Monday, November 21, 2016

Meet George Jetson – your new AI-empowered chief procurement officer

The next BriefingsDirect technology innovation thought leadership discussion explores how rapid advances in artificial intelligence (AI) and machine learning are poised to reshape procurement -- like a fast-forwarding to a once-fanciful vision of the future.

Whereas George Jetson of the 1960s cartoon portrayed a world of household robots, flying cars, and push-button corporate jobs -- the 2017 procurement landscape has its own impressive retinue of decision bots, automated processes, and data-driven insights.

We won’t need to wait long for this vision of futuristic business to arrive. As we enter 2017, applied intelligence derived from entirely new data analysis benefits has redefined productivity and provided business leaders with unprecedented tools for managing procurement, supply chains, and continuity risks.

Listen to the podcast. Find it on iTunes. Get the mobile app. Read a full transcript or download a copy.

To learn more about the future of predictive -- and even proactive -- procurement technologies, please welcome Chris Haydon, Chief Strategy Officer at SAP Ariba. The discussion is moderated by BriefingsDirect's Dana Gardner, Principal Analyst at Interarbor Solutions.

Here are some excerpts:

Gardner: It seems like only yesterday that we were content to gain a common view of the customer or develop an end-to-end bead on a single business process. These were our goals in refining business in general, but today we've leapfrogged to a future where we're using words like “predictive” and “proactive” to define what business function should do and be about. Chris, what's altered our reality to account for this rapid advancement from visibility into predictive -- and on to proactive?

Haydon: There are a couple of things. The acceleration of the smarts, the intelligence, or the artificial intelligence, whatever the terminology that you identify with, has really exploded. It’s a lot more real, and you see these use-cases on television all the time. The business world is just looking to go in and adopt that.

And then there’s this notion of the Lego block of being able to string multiple processes together via an API is really exciting -- that coupled with the ability to have insight. The last piece, the ability to make sense of big data, either from a visualization perspective or from a machine-learning perspective, has accelerated things.

These trends are starting to come together in the business-to-business (B2B) world, and today, we're seeing them manifest themselves in procurement.

Gardner: What is it about procurement as a function that’s especially ripe for taking advantage of these technologies?

Transaction intense

Haydon: Procurement is obviously very transaction-intense. Historically, what transaction intensity means is people, processing, exceptions. When we talk about these trends now, the ability to componentize services, the ability to look at big data or machine learning, and the input on top of this contextualizes intelligence. It's cognitive and predictive by its very nature, a bigger data set, and [improves] historically inefficient human-based processes. That’s why procurement is starting to be at the forefront.

Haydon

Gardner: Procurement itself has changed from the days of when we were highly vertically integrated as corporations. We had long lead times on product cycles and fulfillment. Nowadays, it’s all about agility and compressing the time across the board. So, procurement has elevated its position. Anything more to add?

Haydon: Everyone needs to be closer to the customer, and you need live business. So, procurement is live now. This change in dynamic -- speed and responsiveness -- is closer to your point. It’s also these other dimensions of the consumer experience that now has to be the business-to-business experience. All that means same-day shipping, real-time visibility, and changing dynamically. That's what we have to deliver.

Gardner: If we go back to our George Jetson reference, what is it about this coming year, 2017? Do you think it's an important inception point when it comes to factoring things like the rising role of procurement, the rising role of analytics, and the fact that the Internet of Things (IoT) is going to bring more relevant data to bear? Why now?

Haydon: There are a couple of things. The procurement function is becoming more mature. Procurement leaders have extracted those first and second levels of savings from sourcing and the like. And they have control of their processes.

With cloud-based technologies and more of control of their processes, they're looking now to how they're going to serve their internal customers by being value-generators and risk-reducers.

How do you forward the business, how do you de-risk, how do you get supply continuity, how do you protect your brand? You do that by having better insight, real-time insight into your supply base, and that’s what’s driving this investment.

Gardner: We've been talking about Ariba being a 20-year-old company. Congratulations on your anniversary of 20 years.

Haydon: Thank you.

AI and bots

Gardner: You're also, of course, part of SAP. Not only have you been focused on procurement for 20 years, but you've got a large global player with lots of other technologies and platform of benefits to avail yourselves of. So, that brings me to the point of AI and bots.

It seems to me that right at the time when procurement needs help, when procurement is more important than ever, that we're also in a position technically to start doing some innovative things that get us into those words "predictive" and more "intelligent."

Set the stage for how these things come together.

Haydon: You allude to being part of SAP, and that's really a great strength and advantage for a domain-focused procurement expertise company.

The machine-learning capabilities that are part of a native SAP HANA platform, which we naturally adopt and get access to, put us on the forefront of not having to invest in that part of the platform, but to focus on how we take that platform and put it into the context of procurement.

There are a couple of pretty obvious areas. There's no doubt that when you’ve got the largest B2B network, billions in spend, and hundreds and millions of transactions on invoicing, you apply some machine learning on that. We can start doing a lot smarter matching an exception management on that, pretty straightforward. That's at one end of the chain.
It's not about upstream and downstream, it's about end-to-end process, and re-imagining and reinventing that.

On the other end of the chain, we have bots. Some people get a little bit wired about the word “bot,” “robotics,” or whatever, maybe it's a digital assistant or it's a smart app. But, it's this notion of helping with decisions, helping with real-time decisions, whether it's identifying a new source of supply because there's a problem, and the problem is identified because you’ve got a live network. It's saying that you have a risk or you have a continuity problem, and not just that it's happening, but here's an alternative, here are other sources of a qualified supply.

Gardner: So, it strikes me that 2017 is such a pivotal year in business. This is the year where we're going to start to really define what machines do well, and what people do well, and not to confuse them. What is it about an end-to-end process in procurement that the machine can do better that we can then elevate the value in the decision-making process of the people?

Haydon: Machines can do better in just identifying patterns -- clusters, if you want to use a more technical word. They transform category management and enables procurement to be at the front of their internal customer set by looking not just at their traditional total cost of ownership (TCO), but total value and use. That's a part of that real dynamic change.

What we call end-to-end, or even what SAP Ariba defined in a very loose way when we talked about upstream, it was about outsourcing and contracting, and downstream was about procurement, purchasing, and invoicing. That's gone, Dana. It's not about upstream and downstream, it's about end-to-end process, and re-imagining and reinventing that.

The role of people

Gardner: When we give more power to a procurement professional by having highly elevated and intelligent tools, their role within the organization advances and the amount of improvement they can make financially advances. But I wonder where there's risk if we automate too much and whether companies might be thinking that they still want people in charge of these decisions. Where do we begin experimenting with how much automation to bring, now that we know how capable these machines have been, or is this going to be a period of exploration for the next few years?

Haydon: It will be a period of exploration, just because businesses have different risk tolerances and there are actually different parts of their life cycle. If you're in a hyper growth mode and you're pretty profitable, that's a little bit different than if you're under a very big margin pressure.

For example, maybe if you're in high tech in the Silicon Valley, and some big names that we could all talk about are, you're prepared to be able to go at it, and let it all come.

If you're in a natural-resource environment, every dollar is even more precious than it was a year ago.

That’s also the beauty, though, with technology. If you want to do it for this category, this supplier, this business unit, or this division you can do that a lot easier than ever before and so you go on a journey.
If you're in a hyper growth mode and you're pretty profitable, that's a little bit different than if you're under a very big margin pressure.

