One of the most critical functions in organizations, but often less appreciated than it should be, is the Procurement function. Somehow it rarely has the glitz and glamour associated with its more modish counterparts like Marketing, Finance or Technology.
Probably for this same reason, it is frequently a laggard in the adoption of cutting edge technologies when compared to other groups. While revenue generating strategies and initiatives get most of the limelight, it is important to acknowledge that cost savings can deliver exactly the same impact to bottom-line and profits.
For the uninitiated, Procurement function deals with the end-to-end process of buying every single item that is required to run a company, from a paper clip to heavy machinery, from office snacks to transport providers, from computer hardware to large data center equipment. To put it another way, everything on the cost side of a company’s P&L that deals with ‘cost of doing business’, except for personnel costs (aka salaries and wages), comes under the purview of Procurement organization.
Procurement’s role has several components — not only does it need to source all of these small and big things, it also has to do it in the most efficient manner. This includes things like finding the best quality and best price for each item, assessing suppliers across multiple parameters, estimating demand at an item level, managing inventory optimally, planning logistics, and keeping track of what’s happening in the broader marketplace.
Think about what happens when you need to make an important purchasing decision like buying a new television. Most people would spend a significant amount of time in researching, comparing features, getting multiple quotes, talking to different sellers, comparing notes with friends. Imagine doing this for a whole plethora of items!
Now let’s get back to the topic of technology adoption within Procurement, particularly data related technologies. Humor me for a second and try to answer the following questions for me:
- How much did you spend on eating-out or takeouts in the last month?
- Was this amount greater or smaller than what you spent on eating in the month before last? By how much?
- How much did you spend in tips and taxes?
- Did you spend more on coffee or dessert during this period? How much more?
- Where did you get the best price per ml of coffee?
- How much did you spend in cash vs. credit?
They are all pretty simple, straightforward questions, aren’t they? But how difficult would it be to answer them accurately? My guess is, quite difficult! You need to get together records of all expenses incurred in the last month and previous month. They are likely in different places — receipts, credit card bills, bank statements, even your memory. They are definitely in different formats. You probably need to do some triangulation across receipts and statements. There are potentially different nomenclatures, e.g. clubbing all pies, cakes, ice-creams etc. under ‘dessert’, and so on.
Now scale this up — many, many times over — across tens of thousands of items (that’s usually how many unique items a mid to large sized company buys); across cities, countries, continents; across tens of thousands of suppliers; across millions of lines of individual transactions; across different languages; across different nomenclatures and standards, across hundreds of millions or billions of dollars!
This is what Procurement functions deal with. That is why data is so core. And that is why leveraging AI or Data Science is absolutely, unequivocally critical. Unfortunately, that last piece is where most Procurement organizations are still playing catch-up.
We all know about ERP (Enterprise Resource Planning) systems — large, often monolithic, systems that enable organizations to manage their core business processes like procurement, inventory, sales, accounting, HR etc. They have been around for almost three decades and have played a key role in enabling companies to capture data. Back in the day Procurement functions were at the forefront of the adoption of traditional ERP systems, but they have struggled to make the next leap as far as technology goes.
While Procurement practitioners have access to loads of data, the challenge is this data is notoriously dirty. It comes from an amalgam of systems (different ERP platforms, legacy systems, even excel sheets) that have different structures, terminologies and granularity. Even where the source system has been streamlined, ERP data is fraught with problems — missing information, incorrect information, manual entry errors etc. And these errors multiply and mushroom overtime as businesses evolve, getting magnified to a degree where any analysis done on the data ends up literally being a case of ‘Garbage In — Garbage Out’.
So what happens ultimately is that Procurement managers end up needing to spend an inordinate amount of their time in just wrestling with this data, trying to eke out some semblance of an insight that they can use to improve things. Very little of their time is then left to actually go out into the market and capture the potential benefits.
Enter Artificial Intelligence, and the equation changes entirely! A well thought out strategy to leverage AI in procurement can unlock tremendous ROI for Procurement functions. They can have the data wrangling taken care of through sophisticated Machine Learning algorithms. They can leverage Natural Language Processing (NLP) techniques to get precious gems of information hiding in the free text fields of data capture systems, which can literally be a gold mine! They can have insights delivered to them day in and day out around where the savings and other opportunities are hiding, so they can focus solely on going after them and delivering bottom-line, hard dollars, impact.
Source: Towards Data Science