As an expert in optimization and analytics, I’m regularly asked, with respect to spend analysis, where do I start, which solution is best, how do I get fast savings, etc. All good questions, and as per the dozens of posts on this subject over the years, they all have multiple answers, highly conditional on the skill of the analyst, the solutions available, the data available for the solutions to act on, the categories that can be impacted in the short term (as some contracts are effectively unbreakable and some vendor contracts all but prevent cost recovery), etc.
More sophisticated analysts ask what built-in reports are necessary and the most useful, how much real-time slicing and dicing capability they really need, how much can be automated, etc. Also good questions, but questions which, depending on the particulars of the situation, also have multiple answers which are dependent on what the organization is buying, who the organization is buying from, where the products and services are coming from and going to, how much data is available with respect to the product and service cost components and composition, how powerful the data amalgamation and computation engines are, how variable their spend patterns are, etc.
Which leads many people to believe that it’s almost impossible to get it right, when the reality is that, with the right type of tool and the right mind set, it’s easy to find success, but very hard to predict where and when you will find success. It’s like being a skilled professional during a market boom where the demand for individuals with your skills outstrips supply. While you don’t know who will offer you a job, if you have good references, work hard, and apply yourself to finding a job, offers will come. Spend analysis often works the same way – if you have the right tools, work hard, and apply yourself to finding opportunities, they will present themselves. It’s just a matter of digging until they materialize.
As I’ve said time and time again, you’re not likely to find your best opportunities in the canned top-N reports which for many years were the staple of spend analysis solution vendors. Because, as I’ve pointed out for years, if you go to manufacturing and ask a handful of buyers who the top 10 suppliers are, you’re going to get a set of overlapping lists which will easily allow you to identify the top 7 or 8 suppliers. The same for categories and commodities. And the AP system tells you who the top departmental and individual spenders are. You may not know the exact spends or exact volumes, but you can quickly get a pretty good idea and focus your negotiation efforts on those categories while you work towards improving your data. The net result is that these reports will only identify a few opportunities that you missed, which you will quickly identify, attack, and capture within 6 to 18 months. That’s why the year-over-year return of most spend analysis efforts falls flat within two years. The biggest opportunities are not in what you know, they’re in what you don’t know. They’re somewhere in the next 20 or 30 suppliers or the next 20 or 30 commodities that can be consolidated into moderately high spend categories that when effectively rationalized and sourced can deliver 10%, 20%, and even 30% savings as well as process efficiencies that deliver further organizational savings still.
A consolidation that will only happen if you can define category groups and supplier families and analyze collective spend and compare it to market rates and identify patterns of overspending not detectable using only top n spend reports or traditional organizational categorizations and supplier codes (where the same supplier has four different entries in the vendor master because, as a non-top n supplier, no one bothered to make sure all spend associated with that supplier was with a single vendor entry in the ERP). And you can only do this if you can slice, dice, and rearrange the data on the fly in non-traditional ways only you, as a creative intelligent being, can do. A rule-based system can only check for previously identified transgressions. It can’t check for unknown transgressions or future transgressions by organizational buyers.
So if you want a short answer to the question of how do you get spend analysis right, you make sure you buy a tool that gives you a lot of different flexibility in the amalgamation, organization, and analysis of the data available to you. It should let you build, in technical terms, “multiple cubes”, allow you to create multiple roll-ups and reports on these cubes, using multiple analysis techniques. It should let you build a cube, appropriate roll-ups, and reports that answer any question you care to ask. If you see a spending pattern or trend that is not in line with what is expected, you need to be able to dive into the data and find out why. That’s how you save money. Flexibility, a toolkit of techniques, and the creativity to apply them in different ways until you stumble on the big savings or value generation opportunity that everyone else has missed.