Gardner: That’s an important point that people might not appreciate, that there's a tolerance for your appetite for automation, intelligence, using machine learning, and AI. They might even change, given the context of the certain procurement activity you're doing within the same company. Maybe you could help people who are a little bit leery of this, thinking that they're losing control. It sounds to me like they're actually gaining more control.

Haydon: They gain more control, because they can do more and see more. To me, it’s layered. Does the first bot automatically requisition something -- yes or no? So, you put tolerances on it. I'm okay to do it if it is less than $50,000, $5,000, or whatever the limit is, and it's very simple. If the event is less than $5,000 and it’s within one percent of the last time I did it, go and do it. But tell me that you are going to do it or let’s have a cooling-off period.

If you don't tell me or if you don’t stop me, I'm going to do it, and that’s the little bit of this predictive as well. So you still control the gate, you just don’t have to be involved in all the sub-processes and all that stuff to get to the gate. That’s interesting.

Gardner: What’s interesting to me as well, Chris, is because the data is such a core element of how successful this is, it means that companies in a procurement intelligence drive will want more data, so they can make better decisions. Suppliers who want to be competitive in that environment will naturally be incentivized to provide more data, more quickly, with more openness. Tell us some of the implications for intelligence brought to procurement on the supplier? What we should expect suppliers to do differently as a result?

Notion of content

Haydon: There's no doubt that, at a couple of levels, suppliers will need to let the buyers know even more about themselves than they have ever known before.

That goes to the notion of content. It’s like there is unique content to be discovered, which is whom am I, what do I do well and demonstrate that I do well. That’s being discovered. Then, there is the notion of being able to transact. What do I need to be able to do to transact with you efficiently whether that's a payment, a bank account, or just the way in which I can consume this?

Then, there is also this last notion of the content. What content do I need to be able to provide to my customer, aka the end user, for them to be able to initiate the business with them?

These three dimensions of being discovered, how to be dynamically transacted with, and then actually providing the content of what you do even as a material of service to the end user via the channel. You have to have all of these dimensions right.
If you don't have the context of the business process between a buyer and a seller and what they are trying to affect through the network, how does it add value?

That’s why we fundamentally believe that a network-based approach, when it's end to end, meaning a supplier can do it once to all of the customers across the [Ariba] Discovery channel, across the transactional channel, across the content channel is really value adding. In a digital economy, that's the only way to do it.

Gardner: So this idea of the business network, which is a virtual repository for all of this information isn't just quantity, but it's really about the quality of the relationship. We hear about different business networks vying for attention. It seems to me that understanding that quality aspect is something you shouldn't lose track of.

Haydon: It’s the quality. It’s also the context of the business process. If you don't have the context of the business process between a buyer and a seller and what they are trying to affect through the network, how does it add value? The leading-practice networks, and we're a leading-practice network, are thinking about Discovery. We're thinking about content; we're thinking about transactions.

Gardner: Again, going back to the George Jetson view of the future, for organizations that want to see the return on their energy and devotion to these concepts around AI, bots, and intelligence. What sort of low-hanging fruit do we look for, for assuring them that they are on the right path? I'm going to answer my own question, but I want you to illustrate it a bit better, and that’s risk and compliance and being able to adjust to unforeseen circumstances seems to me an immediate payoff for doing this.

Severance of pleadings

Haydon: The United Kingdom is enacting a law before the end of the year for severance of pleadings. It’s the law, and you have to comply. The real question is how you comply.

You eye your brand, you eye your supply chain, and having the supply-chain profile information at hand right now is top of mind. If you're a Chief Procurement Officer (CPO) and you walk into the CEO’s office, the CEO could ask, "Can you tell me that I don’t have any forced labor, I don’t have any denied parties, and I'm Office of Foreign Assets Control (OFAC) compliant? Can you tell me that now?"

You might be able to do it for your top 50 suppliers or top 100 suppliers, and that’s great, but unfortunately, a small, $2,000 supplier who uses some forced labor in any part of the world is potentially a problem in this extended supply chain. We've seen brands boycotted very quickly. These things roll.

So yes, I think that’s just right at the forefront. Then, it's applying intelligence to that to give that risk threshold and to think about where those challenges are. It's being smart and saying, "Here is a high risk category. Look at this category first and all the suppliers in the category. We're not saying that the suppliers are bad, but you better have a double or triple look at that, because you're at high risk just because of the nature of the category."
Think larger than yourself in trying to solve that problem differently. Those cloud deployment models really help you.

Gardner: Technically, what should organizations be thinking about in terms of what they have in place in order for their systems and processes to take advantage of these business network intelligence values? If I'm intrigued by this concept, if I see the benefits in reducing risk and additional efficiency, what might I be thinking about in terms of my own architecture, my own technologies in order to be in the best position to take advantage of this?

Haydon: You have to question how much of that you think you can build yourself. If you think you're asking different questions than most of your competitors, you're probably not. I'm sure there are specific categories and specific areas on tight supplier relationships and co-innovation development, but when it comes to the core risk questions, more often, they're about an industry, a geography, or the intersection of both.

Our recommendation to corporations is never try and build it yourself. You might need to have some degree of privacy, but look to have it as more industry-based. Think larger than yourself in trying to solve that problem differently. Those cloud deployment models really help you.

Gardner: So it really is less of a technical preparatory thought process than process being a digital organization, availing yourself of cloud models, being ready to think about acting intelligently and finding that right demarcation between what the machines do best and what the people do best.

More visible

Haydon: By making things digital they are actually more visible. You have to be able to balance the pure nature of visibility to get at the product; that's the first step. That’s why people are on a digital journey.

Gardner: Machines can’t help you with a paper-based process, right?

Haydon: Not as much. You have to scan it and throw it in. Then, you are then digitizing it.

Gardner: We heard about Guided Buying last year from SAP Ariba. It sounds like we're going to be getting a sort of "Guided Buying-Plus" next year and we should keep an eye on that.

Haydon: We're very excited. We announced that earlier this year. We're trying to solve two problems quickly through Guided Buying.
Our Guided Buying has a beautiful consumer-based look and feel, but with embedded compliance. We hide the complexity. We just show the user what they need to know at the time, and the flow is very powerful.

One is the nature of the ad-hoc user. We're all ad-hoc users in the business today. I need to buy things, but I don’t want to read the policy, I don’t want to open the PDF on some corporate portal on some threshold limit that, quite honestly, I really need to know about once or twice a year.

So our Guided Buying has a beautiful consumer-based look and feel, but with embedded compliance. We hide the complexity. We just show the user what they need to know at the time, and the flow is very powerful.

Listen to the podcast. Find it on iTunes. Get the mobile app. Download the transcript. Sponsor: SAP Ariba.

You may also be interested in:

Friday, November 18, 2016

Strategic view across more data delivers digital business boost for AmeriPride

The next BriefingsDirect Voice of the Customer digital transformation case study explores how linen services industry leader AmeriPride Services uses big data to gain a competitive and comprehensive overview of its operations, finances and culture.

We’ll explore how improved data analytics allows for disparate company divisions and organizations to come under a single umbrella -- to become more aligned -- and to act as a whole greater than the sum of the parts. This is truly the path to a digital business.

