 Live from New York, it's theCUBE, covering Inforum 2016, brought to you by Inforum. Now, here are your hosts, Dave Vellante and George Gilbert. Welcome back to New York City, everybody. This is theCUBE. We're here live at the Javits Center. Lee Martin is here. She's the Senior Director of Infor Dynamic Science Labs. Lee, thanks for coming on theCUBE. Thanks for having me. So we've heard a lot about this kind of science, cool name, dynamic science lab. Tell us about it. Sure, that's true. On main stage this morning, we were mentioned in several areas. So we are a team of data scientists, essentially. We're based in Cambridge, Massachusetts, right near MIT. We recruit heavily from the local schools, including MIT and Harvard. We have about 20 scientists in our team today. And they are typically PhDs or masters in operations research, applied math, statistics, those types of areas. Right, so we always talk about theCUBE, about what makes a data scientist, and you mentioned those areas. And also, just the ability to do, you know, multitask in those areas, and a love of data, and data wrangling, and. Yeah, yeah, so we do, we typically start our engagements with customers. We do that through a proof of concept phase. So if we're developing a new product, or we're trying to do a new science-based enhancement, we work with customers, and the first thing we want to do is bring in data. So we'll identify a particular problem in an industry that we're trying to solve. We'll speak to a variety of customers. We'll find a few good matches in there where there's a customer that is interested in working with us on a proof of concept and giving us their data. And then we'll work with them over a series of a few months to look at their data. And that's always the first step, right, is getting an understanding of their data. And for us because we work across all the industries at Infor, we're touching lots of different types of data. So the folks in our team don't necessarily come with those industry backgrounds, and some of it's learning as we go. So getting access to the data is a way for us to start understanding that industry. I remember when hearing a story in one of the big data conferences that we were doing, in the early days of the automobile, the industry elite were concerned that there wouldn't be enough, you know, what we now call chauffeurs to drive people around. Because, you know, the people, the way they were transported was somebody would, you know, in a horse and carriage, would be driving, right? So there weren't enough chauffeurs to go around. Then they applied that little story to data scientists. Like, there's not enough data scientists to go around, so it's a barrier to actually growth in that business. This notion of, you know, data science for the many, the citizen data scientists, all kinds of cliches. Where are we at with that? Is it really still a community of just high powered data scientists? Can we permeate that throughout organizations? It's a good question, because there is a lot of competition for data scientists out there right now. It's really hard to actually recruit folks to come into our team. And, you know, I've been, prior to coming to Infor, I worked at other analytics teams. So I've been around in the industry for a while now. And certainly when I first started, data science was sort of a little bit more of a, it felt like a niche field. But now it's everywhere, and it's very broad. So when you hear the terms data scientists and big data, it actually, people use it in a variety of different ways and in a very broad meaning. So for some people, it can mean BI related things. For us, we think of it more of a traditional, really data oriented using predictive analytics, using mathematical modeling, optimization forecast. So, you know, new methods and old methods. So forecasting has been around for a long time. And machine learning is relatively new, you know, sort of in the scale of how people talk about things. So it definitely has grown over time, and we really do see a big competition for it now in terms of recruiting, and a lot of customers asking for it. So it definitely has grown. I just want to follow up on Dave's question about the demand for data scientists. McKinsey did that study a few years ago, so we're going to be many hundreds of thousands, you know, short. And if you want, you know, to work with the data scientists, you're going to have to work with us or IBM or, you know, because they're all going to work for us. How far along are we in using machine learning to help the data scientists become more productive? I think machine learning is one of those things we've definitely made progress on. My sense is that it also has become a bit of a buzzword. And so it's, the term is thrown around a lot. There are some places where really machine learning is happening. In fact, we have a project that we have been working on for the past couple of months that ultimately down the line, we'll use machine learning. But you got to start somewhere, right? Especially with a customer who's trying to solve a very specific problem, you want to start with solving that problem and then branch out. And maybe machine learning isn't the first thing you bring to the table, because it tends to be slightly more advanced than some other techniques. So I do think it's being used. I do think, you know, it's a great tool to have in your tool belt. And I think it can bring a lot to the industry, but I also think it's, you know, there's a lot of use of the term and it's not always sort of what we think of it as in a more traditional sense. It