Giana Manzi: William thanks so much for joining today. I know we are here to discuss all things procurement analytics and data. So when you think about where the biggest gaps lie in procurement analytics, where do you find that biggest gap to be?
William Cheung: I would say that it depends on the organization and there's what I would say are two very distinct phases in the evolution of procurement analytics for a company. The first one is some of your more fundamental phase for the mental space, I would call it where you're really trying to understand how much spend do we have and there's various approaches, right? You could be scrappy where you get all your data feeds as raw as possible and then you could do a very rough analysis based on GL accounts, and, by cost center you could almost get an idea of what kind of spend you could get. There's really a lot of tools these days that allow you to manipulate very large sets of data at a desktop level. So the investment to get to that first bump where you have something, it's not a very large investment. And then as you progress through that evolution you get to a point where now you have an actual system, a visibility system that allows you to compile all these different data sources to help you categorize that spending. And once you have that category level of data that will be the top of that phase one of data analytics where you understand at least at a high level how much spend, where the spend is, and what the vendors are for for those categories. Now the next level that I would refer to is what I really consider the advanced analytics phase.
Giana Manzi: Okay in terms of opportunity—where do you see kind of an underutilized set of data? What do you think people aren't looking at as much as they should to give them really valuable insights?
William Cheung: Yeah, you know. Interesting question and I think operationally I would say that the most underutilized datasets are the ones that don't necessarily translate to hard savings on a tracker right as part of your sourcing pipeline. You understand spend and then if you address that aerial spend you get x amount or percentage in savings, but as we think about everybody that's involved in the P2P process… The most overlooked data set is anything related to that P2P process, right? So if you're talking about how long does it take for a requester to create a PO? How long does it take for a requisition to be approved and properly processed so that you can turn into a purchase order? What it really costs for a company to receive a paper invoice and translating into a payment including payment terms and all the available possibilities with dynamic discounting and discounted terms? So all this data, it's there in this system but to be able to capture all those elements relating to each individual transaction and painting the bigger picture with it, it's something I definitely see a great opportunity out there. It's, unfortunately, not always the focus because it's very large data is difficult to interpret and ultimately it involves changing people's behavior. So it involves change management and depending on the size of the organization it could be quite a process to implement that change management. So, it goes beyond just the analytics skills to be able to put that together, but I definitely think that the better the analytics are in that regard it will make the business case easier to implement.
So it goes both ways. They share responsibility to make that happen.
Giana Manzi: Right, right. Okay, so kind of going beyond the traditional procurement and sourcing metrics in terms of cost savings and actually finding a way to quantify efficiencies and you know kind of that human element of procurement that definitely poses its own challenges. Yep.
Where do you see things going in terms of procurement analytics, procurement analytics technology? Have you seen any sort of trends more recently that have been interesting to you and you hope to see continue?
William Cheung: Sure, I would say that at a very high level one of the great things about procurement is that, or being in procurement analytics, is that there is not..there isn’t need for… I'm going to say this now and then later on I'm going to contradict myself but I'll say first… is that there's really no specific procurement tools that you need, with all the different BI options are out there. Really you just have to pick one and go deep with that, right? Then, in addition to BI, I would say not just the visualization part but most importantly the data manipulation part. Long gone are the days where you open an Excel file by which is not only limited by number of rows, but once you start putting a lot of vLOOKUPS and various formulas into into the Excel workbook. It becomes very burdensome on the desktop. Now you have tools where you can combine and manipulate and apply formulas to very large data sets within a workflow and it's at the desktop level. So you have a lot more manipulation power than we have before and that has been a great helper in producing various types of analytics. Now I'm going to contradict myself in that the same way that there's very simple technology or not simple, but very straightforward technology that you can access to quickly ramp up your ability to produce analytics.
The next level release is in the AI and now we're talking about very specific tools that are procurement, that target procurement, right? So procurement specific tools that use AI to simplify and almost automate that task flow, as I was explaining before looking at various categories and realizing for example that you have too many vendors, but then each category might have slightly different thresholds of what you would consider “many vendors.” So to be able to automate some of that analysis and let the machine learning for example say okay, so for something that's more commoditized like computer parts, for example, anything more than one is really too many vendors versus something very specific like Legal Services where you have litigation and the areas of expertise. You have firms that are very narrow in their area of expertise, so one hundred, two hundred, a thousand vendors might not be too many. So if your tool can understand that and really guide the category managers towards where some of the more tangible or actual opportunities are. I really see that as the future. And it really depends on the level of evolution for the company and where they are in that maturity level.