Listen to the podcast. Find it on iTunes. Get the mobile app. Read a full transcript or download a copy.

Here to describe how digital transformation has been supported by innovations at the big data core, we’re joined by Steven John, CIO, and Tony Ordner, Information Team Manager, both at at AmeriPride Services in Minnetonka, Minnesota. The discussion is moderated by BriefingsDirect's Dana Gardner, Principal Analyst at Interarbor Solutions.

Here are some excerpts:

Gardner: Let’s discuss your path to being a more digitally transformed organization. What were the requirements that led you to become more data-driven, more comprehensive, and more inclusive in managing your large, complex organization?

John


John: One of the key business drivers for us was that we're a company in transition -- from a very diverse organization to a very centralized organization. Before, it wasn't necessarily important for us to speak the same data language, but now it's critical. We’re developing the lexicon, the Rosetta Stone, that we can all rely on and use to make sure that we're aligned and heading in the same direction.

Gardner: And Tony, when we say “data,” are we talking about just databases and data within applications? Or are we being even more comprehensive -- across as many information types as we can?

Ordner: It’s across all of the different information types. When we embarked on this journey, we discovered that data itself is great to have, but you also have to have processes that are defined in a similar fashion. You really have to drive business change in order to be able to effectively utilize that data, analyze where you're going, and then use that to drive the business. We're trying to institute into this organization an iterative process of learning.

Gardner: For those who are not familiar with AmeriPride Services, tell us about the company. It’s been around for quite a while. What do you do, and how big of an umbrella organization are we talking about?

Long-term investments

John: The company is over 125 years old. It’s family-owned, which is nice, because we're not driven by the quarter. We can make longer-term investments through the family. We can have more of a future view and have ambition to drive change in different ways than a quarter-by-quarter corporation does.

We're in the laundry business. We're in the textiles and linen business. What that means is that for food and beverage, we handle tablecloths, napkins, chef coats, aprons, and those types of things. In oil and gas, we provide the safety garments that are required. We also provide the mats you cross as you walk in the door of various restaurants or retail stores. We're in healthcare facilities and meet the various needs of providing and cleansing the garments and linens coming out of those institutions. We're very diverse. We're the largest company of our kind in Canada, probably about fourth in the US, and growing.
Become a Member of myVertica
Gain Access to the Free 
HPE Vertica Community Edition
Gardner: And this is a function that many companies don't view as core and they're very happy to outsource it. However, you need to remain competitive in a dynamic world. There's a lot of innovation going on. We've seen disruption in the taxicab industry and the hospitality industry. Many companies are saying, “We don’t want to be a deer in the headlights; we need to get out in front of this.”

Tony, how do you continue to get in front of this, not just at the data level, but also at the cultural level?

Ordner: Part of what we're doing is defining those standards across the company. And we're coming up with new programs and new ways to get in front and to partner with the customers.

Ordner
As part of our initiative, we're installing a lot of different technology pieces that we can use to be right there with the customers, to make changes with them as partners, and maybe better understand their business and the products that they aren't buying from us today that we can provide. We’re really trying to build that partnership with customers, provide them more ways to access our products, and devise other ways they might not have thought of for using our products and services.

With all of those data points, it allows us to do a much better job.

Gardner: And we have heard from Hewlett Packard Enterprise (HPE) the concept that it's the “analytics that are at the core of the organization,” that then drive innovation and drive better operations. Is that something you subscribe to, and is that part of your thinking?

John: For me, you have to extend it a little bit further. In the past, our company was driven by the experience and judgment of the leadership. But what we discovered is that we really wanted to be more data-driven in our decision-making.

Data creates a context for conversation. In the context of their judgment and experience, our leaders can leverage that data to make better decisions. The data, in and of itself, doesn’t drive the decisions -- it's that experience and judgment of the leadership that's that final filter.

We often forget the human element at the end of that and think that everything is being driven by analytics, when analytics is a tool and will remain a tool that helps leaders lead great companies.

Gardner: Steven, tell us about your background. You were at a startup, a very successful one, on the leading edge of how to do things different when it comes to apps, data, and cloud delivery.

New ways to innovate

John: Yes, you're referring to Workday. I was actually Workday’s 33rd customer, the first to go global with their product. Then, I joined Workday in two roles: as their Strategic CIO, working very closely with the sales force, helping CIOs understand the cloud and how to manage software as a service (SaaS); and also as their VP of Mid-Market Services, where we were developing new ways to innovate, to implement in different ways and much more rapidly.

And it was a great experience. I've done two things in my life, startups and turnarounds, and I thought that I was kind of stepping back and taking a relaxing job with AmeriPride. But in many ways, it's both; AmeriPride’s both a turnaround and a startup, and I'm really enjoying the experience.

Gardner: Let’s hear about how you translate technology advancement into business advancement. And the reason I ask it in that fashion is that it seems as a bit of a chicken and the egg, that they need to be done in parallel -- strategy, ops, culture, as well as technology. How are you balancing that difficult equation?

John: Let me give you an example. Again, it goes back to that idea of, if you just have the human element, they may not know what to ask, but when you add the analytics, then you suddenly create a set of questions that drive to a truth.
Become a Member of myVertica
Gain Access to the Free 
HPE Vertica Community Edition
We're a route-based business. We have over a 1,000 trucks out there delivering our products every day. When we started looking at margin we discovered that our greatest margin was from those customers that were within a mile of another customer.

So factoring that in changes how we sell, that changes how we don't sell, or how we might actually let some customers go -- and it helps drive up our margin. You have that piece of data, and suddenly we as leaders knew some different questions to ask and different ways to orchestrate programs to drive higher margin.

Gardner: Another trend we've seen is that putting data and analytics, very powerful tools, in the hands of more people can have unintended, often very positive, consequences. A knowledge worker isn't just in a cube and in front of a computer screen. They're often in the trenches doing the real physical work, and so can have real process insights. Has that kicked in yet at AmeriPride, and are you democratizing analytics?

Ordner: That’s a really great question. We've been trying to build a power-user base and bring some of these capabilities into the business segments to allow them to explore the data.

You always have to keep an eye on knowledge workers, because sometimes they can come to the wrong conclusions, as well as the right ones. So it's trying to make sure that we maintain that business layer, that final check. It's like, the data is telling me this, is that really where it is?

I liken it to having a flashlight in a dark room. That’s what we are really doing with visualizing this data and allowing them to eliminate certain things, and that's how they can raise the questions, what's in this room? Well, let me look over here, let me look over there. That’s how I see that.

Too much information

John: One of the things I worry about is that if you give people too much information or unstructured information, then they really get caught up in the academics of the information -- and it doesn’t necessarily drive a business process or drive a business result. It can cause people to get lost in the weeds of all that data.

You still have to orchestrate it, you still have to manage it, and you have to guide it. But you have to let people go off and play and innovate using the data. We actually have a competition among our power-users where they go out and create something, and there are judges and prizes. So we do try to encourage the innovation, but we also want to hold the reins in just a little bit.

Gardner: And that gets to the point of having a tight association between what goes on in the core and what goes on at the edge. Is that something that you're dabbling in as well?

John: It gets back to that idea of a common lexicon. If you think about evolution, you don't want a Madagascar or a Tasmania, where groups get cut off and then they develop their own truth, or a different truth, or they interpret data in a different way -- where they create their own definition of revenue, or they create their own definition of customer.