feels like Infor is trying to be very deliberate about the dynamic science lab in that it's not just theoretical. You're trying to actually embed it into applications to have a productive outcome. I wonder, I mean, it seems blatantly obvious, but maybe it didn't just happen like that overnight, but can you talk about the intent and how you're actually purposefully driving innovation through your products, ultimately to your customer? Right, so when Duncan introduced dynamic science labs back in New Orleans, he was talking about the concept that Infor, with its customers, has access to such a large amount, such a large amount of data. So if you think of all the data and ERPs and the edge apps and other areas that we have, and our customers want to harness that. So we have really spent a lot of time understanding where issues exist with customers, so things that they want to really focus on, problems in their business, areas that they want to improve on, and really trying to start with those and then building out a solution from there. So this year we've gone live, we've gone GA on three solutions out of our team. Now we're a team of 20, we're now 20 folks. So having three solutions out there is very exciting for us, which is probably why we were mentioned a lot on main stage today. So we have an inventory intelligence for healthcare solution, we have a CRM sales logic solution, and we have pricing science for distribution. So we have other projects beyond that that we are working on, but those are three core solutions that we're really excited to have those available here today, but they're still in early stages. So in all cases we're sort of working with early adopters who are helping us to refine our approaches and our algorithms, but our goal is really to try to help solve those particular problems using science. So let's take one as an example, take the CRM sales logic. Everybody's familiar with CRM, something that resonates. How would you anticipate that DSL will differentiate in for and add value for customers in ways that plain vanilla CRM will? Well, so in the case of CRM, I think we are bringing, this is the first step, right? As in a lot of the solutions that we're starting with, let's get down some basics, and then let's move on to the more complicated things. So I was talking about machine learning before, in the case of CRM sales intelligence, we're starting with next likely purchase and lead scoring. Now that's not something that nobody else has, there are solutions out there that have that, but what we really want to do is bring our in for customers to advanced, to the sort of arena of advanced math and science so that eventually we do get ahead of others. But in some cases, we're really trying to sort of start with what are the things that customers want today, and that's something in the CRM space that they felt like they were missing, and so we're able to go in and add that functionality. So in the case of CRM, we're going to build on that, and we're going to get more advanced, but we have a good solid platform to start with now. A lot of it too is the data itself, the quality of the data, but it starts with the premise that basically you said, Duncan sort of had this premise when you guys launched this two years ago, of we have a ton of data. Are you finding that you've got the right data, and is the quality of that data in a form that is acceptable, or is that big part of the effort? Determining that and wrangling that, or maybe talking about that a little bit? Certainly managing data is a really, and cleaning data for the purposes that we need to use it for is really difficult. So we do spend a lot of time with customers, and oftentimes we'll go visit with customers who want to do a project, and their question isn't actually, well can we work with you on a project? Their question is what data do I need to start thinking about now for this project that I want to do in the future? So that's in particular when you think about IoT, that is a really hot question that we get from people, is what kind of things should I be gathering around IoT and thinking about in terms of IoT so that you have something to analyze down the road? I want to get back to this question of how much data, and is it an advantage? By way of relaying something I heard from one of the leading AI researchers at Stanford, where she became part of Toyota's autonomous driving program. I mean, she isn't going to work for them, but she's part of their funded institute. She said she thought they would be a leader because you just put cameras and sensors on Toyota cars, they'll have more data than just about anyone else. And so my question then, taking that to loss and for is will you have, let's say from the CRM data, will you have more data that's relevant for this next best offer than anyone else? Or understanding that more data is more important than better algorithms, how do you make that advantage yours? Yeah, so I am a fan of the idea of knowing where you're going before you start down the data science path. I mean, I think, I sort of think of it, there's two ways you can go. You can sort of say, here's all my data, go see if you can find something interesting. And the other thing is, here's this problem I'm trying to solve, analyze this data, and let's see if we can answer the question. So I think most of our work falls in the second camp of what's the business problem you're trying to solve, and do you have the data to support that? This other path is very valuable, but it's longer, it's much more expensive, it's more time intensive to do. But I think, I guess