.  The order is automatically routed to the supplier with the best available discounts and the admin is alerted of the change.
Scenario 3: Your pricing for corrugated boxes is tied to the 42# Kraft index, and changes ever quarter to reflect current market conditions.  Pricing in your order placement system is updated automatically every quarter using the most recent index data, and provides a forecast out for the year based on trending.  It also automatically checks invoices from the supplier to ensure the right price was charged, and kicks back any invoices that reflect the wrong pricing based on current index information.

Scenario 4: Your company recently acquired a business that distributes nuts and bolts.  Your spend cube automatically categorizes orders placed out of that business unit as “Direct Materials” while keeping orders for existing business units as “MRO”.
All of these scenarios illustrate what can be done using predictive analytics (and good data).  Predictive analytics are the future of business, and the future is now.  As I write this blog, pharmaceutical companies are developing chips that will go into a pill to monitor a patient’s condition.  The data will be transmitted to a patch on your arm, and then broadcast to your doctor.  The doctor will be able to monitor your condition in real time, catching trends in your body’s behavior, and potentially correcting problems before they occur.

Insurance companies are using predictive analytics to determine the cost to cover a population, and adjusting their premiums based on the types of exposures they know they are going to incur.
The United States Government is tying systems together on a local, state and federal level, working with healthcare providers and correctional facilities to better predict and prevent catastrophic events, such as mass shootings, by looking for patterns in behavior of individuals that go in and out of those systems. 

“Predictive Analytics” as an industry is going through explosive growth and will likely be the foundation of all software tools, operating systems and electronics in the future.   This rapid pace of growth, coupled with the realization of how much predictive analytics can do, is staggering.  So why don’t you hear anything besides superficial conjecture about the use of these tools in strategic sourcing and spend management?
Probably because most companies are still trying to figure out how to add UN classification codes into their PO data. 

Procurement, Finance, and the rest of the organization have, for the most part, failed to understand the importance of proper systems when it comes to the stuff you buy (as opposed to the stuff you sell).  Further, our industry has failed to develop true best practices on how that data should look, and what should be done with it.  Spend cubes, supplier profiling and UN categories don’t help drive decisions because they only look back at the past, without providing insight into the future.  Information quickly becomes obsolete, and without continuous updates and access to real-time changes, it becomes irrelevant, leading us to bad decisions and manual processes.
It’s clear that the power of predictive analytics can be used to help supply chain and strategic sourcing departments improve the value they bring to their internal (and external) customers.  Until we as an industry can start gaining clarity on what types of data we should be capturing and how we can use it, we will never evolve to a point where these tools are able to help us.

In my next post, I will detail what we need to make the shift to predictive analytics in spend management, including developing a visionary (but also realistic) strategy, hiring the right talent to continuously manage the data (Hint: don’t include the word “buyer” or “contract” in the job description) and most importantly, getting the C-Suite to agree to the investment.
Share To:

Joe Payne

Post A Comment:

1 comments so far,Add yours

  1. Great post!

    Although I definitely see the value of 'predictive analytics', I think the definition itself for the content of the blogposts is to narrow. It’s just scenario 3 in which there is something about 'predictive analytics': ‘provides a forecast out for the year based on’.
    I think 'predictive analytics' is (a very important) part of a greater whole, which has to do with: Operational decision making in procurement based on realtime information and strategic decision making based on predictive analytics.

    For more information click on my link.