If you think about it as orbits, you have to have a balance. Maybe you only need to touch certain people in the outer orbit once a month, but you have to touch them once a month to make sure they're connected. The thing about orbits and keeping people in the proper orbits is that if you don't, then one of two things happens, based on gravity. They either spin out of orbit or they come crashing in. The idea is to figure out what's the right balance for the right groups to keep them aligned with where we are going, what the data means, and how we're using it, and how often.

Gardner: Let’s get back to the ability to pull together the data from disparate environments. I imagine, like many organizations, that you have SaaS apps. Maybe it’s for human capital management or maybe it’s for sales management. How does that data then get brought to bear with internal apps, some of them may even be on a mainframe still, or virtualized apps from older code basis and so forth? What’s the hurdle and what words of wisdom might you impart to others who are earlier in this journey of how to make all that data common and usable?

Ordner: That tends to be a hurdle. As to the data acquisition piece, as you set these things up in the cloud, a lot of the times the business units themselves are doing these things or making the agreements. They don't put into place the data access that we've always needed. That’s been our biggest hurdle. They'll sign the contracts, not getting us involved until they say, "Oh my gosh, now we need the data." We look at it and we say, "Well, it’s not in our contracts and now it’s going to cost more to access the data." That’s been our biggest hurdle for the cloud services that we've done.

Once you get past that, web services have been a great thing. Once you get the licensing and the contract in place, it becomes a very simple process, and it becomes a lot more seamless.

Gardner: So, maybe something to keep in mind is always think about the data before, during, and after your involvement with any acquisition, any contract, and any vendor?

Ordner: Absolutely.

You own three things

John: With SaaS, at the end of the day, you own three things: the process design, the data, and the integration points. When we construct a contract, one of the things I always insist upon is what I refer to as the “prenuptial agreement.”

What that simply means is, before the relationship begins, you understand how it can end. The key thing in how it ends is that you can take your data with you, that it has a migration path, and that they haven't created a stickiness that traps you there and you don't have the ability to migrate your data to somebody else, whether that’s somebody else in the cloud or on-premise.

Gardner: All right, let’s talk about lessons learned in infrastructure. Clearly, you've had an opportunity to look at a variety of different platforms, different requirements that you have had, that you have tested and required for your vendors. What is it about HPE Vertica, for example, that is appealing to you, and how does that factor into some of these digital transformation issues?

Ordner: There are two things that come to mind right away for me. One is there were some performance implications. We were struggling with our old world and certain processes that ran 36 hours. We did a proof of concept with HPE and Vertica and that ran in something like 17 minutes. So, right there, we were sold on performance changes.

As we got into it and negotiated with them, the other big advantage we discovered is that the licensing model with the amount of data, versus the core model that everyone else runs in the CPU core. We're able to scale this and provide that service at a high speed, so we can maintain that performance without having to take penalties against licensing. Those are a couple of things I see. Anything from your end, Steven?

John: No, I think that was just brilliant.

Gardner: How about on that acquisition and integration of data. Is there an issue with that that you have been able to solve?

Ordner: With acquisition and integration, we're still early in that process. We're still learning about how to put data into HPE Vertica in the most effective manner. So, we're really at our first source of data and we're looking forward to those additional pieces. We have a number of different telematics pieces that we want to include; wash aisle telematics as well as in-vehicle telematics. We're looking forward to that.

There's also scan data that I think will soon be on the horizon. All of our garments and our mats have chips in them. We scan them in and out, so we can see the activity and where they flow through the system. Those are some of our next targets to bring that data in and take a look at that and analyze it, but we're still a little bit early in that process as far as multiple sources. We're looking forward to some of the different ways that Vertica will allow us to connect to those data sources.

Gardner: I suppose another important consideration when you are picking and choosing systems and platforms is that extensibility. RFID tags are important now; we're expecting even more sensors, more data coming from the edge, the information from the Internet of Things (IoT). You need to feel that the systems you're putting in place now will scale out and up. Any thoughts about the IoT impact on what you're up to?

Overcoming past sins

John: We have had several conversations just this week with HPE and their teams, and they are coming out to visit with us on that exact topic. Being about a year into our journey, we've been doing two things. We've been forming the foundation with HPE Vertica and we've been getting our own house in order. So, there's a fair amount of cleanup and overcoming the sins of the past as we go through that process.

But Vertica is a platform; it's a platform where we have only tapped a small percentage of its capability. And in my personal opinion, even HPE is only aware of a portion of its capability. There are a whole set of things that it can do, and I don’t believe that we have discovered all of them.
Become a Member of myVertica
Gain Access to the Free 
HPE Vertica Community Edition
With that said, we're going to do what you and Tony just described; we're going to use the telematics coming out of our trucks. We're going to track safety and seat belts. We're going to track green initiatives, routes, and the analytics around our routes and fuel consumption. We're going to make the place safer, we're going to make it more efficient, and we're going to get proactive about being able to tell when a machine is going to fail and when to bring in our vendor partners to get it fixed before it disrupts production.

Gardner: It really sounds like there is virtually no part of your business in the laundry services industry that won't be in some way beneficially impacted by more data, better analytics delivered to more people. Is that fair?

Ordner: I think that’s a very fair statement. As I prepared for this conference, one of the things I learned, and I have been with the company for 17 years, is that we've done a lot technology changes, and technology has taken an added significance within our company. When you think of laundry, you certainly don't think of technology, but we've been at the leading edge of implementing technology to get closer to our customers, closer to understanding our products.

[Data technology] has become really ingrained within the industry, at least at our company.

John: It is one of those few projects where everyone is united, everybody believes that success is possible, and everybody is willing to pay the price to make it happen.

Listen to the podcast. Find it on iTunes. Get the mobile app. Read a full transcript or download a copy. Sponsor: Hewlett Packard Enterprise.

You may also be interested in:

Monday, November 7, 2016

Swift and massive data classification advances score a win for better securing sensitive information

The next BriefingsDirect Voice of the Customer digital transformation case study explores how -- in an era when cybersecurity attacks are on the rise and enterprises and governments are increasingly vulnerable -- new data intelligence capabilities are being brought to the edge to provide better data loss prevention (DLP).

We'll learn how Digital Guardian in Waltham, Massachusetts analyzes both structured and unstructured data to predict and prevent loss of data and intellectual property (IP) with increased accuracy.

Listen to the podcast. Find it on iTunes. Get the mobile app. Read a full transcript or  download a copy.
 
To learn how data recognition technology supports network and endpoint forensic insights for enhanced security and control, we're joined by Marcus Brown, Vice President of Corporate Business Development for Digital Guardian. The discussion is moderated by BriefingsDirect's Dana Gardner, Principal Analyst at Interarbor Solutions.

Here are some excerpts:

Gardner: What are some of the major trends making DLP even more important, and even more effective?

Brown: Data protection has very much to come to the forefront in the last couple of years. Unfortunately, we wake up every morning and read in the newspapers, see on television, and hear on the radio a lot about data breaches. It’s pretty much every type of company, every type of organization, government organizations, etc., that’s being hit by this phenomenon at the moment.

Brown

So, awareness is very high, and apart from the frequency, a couple of key points are changing. First of all, you have a lot of very skilled adversaries coming into this, criminals, nation-state actors, hactivists, and many others. All these people are well-trained and very well resourced to come after your data. That means that companies have a pretty big challenge in front of them. The threat has never been bigger.