what I'm trying to ask is if you know you want to go after this next best offer or something equivalent, do you compare the data you are likely to be able to bring to bear on that modeling exercise? Do you compare that with competitors on the assumption that the more data, the better. That someone who has big data advantages even if they have inferior algorithms, they will probably have an advantage. It's a fair question. I do think that at the out start, at the start, you're probably, you have a smaller data set perhaps because you're trying to answer this very specific question, but I do think the advantage of Infor is that if you have a suite of Infor products and you see down the road that you're able to pull in that the data from those various products, then you have the ability to look across all of those products and start to get a bigger picture of things. Have more context. Exactly, so I think to start with, we're trying to solve specific problems and we often know or we find out during the proof of concept phase what the data is to solve that problem, but I do think down the road, we can take in data from all those different Infor applications and combine it and find interesting things that customers today can't do and especially if you're using solutions from lots of different vendors, it will be hard to get to the point where you can put all that data together. That's the advantage about Infor is if you have all these solutions, you have the opportunity to bring that data together. Get value out of the existing data. There's nothing to stop you in the future from bringing in external data sources Exactly. Leveraging all the API that are out there, but want to start with what you've got. Start with what we have, exactly. The data that our customers have today is the best and easiest thing to get at. So are you prioritizing which are the problems that you'll probably be most successful with based on the primary data and the context or do customers come to you and say these are the highest value problems? Yeah, it's a combination of us talking to customers and saying so inventory intelligence for healthcare, for example. The senior scientist who designed that solution, Don Rose, she and I spent six months when we first joined Infor two years ago and speaking with a whole bunch of loss in customers to understand all the challenges that they have, things they'd like to see or like to work on. And we came up with this whole list of things and we decided inventory optimization was a good place to start because we had some experience there from our retail, prior retail experience. And so we could bring some knowledge to that space even though neither of us were healthcare experts. Excellent, so okay, so I guess you guys do Infor every two years, right? I'm waiting for the annual, but that's cool. So it gives you enough time to sort of introduce things and then bake them out. So we're seeing a lot of progress from New Orleans. So okay, so at Infor 18, what should we be looking for? For DSL? Well, I know the announcement came out today about IoT and I think we are not the sole group thinking about IoT at Infor, but it's certainly an area that will be interesting for us down the line once we start working with customers who have some IoT data. I think RFID, we're talking to customers about RFID projects and similar to IoT, how can you use our RFID information? We talked about I think the network aspect of things. So we are in fact working with GT Nexus right now. It was sort of a subtlety of the presentation today on main stage, but we are doing a proof of concept with them around ETAs. So being better able to forecast when an item will leave one continent and arrive at the other continent. So it's a first step of our working relationship with them, but we are going to be working with them on a roadmap to move beyond that. And so I think we'll see even more advanced projects going on around network and GT Nexus. And are there any natural industry affinities that you're finding? So one of the great things about our team is the fact that we work across industries, right? So there's a lot of science going on at Infor, not just in dynamic science labs. We happen to get a lot of mention on main stage, but there's a lot of science going on outside of just our team. And there is the ability to take one concept and bring it to another area. So whether we can take inventory optimization for healthcare and modify that to do spare parts optimization in manufacturing or distribution, we definitely see those kinds of opportunities down the road. Dave, you better take one more because then- Okay, we're out of time. So if you want to get one more in, now's your chance. All right, we got to wrap. Last thoughts, Lee, on Inforum 16. Some of the things that customers are talking about. Because what's the reaction been to dynamic science labs? Well, I was really pleased to see that we had a session immediately after the main stage this morning and we had a full house. We had some people standing in the background. So I think it's really exciting, like I said, to have these three solutions out there, have our team be mentioned on main stage, being able to really do some good work with our customers. We have several customers here presenting this week about the work we've done with them. So that's great to see and we're really excited about the next steps. Great, all right, well thanks, Lee, for coming on theCUBE. All right, thanks for having me. We appreciate it. All right, keep it right there. Everybody will be back with our next guest right after this brief break.