I will leave for another time the discussion about Big Data and how to best manage it – for the present let’s just assume you’ve got it and that it’s under control. You’ve got your enterprise data warehouse and your big honking data cube and your one version of the truth. Perhaps you are, as John Miller of Arkonas puts it in his recent “One Eighty” newsletter, a “data hoarder”, and it’s time for you to make use of that data by becoming an Aspiring, or even a Strategic, Data User.
Just as there’s gold in them thar hills, there’s insight hidden in that data, insight that can make a difference, and there are some pretty straightforward applications that can have quite an immediate impact.
To introduce the first of my insight suggestions I will need to tell you about my new girlfriend at SAS. Her surname is SAS® Operations Research, but to distinguish her from her equally attractive sisters I need to also share with you her first name: SAS® Inventory Optimization. I’m still friends with my old SAS girlfriend, SAS® Forecast Server, but ever since SAS embedded its high performance forecasting capabilities in my SAS® Financial Management solution (see my post, “The Future isn’t what it used to be”) I just haven’t been seeing as much of her as I used to - what the Customer Intelligence people call “churn”.
Excellence in working capital management is one of the touchstones of good financial management, and half of that working capital battle lies in inventory, where it’s about getting those inventory turns down to best-in-class for your industry. Trying to accomplish this just from the office of finance is like trying to battle back from an 0-and-2 count against Sandy Koufax – you just don’t have many options/levers/drivers at your disposal outside of withholding inventory purchase authority. Once the excess inventory has entered the channel there is not much you can do but wait until the snake has digested its meal. Improving inventory turns is a challenge best addressed by getting the right tools (inventory optimization) into the hands of the supply chain team who can prevent the problem from developing in the first place.
Inventory optimization is one of those extremely powerful analytic tools that truly belongs in every organization where inventory management is more than just a trivial exercise. There is so much money, so much working capital, and so much physical inventory in play that even small improvements, measured in tenths, can have a big impact on cash.
I was introduced to my new paramour at a recent presentation to a large U.S. financial institution. The impressive demonstration came in two parts. First, the live demo showed the example of a bank needing to keep hundreds of ATM’s in a metropolitan area filled with cash (what a bank would call ‘inventory’), each ATM with its own unique hourly usage profile, minimum cash threshold, and current cash balance, with further constraints coming from the number and location of the armored trucks available to service the machines.
The calculation from the linear programming algorithm came back in under thirty seconds, having computed the optimal timing for reloading each of the 250 machines for the entire week. Diving into the details, you get to see both how powerful and un-intuitive this analytic tool is, and how futile it would be to try to attempt this any other way. Many ATMs are refilled long before they get anywhere near their minimum threshold, because to do otherwise might mean not being able to get back to that one soon enough the next time to prevent a cash-out, or having some other machine run out while this one is being unnecessarily serviced twice.
Next comes the distribution routing diagram, showing the optimal route for each truck to meet the above mentioned criteria and constraints. Click on the map and the model shows you visually, hour-by-hour, where each truck is, what its route is that during particular hour, and what its route for the entire shift looks like.
If you are wondering what the cash “inventory” levels look like in this example, we’re talking in the range of $50 million or so at any point in time, give or take. Multiply that statewide, region-wide or nationwide and you can immediately appreciate my prior point about how impactful the improvement of even a tenth of a turn would be. Given a reliable front-end forecast, the rule of thumb for applying inventory optimization is in the range of a 5-15% improvement; somewhere between a half a turn and one turn depending on your current performance (optimizing an unreliable forecast input is simply a fruitless exercise).
An impossible task without inventory optimization software, but just another day at the office for this bank’s operations team. Furthermore, given the variability in the initial assumptions, forecast, inputs and constraints, the output isn’t merely good, or better – it’s optimal! How often can you say that about any part of your organization? The data is available, the tools are available; it’s just a matter of time before every organization, no matter what its inventory consists of - cash, computers, clothing or coal - habitually employs operations research/optimization analytics as part of its tactical arsenal.
By Leo Sadovy, EPM Contributor, from: http://blogs.sas.com/content/valuealley/2012/10/09/insight-from-analytics-inventory-optimization/?utm_source=feedburner&utm_medium=email&utm_campaign=Feed%3A+ValueAlley+%28Value+Alley%29