In terms of data protection, there are a couple of key trends at the cyber-security level. People have been aware of the so-called insider threat for a long time. This could be a disgruntled employee or it could be someone who has been recruited for monetary gain to help some organization get to your data. That’s a difficult one, because the insider has all the privilege and the visibility and knows where the data is. So, that’s not a good thing.

Then, you have employees, well-meaning employees, who just make mistakes. It happens to all of us. We touch something in Outlook, and we have a different email address than the one we were intending, and it goes out. The well-meaning employees, as well, are part of the insider threat.

Outside threats

What’s really escalated over the last couple of years are the advanced external attackers or the outside threat, as we call it. These are well-resourced, well-trained people from nation-states or criminal organizations trying to break in from the outside. They do that with malware or phishing campaigns.

About 70 percent of the attacks stop with the phishing campaign, when someone clicks on something that looked normal. Then, there's just general hacking, a lot of people getting in without malware at all. They just hack straight in using different techniques that don’t rely on malware.

People have become so good at developing malware and targeting malware at particular organizations, at particular types of data, that a lot of tools like antivirus and intrusion prevention just don’t work very well. The success rate is very low. So, there are new technologies that are better at detecting stuff at the perimeter and on the endpoint, but it’s a tough time.

There are internal and external attackers. A lot of people outside are ultimately after the two main types of data that companies have. One is a customer data, which is credit card numbers, healthcare information, and all that stuff. All of this can be sold on the black market per record for so-and-so many dollars. It’s a billion-dollar business. People are very motivated to do this.
Learn More About HPE IDOL
Advanced Enterprise Search and Analytics
For Unstructured Data
Most companies don’t want to lose their customers’ data. That’s seen as a pretty bad thing, a bad breach of trust, and people don’t like that. Then, obviously, for any company that has a product where you have IP, you spent lots of money developing that, whether it’s the new model of a car or some piece of electronics. It could be a movie, some new clothing, or whatever. It’s something that you have developed and it’s a secret IP. You don’t want that to get out, as well as all of your other internal information, whether it’s your financials, your plans, or your pricing. There are a lot of people going after both of those things, and that’s really the challenge.

In general, the world has become more mobile and spread out. There is no more perimeter to stop people from getting in. Everyone is everywhere, private life and work life is mixed, and you can access anything from anywhere. It’s a pretty big challenge.

Gardner: Even though there are so many different types of threats, internal, external, and so forth, one of the common things that we can do nowadays is get data to learn more about what we have as part of our inventory of important assets.

While we might not be able to seal off that perimeter, maybe we can limit the damage that takes place by early detection of problems. The earlier that an organization can detect that something is going on that shouldn’t be, the quicker they can come to the rescue. How does the instant analysis of data play a role in limiting negative outcomes?

Can't protect everything

Brown: If you want to protect something, you have to know it’s sensitive and that you want to protect it. You can’t protect everything. You're going to find which data is sensitive, and we're able to do that on-the-fly to recognize sensitive data and nonsensitive data. That’s a key part of the DLP puzzle, the data protection puzzle.

We work for some pretty large organizations, some of the largest companies and government organizations in the world, as well as lot of medium- and smaller-sized customers. Whatever it is we're trying to protect, personal information or indeed the IP, we need to be in the right place to see what people are doing with that data.

Our solution consists of two main types of agents. Some agents are on endpoint computers, which could be desktops or servers, Windows, Linux, and Macintosh. It’s a good place to be on the endpoint computer, because that’s where people, particularly the insider, come into play and start doing something with data. That’s where people work. That’s how they come into the network and it’s how they handle a business process.

So the challenge in DLP is to support the business process. Let people do with data what they need to do, but don’t let that data get out. The way to do that is to be in the right place. I already mentioned the endpoint agent, but we also have network agents, sensors, and appliances in the network that can look at data moving around.

The endpoint is really in the middle of the business process. Someone is working, they're working with different applications, getting data out of those applications, and they're doing whatever they need to do in their daily work. That’s where we sit, right in the middle of that, and we can see who the user is and what application they're working with it. It could be an engineer working with the computer-aided design (CAD) or the product lifecycle management (PLM) system developing some new automobile or whatever, and that’s a great place to be.

We rely very heavily on the HPE IDOL technology for helping us classify data. We use it particularly for structured data, anything like a credit card number, or alphanumeric data. It could be also free text about healthcare, patient information, and all this sort of stuff.

We use IDOL to help us scan documents. We can recognize regular expressions, that’s a credit card number type of thing, or Social Security. We can also recognize terminology. We rely on the fact that IDOL supports hundreds of languages and many different subject areas. So, using IDOL, we're able to recognize a whole lot of anything that’s written in textual language.

Our endpoint agent also has some of its own intelligence built in that we put on top of what we call contextual recognition or contextual classification. As I said, we see the customer list coming out of Salesforce.com or we see the jet fighter design coming out of the PLM system and we then tag that as well. We're using IDOL, we're using some of our technology, and we're using our vantage point on the endpoint being in the business process to figure out what the data is.

We call that data-in-use monitoring and, once we see something is sensitive, we put a tag on it, and that tag travels with the data no matter where it goes.

An interesting thing is that if you have someone making a mistake, an unintentional, good-willed employee, accidentally attaching the wrong doc to something that it goes out, obviously it will warn the user of that.

We can stop that

If you have someone who is very, very malicious and is trying to obfuscate what they're doing, we can see that as well. For example, taking a screenshot of some top-secret diagram, embedding that in a PowerPoint and then encrypting the PowerPoint, we're tagging those docs. Anything that results from IP or top-secret information, we keep tagging that. When the guy then goes to put it on a thumb drive, put it on Dropbox, or whatever, we see that and stop that.

So that’s still a part of the problem, but the two points are classify it, that’s what we rely on IDOL a lot for, and then stop it from going out, that’s what our agent is responsible for.

Gardner: Let’s talk a little bit about the results here, when behaviors, people and the organization are brought to bear together with technology, because it’s people, process and technology. When it becomes known in the organization that you can do this, I should think that that must be a fairly important step. How do we measure effectiveness when you start using a technology like Digital Guardian? Where does that become explained and known in the organization and what impact does that have?

Brown: Our whole approach is a risk-based approach and it’s based on visibility. You’ve got to be able to see the problem and then you can take steps and exercise control to stop the problems.
Learn More About HPE IDOL
Advanced Enterprise Search and Analytics
For Unstructured Data
When you deploy our solution, you immediately gain a lot of visibility. I mentioned the endpoints and I mentioned the network. Basically, you get a snapshot without deploying any rules or configuring in any complex way. You just turn this on and you suddenly get this rich visibility, which is manifested in reports, trends, and all this stuff. What you get, after a very short period of time, is a set of reports that tell you what your risks are, and some of those risks may be that your HR information is being put on Dropbox.

You have engineers putting the source code onto thumb drives. It could all be well-meaning, they want to work on it at home or whatever, or it could be some bad guy.

One the biggest points of risk in any company is when an employee resigns and decides to move on. A lot of our customers use the monitoring and the reporting we have at that time to actually sit down with the employee and say, "We noticed that you downloaded 2,000 files and put them on a thumb drive. We’d like you to sign this saying that you're going to give us that data back."

That’s a typical use case, and that’s the visibility you get. You turn it on and you suddenly see all these risks, hopefully, not too many, but a certain number of risks and then you decide what you're going to do about it. In some areas you might want to be very draconian and say, "I'm not going to allow this. I'm going to completely block this. There is no reason why you should put the jet fighter design up on Dropbox."

Gardner: That’s where the epoxy in the USB drives comes in.

Warning people

Brown: Pretty much. On the other hand, you don’t want to stop people using USB, because it’s about their productivity, etc. So, you might want to warn people, if you're putting some financial data on to a thumb drive, we're going to encrypt that so nothing can happen to it, but do you really want to do this? Is this approach appropriate? People get a feeling that they're being monitored and that the way they are acting maybe isn't according to company policy. So, they'll back out of it.

In a nutshell, you look at the status quo, you put some controls in place, and after those controls are in place, within the space of a week, you suddenly see the risk posture changing, getting better, and the incidence of these dangerous actions dropping dramatically.

Very quickly, you can measure the security return on investment (ROI) in terms of people’s behavior and what’s happening. Our customers use that a lot internally to justify what they're doing.

Generally, you can get rid of a very large amount of the risk, say 90 percent, with an initial pass, or initial first two passes of rules to say, we don’t want this, we don’t want that. Then, you're monitoring the status, and suddenly, new things will happen. People discover new ways of doing things, and then you’ve got to put some controls in place, but you're pretty quickly up into the 90 percent and then you fine-tuning to get those last little bits of risk out.

Gardner: Because organizations are becoming increasingly data-driven, they're getting information and insight across their systems and their applications. Now, you're providing them with another data set that they could use. Is there some way that organizations are beginning to assimilate and analyze multiple data sets including what Digital Guardian’s agents are providing them in order to have even better analytics on what’s going on or how to prevent unpleasant activities?

Brown: In this security world, you have the security operations center (SOC), which is kind of the nerve center where everything to do with security comes into play. The main piece of technology in that area is the security information and event management (SIEM) technology. The market leader is HPE’s ArcSight, and that’s really where all of the many tools that security organizations use come together in one console, where all of that information can be looked at in a central place and can also be correlated.

We provide a lot of really interesting information for the SIEM for the SOC. I already mentioned we're on the endpoint and the network, particularly on the endpoint. That’s a bit of a blind spot for a lot of security organizations. They're traditionally looking at firewalls, other network devices, and this kind of stuff.

We provide rich information about the user, about the data, what’s going on with the data, and what’s going on with the system on the endpoint. That’s key for detecting malware, etc. We have all this rich visibility on the endpoint and also from the network. We actually pre-correlate that. We have our own correlation rules. On the endpoint computer in real time, we're correlating stuff. All of that gets populated into ArcSight.

At the recent HPE Protect Show in National Harbor in September we showed the latest generation of our integration, which we're very excited about. We have a lot of ArcSight content, which helps people in the SOC leverage our data, and we gave a couple of presentations at the show on that.

Gardner: And is there a way to make this even more protected? I believe encryption could be brought to bear and it plays a role in how the SIEM can react and behave.

Seamless experience

Brown: We actually have a new partnership, related to HPE's acquisition of Voltage, which is a real leader in the e-mail security space. It’s all about applying encryption to messages and managing the keys and making that user experience very seamless and easy to use.

Adding to that, we're bundling up some of the classification functionality that we have in our network sensors. What we have is a combination between Digital Guardian Network, DOP, and the HPE Data Security Encryption solution, where an enterprise can define a whole bunch of rules based on templates.

We can say, "I need to comply with HIPAA," "I need to comply with PCI," or whatever standard it is. Digital Guardian on the network will automatically scan all the e-mail going out and automatically classify according to our rules which e-mails are sensitive and which attachments are sensitive. It then goes on to the HPE Data Security Solution where it gets encrypted automatically and then sent out.

It’s basically allowing corporations to apply standard set of policies, not relying on the user to say they need to encrypt this, not leaving it to the user’s judgment, but actually applying standard policies across the enterprise for all e-mail making sure they get encrypted. We are very excited about it.
Learn More About HPE IDOL
Advanced Enterprise Search and Analytics
For Unstructured Data
Gardner: That sounds key -- using encryption to the best of its potential, being smart about it, not just across the waterfront, and then not depending on a voluntary encryption, but doing it based on need and intelligence.
 
Brown: Exactly.

Gardner: For those organizations that are increasingly trying to be data-driven, intelligent, taking advantage of the technologies and doing analysis in new interesting ways, what advice might you offer in the realm of security? Clearly, we’ve heard at various conferences and other places that security is, in a sense, the killer application of big-data analytics. If you're an organization seeking to be more data-driven, how can you best use that to improve your security posture?

Brown: The key, as far as we’re concerned, is that you have to watch your data, you have to understand your data, you need to collect information, and you need visibility of your data.

The other key point is that the security market has been shifting pretty dramatically from more of a network view much more toward the endpoint. I mentioned earlier that antivirus and some of these standard technologies on the endpoint aren't really cutting it anymore. So, it’s very important that you get visibility down at the endpoint and you need to see what users are doing, you need to understand what your systems are running, and you need to understand where your data is.

So collect that, get that visibility, and then leverage that visibility with analytics and tools so that you can profit from an automated kind of intelligence.

Listen to the podcast. Find it on iTunes. Get the mobile app. Read a full transcript or  download a copy. Sponsor: Hewlett Packard Enterprise.

You may also be interested in:

Tuesday, November 1, 2016

2016 election campaigners look to big data analysis to gain an edge in intelligently reaching voters

The next BriefingsDirect Voice of the Customer digital transformation case study explores how data-analysis services startup BlueLabs in Washington, DC helps presidential election campaigns better know and engage with potential voters.

We'll learn how BlueLabs relies on high-performing analytics platforms that allow a democratization of querying, of opening the value of vast data resources to discretely identify more of those in the need to know.

Listen to the podcast. Find it on iTunes. Get the mobile app. Read a full transcript or download a copy.

Here to describe how big data is being used creatively by contemporary political organizations for two-way voter engagement, we're joined by Erek Dyskant Co-Founder and Vice President of Impact at BlueLabs Analytics in Washington. The discussion is moderated by BriefingsDirect's Dana Gardner, Principal Analyst at Interarbor Solutions.

Here are some excerpts:

Gardner: Obviously, this is a busy season for the analytics people who are focused on politics and campaigns. What are some of the trends that are different in 2016 from just four years ago. It’s a fast-changing technology set, it's also a fast-changing methodology. And of course, the trends about how voters think, react, use social, and engage are also dynamic. So what's different this cycle?

Dyskant: From a voter-engagement perspective, in 2012, we could reach most of our voters online through a relatively small set of social media channels -- Facebook, Twitter, and a little bit on the Instagram side. Moving into 2016, we see a fragmentation of the online and offline media consumption landscape and many more folks moving toward purpose-built social media platforms.

If I'm at the HPE Conference and I want my colleagues back in D.C. to see what I'm seeing, then maybe I'll use Periscope, maybe Facebook Live, but probably Periscope. If I see something that I think one of my friends will think is really funny, I'll send that to them on Snapchat.
Join myVertica
To Get the Free
HPE Vertica Community Edition
Where political campaigns have traditionally broadcast messages out through the news-feed style social-media strategies, now we need to consider how it is that one-to-one social media is acting as a force multiplier for our events and for the ideas of our candidates, filtered through our campaign’s champions.

Gardner: So, perhaps a way to look at that is that you're no longer focused on precincts physically and you're no longer able to use broadcast through social media. It’s much more of an influence within communities and identifying those communities in a new way through these apps, perhaps more than platforms.

Social media

Dyskant: That's exactly right. Campaigns have always organized voters at the door and on the phone. Now, we think of one more way. If you want to be a champion for a candidate, you can be a champion by knocking on doors for us, by making phone calls, or by making phone calls through online platforms.

You can also use one-to-one social media channels to let your friends know why the election matters so much to you and why they should turn out and vote, or vote for the issues that really matter to you.

Gardner: So, we're talking about retail campaigning, but it's a bit more virtual. What’s interesting though is that you can get a lot more data through the interaction than you might if you were physically knocking on someone's door.

Dyskant: The data is different. We're starting to see a shift from demographic targeting. In 2000, we were targeting on precincts. A little bit later, we were targeting on combinations of demographics, on soccer moms, on single women, on single men, on rural, urban, or suburban communities separately.

Dyskant

Moving to 2012, we've looked at everything that we knew about a person and built individual-level predictive models, so that we knew each person's individual set of characteristics made that person more or less likely to be someone that our candidate would have an engaging conversation through a volunteer.

Now, what we're starting to see is behavioral characteristics trumping demographic or even consumer data. You can put whiskey drinkers in your model, you can put cat owners in your model, but isn't it a lot more interesting to put in your model that fact that this person has an online profile on our website and this is their clickstream? Isn't it much more interesting to put into a model that this person is likely to consume media via TV, is likely to be a cord-cutter, is likely to be a social media trendsetter, is likely to view multiple channels, or to use both Facebook and media on TV?

That lets us have a really broad reach or really broad set of interested voters, rather than just creating an echo chamber where we're talking to the same voters across different platforms.

Gardner: So, over time, the analytics tools have gone from semi-blunt instruments to much more precise, and you're also able to better target what you think would be the right voter for you to get the right message out to.

One of the things you mentioned that struck me is the word "predictive." I suppose I think of campaigning as looking to influence people, and that polling then tries to predict what will happen as a result. Is there somewhat less daylight between these two than I am thinking, that being predictive and campaigning are much more closely associated, and how would that work?

Predictive modeling

Dyskant: When I think of predictive modeling, what I think of is predicting something that the campaign doesn't know. That may be something that will happen in the future or it may be something that already exists today, but that we don't have an observation for it.

In the case of the role of polling, what I really see about that is understanding what issues matter the most to voters and how it is that we can craft messages that resonate with those issues. When I think of predictive analytics, I think of how is it that we allocate our resources to persuade and activate voters.

Over the course of elections, what we've seen is an exponential trajectory of the amount of data that is considered by predictive models. Even more important than that is an exponential set of the use cases of models. Today, we see every time a predictive model is used, it’s used in a million and one ways, whereas in 2012 it might have been used in 50, 20, or 100 sessions about each voter contract.

Gardner: It’s a fascinating use case to see how analytics and data can be brought to bear on the democratic process and to help you get messages out, probably in a way that's better received by the voter or the prospective voter, like in a retail or commercial environment. You don’t want to hear things that aren’t relevant to you, and when people do make an effort to provide you with information that's useful or that helps you make a decision, you benefit and you respect and even admire and enjoy it.

Dyskant: What I really want is for the voter experience to be as transparent and easy as possible, that campaigns reach out to me around the same time that I'm seeking information about who I'm going to vote for in November. I know who I'm voting for in 2016, but in some local actions, I may not have made that decision yet. So, I want a steady stream of information to be reaching voters, as they're in those key decision points, with messaging that really is relevant to their lives.
I want a steady stream of information to be reaching voters, as they're in those key decision points, with messaging that really is relevant to their lives.

I also want to listen to what voters tell me. If a voter has a conversation with a volunteer at the door, that should inform future communications. If somebody has told me that they're definitely voting for the candidate, then the next conversation should be different from someone who says, "I work in energy. I really want to know more about the Secretary’s energy policies."

Gardner: Just as if a salesperson is engaging with process, they use customer relationship management (CRM), and that data is captured, analyzed, and shared. That becomes a much better process for both the buyer and the seller. It's the same thing in a campaign, right? The better information you have, the more likely you're going to be able to serve that user, that voter.

Dyskant: There definitely are parallels to marketing, and that’s how we at BlueLabs decided to found the company and work across industries. We work with Fortune 100 retail organizations that are interested in how, once someone buys one item, we can bring them back into the store to buy the follow-on item or maybe to buy the follow-on item through that same store’s online portal. How it is that we can provide relevant messaging as users engage in complex processes online? All those things are driven from our lessons in politics.

Politics is fundamentally different from retail, though. It's a civic decision, rather than an individual-level decision. I always want to be mindful that I have a duty to voters to provide extremely relevant information to them, so that they can be engaged in the civic decision that they need to make.

Gardner: Suffice it to say that good quality comparison shopping is still good quality comparison decision-making.

Dyskant: Yes, I would agree with you.

Relevant and speedy

Gardner: Now that we've established how really relevant, important, and powerful this type of analysis can be in the context of the 2016 campaign, I'd like to learn more about how you go about getting that analysis and making it relevant and speedy across large variety of data sets and content sets. But first, let’s hear more about BlueLabs. Tell me about your company, how it started, why you started it, maybe a bit about yourself as well.

Dyskant: Of the four of us who started BlueLabs, some of us met in the 2008 elections and some of us met during the 2010 midterms working at the Democratic National Committee (DNC). Throughout that pre-2012 experience, we had the opportunity as practitioners to try a lot of things, sometimes just once or twice, sometimes things that we operationalized within those cycles.

Jumping forward to 2012 we had the opportunity to scale all that research and development to say that we did this one thing that was a different way of building models, and it worked for in this congressional array. We decided to make this three people’s full-time jobs and scale that up.

Moving past 2012, we got to build potentially one of the fastest-growing startups, one of the most data-driven organizations, and we knew that we built a special team. We wanted to continue working together with ourselves and the folks who we worked with and who made all this possible. We also wanted to apply the same types of techniques to other areas of social impact and other areas of commerce. This individual-level approach to identifying conversations is something that we found unique in the marketplace. We wanted to expand on that.
Join myVertica
To Get the Free
HPE Vertica Community Edition
Increasingly, what we're working on is this segmentation-of-media problem. It's this idea that some people watch only TV, and you can't ignore a TV. It has lots of eyeballs. Some people watch only digital and some people consume a mix of media. How is it that you can build media plans that are aware of people's cross-channel media preferences and reach the right audience with their preferred means of communications?

Gardner: That’s fascinating. You start with the rigors of the demands of a political campaign, but then you can apply in so many ways, answering the types of questions anticipating the type of questions that more verticals, more sectors, and charitable organizations would want to be involved with. That’s very cool.

Let’s go back to the data science. You have this vast pool of data. You have a snappy analytics platform to work with. But, one of the things that I am interested in is how you get more people whether it's in your organization or a campaign, like the Hillary Clinton campaign, or the DNC to then be able to utilize that data to get to these inferences, get to these insights that you want.

What is it that you look for and what is it that you've been able to do in that form of getting more people able to query and utilize the data?

Dyskant: Data science happens when individuals have direct access to ask complex questions of a large, gnarly, but well-integrated data set. If I have 30 terabytes of data across online contacts, off-line contacts, and maybe a sample of clickstream data, and I want to ask things like of all the people who went to my online platform and clicked the password reset because they couldn't remember their password, then never followed up with an e-mail, how many of them showed up at a retail location within the next five days? They tried to engage online, and it didn't work out for them. I want to know whether we're losing them or are they showing up in person.

That type of question maybe would make it into a business-intelligence (BI) report a few months from that, but people who are thinking about what we do every day, would say, "I wonder about this, turn it into a query, and say, "I think I found something." If we give these customers phone calls, maybe we can reset their passwords over the phone and reengage them.

Human intensive

That's just one tiny, micro example, which is why data science is truly a human-intensive exercise. You get 50-100 people working at an enterprise solving problems like that and what you ultimately get is a positive feedback loop of self-correcting systems. Every time there's a problem, somebody is thinking about how that problem is represented in the data. How do I quantify that. If it’s significant enough, then how is it that the organization can improve in this one specific area?

All that can be done with business logic is the interesting piece. You need very granular data that's accessible via query and you need reasonably fast query time, because you can’t ask questions like that when you're going to get coffee every time you run a query.

Layering predictive modeling allows you to understand the opportunity for impact if you fix that problem. That one hypothesis with those users who cannot reset their passwords is that maybe those users aren't that engaged in the first place. You fix their password but it doesn’t move the needle.

The other hypothesis is that it's people who are actively trying to engage with your server and are unsuccessful because of this one very specific barrier. If you have a model of user engagement at an individual level, you can say that these are really high-value users that are having this problem, or maybe they aren’t. So you take data science, align it with really smart individual-level business analysis, and what you get is an organization that continues to improve without having to have at an executive-decision level for each one of those things.

Gardner: So a great deal of inquiry experimentation, iterative improvement, and feedback loops can all come together very powerfully. I'm all for the data scientist full-employment movement, but we need to do more than have people have to go through data scientist to use, access, and develop these feedback insights. What is it about the SQL, natural language, or APIs? What is it that you like to see that allows for more people to be able to directly relate and engage with these powerful data sets?
It's taking that hypothesis that’s driven from personal stories, and being able to, through a relatively simple query, translate that into a database query, and find out if that hypothesis proves true at scale.

Dyskant: One of the things is the product management of data schemas. So whenever we build an analytics database for a large-scale organization I think a lot about an analyst who is 22, knows VLOOKUP, took some statistics classes in college, and has some personal stories about the industry that they're working in. They know, "My grandmother isn't a native English speaker, and this is how she would use this website."

So it's taking that hypothesis that’s driven from personal stories, and being able to, through a relatively simple query, translate that into a database query, and find out if that hypothesis proves true at scale.

Then, potentially take the result of that query, dump them into a statistical-analysis language, or use database analytics to answer that in a more robust way. What that means is that our schemas favor very wide schemas, because I want someone to be able to write a three-line SQL statement, no joins, that enters a business question that I wouldn't have thought to put in a report. So that’s the first line -- is analyst-friendly schemas that are accessed via SQL.

The next line is deep key performance indicators (KPIs). Once we step out of the analytics database, consumers drop into the wider organization that’s consuming data at a different level. I always want reporting to report on opportunity for impact, to report on whether we're reaching our most valuable customers, not how many customers are we reaching.

"Are we reaching our most valuable customers" is much more easily addressable; you just talk to different people. Whereas, when you ask, "Are we reaching enough customers," I don’t know how find out. I can go over to the sales team and yell at them to work harder, but ultimately, I want our reporting to facilitate smarter working, which means incorporating model scores and predictive analytics into our KPIs.

Getting to the core

Gardner: Let’s step back from the edge, where we engage the analysts, to the core, where we need to provide the ability for them to do what they want and which gets them those great results.

It seems to me that when you're dealing in a campaign cycle that is very spiky, you have a short period of time where there's a need for a tremendous amount of data, but that could quickly go down between cycles of an election, or in a retail environment, be very intensive leading up to a holiday season.

Do you therefore take advantage of the cloud models for your analytics that make a fit-for-purpose approach to data and analytics pay as you go? Tell us a little bit about your strategy for the data and the analytics engine.

Dyskant: All of our customers have a cyclical nature to them. I think that almost every business is cyclical, just some more than others. Horizontal scaling is incredibly important to us. It would be very difficult for us to do what we do without using a cloud model such as Amazon Web Services (AWS).

Also, one of the things that works well for us with HPE Vertica is the licensing model where we can add additional performance with only the cost of hardware or hardware provision through the cloud. That allows us to scale up our cost areas during the busy season. We'll sometimes even scale them back down during slower periods so that we can have those 150 analysts asking their own questions about the areas of the program that they're responsible for during busy cycles, and then during less busy cycles, scale down the footprint of the operation.
I do everything I can to avoid aggregation. I want my analysts to be looking at the data at the interaction-by-interaction level.

Gardner: Is there anything else about the HPE Vertica OnDemand platform that benefits your particular need for analysis? I'm thinking about the scale and the rows. You must have so many variables when it comes to a retail situation, a commercial situation, where you're trying to really understand that consumer?

Dyskant: I do everything I can to avoid aggregation. I want my analysts to be looking at the data at the interaction-by-interaction level. If it’s a website, I want them to be looking at clickstream data. If it's a retail organization, I want them to be looking at point-of-sale data. In order to do that, we build data sets that are very frequently in the billions of rows. They're also very frequently incredibly wide, because we don't just want to know every transaction with this dollar amount. We want to know things like what the variables were, and where that store was located.

Getting back to the idea that we want our queries to be dead-simple, that means that we very frequently append additional columns on to our transaction tables. We’re okay that the table is big, because in a columnar model, we can pick out just the columns that we want for that particular query.
Join myVertica
To Get the Free
HPE Vertica Community Edition
Then, moving into some of the in-database machine-learning algorithms allows us to perform more higher-order computation within the database and have less data shipping.

Gardner: We're almost out of time, but I wanted to do some predictive analysis ourselves. Thinking about the next election cycle, midterms, only two years away, what might change between now and then? We hear so much about machine learning, bots, and advanced algorithms. How do you predict, Erek, the way that big data will come to bear on the next election cycle?

Behavioral targeting

Dyskant: I think that a big piece of the next election will be around moving even more away from demographic targeting, toward even more behavioral targeting. How is it that we reach every voter based on what they're telling us about them and what matters to them, how that matters to them? That will increasingly drive our models.

To do that involves probably another 10X scale in the data, because that type of data is generally at the clickstream level, generally at the interaction-by-interaction level, incorporating things like Twitter feeds, which adds an additional level of complexity and laying in computational necessity to the data.

Gardner: It almost sounds like you're shooting for sentiment analysis on an issue-by-issue basis, a very complex undertaking, but it could be very powerful.

Dyskant: I think that it's heading in that direction, yes.

Listen to the podcast. Find it on iTunes. Get the mobile app. Read a full transcript or download a copy. Sponsor: Hewlett Packard Enterprise.

You may also be interested in: