Eric Kavanagh: OK ladies and gentlemen. It is four o’clock Eastern Time, once again, on a Wednesday, it’s time for Hot Technologies. Yes indeed, my name is Eric Kavanagh. I will be your host for today’s web seminar featuring two of our favorite people in the business: Kim Brushaber of IDERA and Mark Madsen of Third Nature. “Using Process Models to Achieve Business Goals.” We’re going to talk about optimizing the business and how you can really use some of these technologies to first understand what’s happening and then remodel what you’re doing and avoid things like redundancies, avoid things like conflicts, maybe in your supply chain or your business processes, wherever they may be, that’s what we’re going to talk about today. So first, we’re going to hear from Kim Brushaber and then we’re going to hear from Mark Madsen. Then we’ll have some nice back-and-forth and feel free to send your questions. Don’t be shy. Send questions by the Q&A component of your webcast console or by the chat window.

With that I’m going to push the first slide here for Kim and I’ll hand it off. Kim, take it away.

Kim Brushaber: Hi there. So I’m going to start out by talking about how you can use some of your business processes to achieve your goals. I thought I advanced the slide – there we go, it may have just been a little slow. So for a business to be successful, it’s got to focus on how the company makes money, keeping customers and keeping the market happy, keeping costs as low as possible and then delivering quality products and making sure that that information that you gather is reliable. Which we’ve used our buzzwords here: revenue growth, customer satisfaction, efficient operations, product and data quality. And some of the key challenges for a business that we’re going to discuss today include silos within your organization; what’s good about them, what’s bad about them because not all silos are bad. How do you keep redundancies out of your process? How you can reduce and eliminate the gaps in your communication and how you can reduce the inefficiencies in your operations.

So, the first kind of silos is the department silos. And silo mentalities are created when departments don’t want to share information with other departments within the company. And while this can be good in the case of sensitive information that few people should know about – so sensitive merger information or acquisition information or maybe information not ready for the sales team to be able to do something with it – in those cases silos can be really good. But it can also be bad because the flow of information is hindered between the groups in the organization and it can cause a lot of the issues that we’re going to discuss here in a moment. You can also have silos that are divided by business objectives and technology objectives. So the business side of the house spends a lot of time looking at ROIs and KPIs and things that are really focused on the business, where on the technology, they really want to look at how am I going to make my products work or how am I going to bring my services to market? And so because there’s very different goals between the two different groups, you can have a natural silo that gets created between the two of them. And then a lot of times silos can be divided by jargon. So the words that you use in your everyday language can be really confusing to one group or another, and here I just put a bunch of fun little buzzwords that are relevant to either one side or the other side of the wall. And of course this doesn’t even begin to cover the spectrum, but a lot of times, those words can cause a silo to be created and cause two different groups of people to be divided because information gets lost in translation. So there are good silos for your business and I’m going to cover a few of the values that silos can bring to an organization.

So they can provide a structure that allows the employees to do their work without fear or distraction. So if you have your people that are in your silo that you need to be talking to and addressing on a daily basis, that can allow you to be able to get your job done more efficiently and more effectively without a lot of interruption. It also facilitates expertise in specific areas of the business. So if you are focusing in really finely on finance and you’re talking to other people that are in finance and all you’re doing all day is speaking about finance, then that creates a really good silo because that group learns the expertise in that area and they don’t have to be responsible for knowing what’s going on in sales or what’s going on in marketing or what’s going on in operations. It also speeds up communication by allowing people to speak the same language. So going back to that jargon, a lot of times that jargon can be a really good thing because it allows people to be able to communicate more swiftly and more effectively. It also keeps accountability and responsibility within the silo. So you know what you’re accountable for within your group and the tasks that you need to deliver and the person you need to report to and it allows you to have greater accountability and greater responsibility rather than— and certainly the silos have a flip side of that where responsibility can get muggy. But within the silo itself, it can create more accountability and responsibility. And then it also fosters a sense of pride and ownership. So you can feel really good about the job that you accomplished at the end of the day and the tasks that you’re required to deliver and these are all really good things about silos.

But there is a sour side of silos, and silos create inefficiencies, they lower morale, they decrease productivity. And so because this is the more negative side of silos, I’m going to use some business process models to go through a variety of bullet points and explain how you can overcome the sour side of silos using the IDERA Business Architect product to show you some of these examples.

So the first one is it creates inefficiencies and redundant processes. So in this example, I’m showing that the marketing organization may have a set of tasks and the sales organization has a different set of tasks. And in this case, if you map them out, you’ll discover that both of them have a task to qualify the lead. And when you realize that, then you can have a conversation cross-functionally between the two different groups to be able to find out “Is my qualifying a lead the same as your qualifying a lead? Are we taking the same steps and the same behaviors? Or does it mean something different between the two different silos?” And if you are doing the same things, you can start to streamline it and give responsibilities to the different groups independently and business processes can really help you kind of map these things out and identify where you’ve got those kinds of issues.

Also, when you are merging companies or if you are merging groups, the merging process, well, you can go through and you can define your process for the various different behaviors. And in this example, Company A’s got some behavior, Company B has some behavior and the merge process takes the elements of A and B, finds the best practices, and then creates a new process that’s going to work very effectively for both groups. So it helps you to become more efficient, more productive, and identify better practices for your business.

Additionally, another sour side of silos is that there can be gaps in the communication between the departments, which is what we were just talking about, where collaboration is not happening but it should be. And so business processes can help you to identify those kinds of gaps. So in this example, sales has a process, a new product gets released and they go out and they sell it. But finance may have an additional process where they need to go in and update the product prices when the product is released. If sales doesn’t know about that, they could still be out there pitching deals with the old product prices and when it comes to the point where finance is starting to review the deal and approve the deal, then a lot of conflict and a lot of back-paneling has to happen to go back to the customer and readjust it. And if you’ve gone and diagramed your process, you will have known this in advance and can fit it in so that sales knows “I need to wait until I get those product price updates before I start to talk to the new customers about the product.”

In this example, BPMN2 has a conversation diagram that allows you to be able to talk between a variety of different departments and identify the handoff points between them. And this is very helpful for reducing redundancies and also allowing for more accountability and responsibility between the departments. So you can say, “OK, so sales management and sales have to work together to approve the deal.” And they both can work off their handoff pieces and what that relies. But the finance department may not necessarily have to be involved in that approval and they know that based on this diagram that’s laid out that says here’s who’s responsible in the different departments that need to work together to accomplish that.

Additionally, rogue processes can come in that don’t benefit the company. So when you’re going through your business processes, you may identify that somebody is doing something that you’re like, “I don’t really understand how that’s effective or how that’s meeting the goal.” So I’ll give you some examples of that. So in this case, product may be going through and they’re doing a new release. They go, they deliver the requirements, development team starts working on those requirements but after the product team starts talking to customers then we come back and decide to revise them. And this would be very, very disruptive for the development team to have to go back and revise the requirements after they’re already in progress building those items. For product, they may just not even think about anything like that. They’re just like, “Oh, I got some new inputs and now I need these things.” And if they don’t talk to the development team, they won’t really understand how much of an impact that might give to the later scope or the delivery of the product. So diagraming out these kinds of pieces can help that silo break down and allow you to be able to understand what elements are helpful to your process and what are detrimental processes.

There also can be a duplication of assets and resources, and this is a big thing when companies are trying to streamline down. So in this case I’ve done kind of a grouping diagram where I’ve identified a variety of different applications and reports that have to be produced and the different players that are associated. And when you start to lay all of these things out, in this example I’ve given a duplication of the editing tools and the call tracking tools and who’s using them. And so you can start to figure out, because a lot of times the independent silos will make these decisions for their team and they don’t necessarily think about the fact that the broader team as a whole could also use that licensing agreement and make it cheaper and more cost effective for all the tools that are being used in the organization. Additionally, business process diagrams can be very helpful for identifying who’s responsible for what information and when. And so in this case, I have data stewards who have said, “OK, these are the people who are responsible for all these data and here’s the tables that they are responsible for dealing with.” And don’t give this information to other people, this is really important in area where there’s sensitive information like medical records or financial data or elements like that that need to be secluded to just a couple of people. So you can help to identify this, which then allows people from other organizations to not have access to that information and secure it down and know where your information is going.

Additionally, since we’re talking a little bit about data, silos can also create poor data quality and data inconsistency. So in this case, I’ve used a business process to help the data team to understand when is a customer a new customer, or when are you updating the customer. So you can go through and diagram out these decision points and the business side who understands the business rules can easily talk to the technical side who has to implement these rules and knows when certain behaviors have to take place. In this example, it’s talking about determining data duplications. So if you have a retail customer and you have a web customer and you’re selling products, you might have completely different systems that are trying to gather the same information. And if you’re trying to deduplicate your information and identify who your customers really are, business process diagrams can really help you nail that down and say, “Oh, well in this case we’re both dealing with an order and in this case we’re both dealing with financials,” and be able to map that information out so that it’s much clearer so that you don’t have those kinds of duplications in your data and you can reduce the redundancies and reduce the deficiencies and bring up the quality of your data.

So additional benefits of having good business processes are that employees can identify issues at the beginning when it’s easier to implement the changes. This is especially true for complex data processes, if you can do the analysis on the design upfront and get all of the teams involved in the conversation, then the processes will flow out a lot smoother and people will be able to react better at the beginning versus if you’re already in the process. New employees are on-boarded more rapidly because they can go and they can review these business processes and understand the tasks that they need to accomplish and where the handoff points are and who they need to talk to for various different things. And decisions can be made in real-time across cross-functional teams. If you’re both drawing out these business process diagrams together, you can find these points where there’s a snag in the process and be able to discuss it and figure out what is the best process for the two of you and where are the best handoff points and who are the best people to be doing each of the different tasks that need to be accomplished.

So some tips for breaking down the silos for business success and being able to achieve your goals: The first one is to focus your business processes on your customer, your products or your services – not the individual departments. So a lot of times people will want to, within their departments, come up with their individual checklist. But if you instead look at the business as a whole and the goals that the business is trying to achieve, you can start to see where things fall out and say, “Do these processes help me get to my goal? Or are they extra processes or are they hindrances in the process and achieving the goal?” You should spend more time discussing the places where the processes connect. So like in that conversation diagram where you’ve got a lot of handoff points, you need to spend a lot more time talking about that and making sure that information is flowing correctly across the different silos.

You can unify your employees by showing in the process, the things that they’re responsible for and how it interacts with the company as a whole. And that gives people a lot more of a sense of purpose towards meeting, towards the goal. You can also collaborate with employees so that they have input on the process that effect their role and job because if the decisions are all made at the top when designing the process, the individuals who are doing the work are going to see steps that are missed and pieces that are missing and be able to discuss those out. And if you’re collaborating with all of your employees when you’re drawing out these processes, you start to figure out those outliers and whether or not those are actual things that should be in the process or not. And then another tip for breaking down the silos is to update your processes regularly to reflect the changing needs and the goals of the organization because the goals and the processes are very fluid and you may find better best practices. You may find new ways that you want to do things and so being able to update that information regularly can really help the organization. And going back to the drawing board with those cross-functional teams can really help to break down the silos and open that communication amongst your team. So that’s the slides that I had prepared.

Eric Kavanagh: Alright. Let me hand it off to the indomitable Mark Madsen. You now have the floor, take it away. And folks, don’t be shy, ask your questions. We’ve got experts on the line here. Mark, it’s all you.

Mark Madsen: OK, thanks Eric. So what you heard just now was about process and process modeling and how it applies. And then from my perspective, coming from the analytics side of the house, I have used business process a lot as ways of explaining and understanding. Now, when you think about analytics, and especially now as we talk about machine learning and other things in addition to BI, it’s still viewed by a broad swath of the market, kind of, I consider, incorrectly. Which is, you send out analysts like gold miners and they rush out into the data and they poke around and they find some nuggets of gold and bring these valuable things back to the organization and then everybody lives happily ever after. Or at least the analyst does because they have a six-figure salaries because that’s what data scientists are all making, in theory.

But the reality is a lot different. The reality is that it takes infrastructure and it takes work and it takes goals and a direction and understanding of business. And those things, they’re required to really understand how to approach problems, how to model for problems and how to solve those problems. And so this idea that you can throw some data and some technology and some smart people at a problem without understanding the context, in particular the process context within which we’re going to apply it, is largely a myth in the same way that most of the Gold Rush was a myth and in fact most of those people went home bankrupt.

There’s also another aspect of this application of analytics to the business, is this idea that it’s all data under glass, right? That somehow analysts or algorithms will surface data and will throw that on a screen in front of somebody. But the problem is we’ve got so much data and you can do so many different things with analytics that it’s easy to overwhelm people. And then you have a secondary problem now which is “I have so much data and I have so many things, which ones do I pay attention to? And how and why do I pay attention to those things?” And that is really the crux of a lot of problems in environments to the point that we’re falling back on requiring experts to curate what information gets displayed to whom and so, far from having self-service data access and self-service dashboards, you end up relying on different experts to help you figure out what goes in the damn things.

And if we talk about where the future is going with, in particular, a lot of the more advanced analytics but the machine learning approaches, AI in business, all this stuff, well there’s a lot of hype around it. There’s a lot of reality to it and a big part of that is embedded. In fact, the modern renaissance in this came through embedding it into process. So taking processes that were automated or automatable, for example the basic idea of recommendation engines in retail on e-commerce sites or on news sites or on music sites is a simple application or algorithm for a task that used to be a human-oriented task. What do you think people are going to like with the question and the merchandise planner or the person who is figuring out what a cross-sale should be or an up-sale should be based on prior data, they would surface that and then punch that into a system and then either marketing or merchandising or some online application would deal with it. And then it got embedded. As you do things, the machine is watching what you’re doing and refining and constantly presenting new, and that’s an embedded analytic. It sits there inside of a process. And if you really want to know where a lot of the future of this work goes, it’s there. It’s not as much helping people by doing more sophisticated analysis. It is by gaining efficiencies across a much broader swath of the business.

And so when you look at things like business intelligence, which is where a lot of the data and analysis market came from, there were statisticians before BI really enabled a lot of people to do a lot of things without statistics, without anything else, by focusing purely on data. The problem was that by focusing purely on the data, it left out a lot of the context. And so what you end up missing is how all of that data, how all these metrics relate. If you think about what goes on a dashboard, you’ll have some bar charts, maybe a graph, a table of numbers. You’ll see a bunch of metrics either individually or collectively and you don’t really see how they relate. So imagine you’re somebody new to something and you go in, you can look at a dashboard and you won’t make heads or tails out of any of the numbers because the numbers themselves don’t tell you anything because they don’t have context. So it might show a number in red but just changing this other number by pulling some other lever might make this better or worse. How do these things relate? That’s the context gets lost in business intelligence and data warehousing and dashboard design because you model data, not process. And that’s the fundamental aspect is that you build repeatability around data and you do that by squeezing most of the process out, focusing on the metrics that are generated off the raw data.

So this screen shows us what is, essentially, a dashboard about lab testing process. There’s an application called Altosoft that does BI in this way. And so what you look at is you see the process and the data not separated, but put back together again. Like that separation was artificial and it was done because we abstracted data, shoved it into databases and built interfaces on top of it. So you usually have two metrics; you have things like the number of tests ordered, which is the first box on this flow, and the last box would be the number of tests completed and filed. And so you’d have these two metrics; you’d put them on a dashboard and you might notice that one is lagging the other significantly. Or maybe you have a third metric which is reprocessed.

So if you’re doing lab tests in a hospital, there are a lot of tests. Many of them are urgent because they’re coming ahead of surgeries or they’re coming out of critical care units or some other thing. So you have processes in place where doctors order them, they go into a laboratory, the laboratory has a process for marking that they’re received, they’re scheduled, they’re going to get done, they’re going to run through the equipment. Sometimes if they sit around too long, because the laboratory is backed up, all the equipment is occupied, they have to be reprocessed. Sometimes the results aren’t valid. Sometimes things like blood samples, they can’t sit for more than 30 minutes or there are breakdowns in the samples and then you have to go and draw blood a second time, which is something you really don’t want to be doing to people. So that means that there’s actually priorities on some of the lab tests over others based on their perishability. So you have other things going on inside the laboratory and you want to avoid those reprocessing problems if at all possible. But you can’t really see the flow of tests through different things because BI itself is typically only about flow in the aggregate metric sense. And so this interface is showing you the data attached to the process so you can see how many come in, how many were received, how many are going on at any one time. I guess it’s not a live demo so you can’t see drill down into the details of the process and the metrics going on inside, what’s happening with the batching or the reprocessing. But this is what gives you a much better view and so a person who at least understands a laboratory can look at this and see what’s going on, as opposed to a bunch of graphs and metrics on a single screen. And so process helps a lot on the interface design side, it doesn’t hide the context.

Process also comes in in other areas. Really, when you talk about BI and data warehousing, before we get into the more advanced analytics, you’re talking about doing one of two things: You’re either talking about analyzing what’s happening within a process and then acting on that, or you’re analyzing the process and then changing it. So the standard sort of organizational use of information is to monitor situations – that’s what your dashboards do and your top 10 and bottom 20 reports. They’re all simple monitoring tools to allow people to see what they need to see and look for deviations. There might be traffic lighting on the dashboard, there might be the bottom 20 report which is essentially a deviation report that’s showing the worst performing something. And then you analyze those things so you look at other data, you look at other things. Maybe you go into a lot more detail around the analysis and then you look at the causes. You might already have a gut feel for this and skip right into action. Oftentimes with a simpler and more well-understood processes that’s exactly what happens. You see a problem, you know what’s going on, you make a decision and you take an action. Usually that is within that process loop on the bottom, you have SAP, it has these things, you see it out of stock in the store so you increase the purchase order for the next round of replenishment and you’re done.

There’s nothing special that happened, but other times, you haven’t seen a problem before so you have to analyze the causes so you really have to dig in to what’s going on. Usually at that point where you start having to analyze cause, you need to understand the process because this is a problem you’ve not seen before, so it’s out of the bounds of the normal process, the day-to-day that’s embedded in our OLTP systems and now you have something that requires some critical thinking. It requires more context because you have a set of problems and a set of possible causes that you have to weed out. You have to reason about this, analyze and gather new information and then change the process. This is happening because we did something. Maybe we didn’t synchronize our marketing campaigns with our replenishment processes so we’re running out of stock. Hopefully that’s not happening in retail, but a lot of retailers used to have these problems when we first instituted BI and data warehousing.

Now, often the causal analysis involves statistics and other more difficult analysis than eyeballing a few numbers, but then you come into the second part, which is you are changing a process. Are you making changes in the right place? Do you understand where to make those process changes? Does the data bear out your intuition or your analysis about what’s going to happen after that change? What other processes are affected? What other numbers in your dashboards that you’re paying attention to will be affected by this? And you’re probably going to be collecting new data which you’re going to feed into the monitoring cycle. So process is actually inherent in understanding at a larger level as you take actions and do things. And the BI world often assumes linear causality. In fact, most management schools are really bad at teaching people how to build performance management and performance metrics around business because they assume straight-line views. And straight-line views are in turn reinforced by simple BI reporting and single metric kind of reporting that you pull up because it doesn’t understand the process of how things influence other things.

So you can use process models not purely as business process models, but you can also apply systems dynamics. You can apply process models and use them in the same way to understand how metrics relate to one another. So in a straight-line view like this diagram – I apologize, I forgot to put the reference to the paper this was from, it’s an old one from the ‘80s, it’s just about systems dynamics and how things are assumed to be and how they really are. So profitability always assumes that if we make quality better than profitability, we’ll get better somehow. Or maybe it will get worse because to make quality better you have to spend more money and that reduces profitability. So there might be a negative on that arrow. Or how leadership or how the alignment of different silos in the organization or process leads to better profitability or lowering costs. There’s always factors and the idea is that any one of those metrics on the left will influence that metric on the right, and it’s all linear.

The diagram on the right side shows a much better example. It shows what’s really happening here, and what’s really happening is that you might change product quality, but there’s a feedback loop between, say, product quality and cost structure which raises the cost structure which lowers the profitability, even at the same time that it also lowers the costs of warranty repairs. And so the math behind this gets a little fuzzy because you can fix something by lowering costs, but you decrease product quality which decreases satisfaction which decreases sales and it increases warranty costs.

Or you could do the inverse. And so you have to more carefully model what’s going to happen as you change any one of these things. And so your metrics about things on the left are in themselves going to be influencing each other and how you change those things, the levers that you pull in the business or your adjustments to business process or practice, are going to influence these. And so process assumes a central role where for a very long time we built very simple things.

And so the next thing is to look at how processes themselves interact. If you take that earlier diagram I had and you, say, change something, you really need to look at how processes interact because a change over here leads to something over there and so this diagram from the earlier presentation about how changes in marketing and changes to data in marketing that lag, what’s going on in sales are actions that lag, meaning your action may come too early or too late to do any good and so it pays to understand how the impacts in one process manifest in another process because everything is always immediated through process.

And so what you have then is just a lot of complexity in business and very often we didn’t capture that. We didn’t capture that when we were working on statistics projects, on machine learning projects, on BI projects and so now you talk about injecting, say, machine learning into a lead scoring process for marketing and sales where it helps you to qualify leads, which affects these two yellow boxes here. Well that lead scoring process that happens somewhere is going to affect both of these. And so it’s going to cause a recalibration or a change in these two processes. If you went into this with the idea that this lead scoring thing is a marketing problem and we’re going to hire a data scientist and they were going to build this lead scoring algorithm for us, it's going to do these things, it’s going to better qualify our leads and prioritize things. How does that affect sales? Is it applied in the right place? Maybe you need to see what's going on across those processes because they both have to change. It's not purely a marketing project. And that's the point of a lot of analytics is that in fact the context and the impacts are a lot brighter and the scope increases, it gets bigger and much hairier.

And you can look at problems at many different levels. So at first you look at it in the context of a marketing problem and then you say, “Oh, well this actually affects marketing and sales. But this project itself has IT impacts, so there’s an IT angle to this which implies that we've got to do other things and by the way this is going to modify SAP which means we got this other process impact.” And so the bounds of complexity will vary and also the level of analysis because process isn't purely just, “Look at this process” or “Look at how those two processes interact.” If you're an executive and you're making much higher order tactical or strategic decisions, you need to see even bigger pictures. So this is a value chain diagram, it's one of my favorites, but it's for the farm-to-retail cheese-making process. So you know in the very far left-hand side you see farms and on the very right-hand side you see retailers and in between you have the transportation that moves physical goods, basically milk and butter, moves dairy products around to various factories which moves to processing plants which moves to distributors and post-processing and packaging plants and all these different things. And that's essentially a supply chain that goes from the production to the consumption.

And what you're seeing in red and green up above is actually the data side of the process interactions between companies, because this is a value chain not for one company but for an industry, although this was actually for a company. You would put yourself into something like this and map this out and there's a lot of different value chain and value system, value mapping things that goes back to Porter in, I think, the late seventies/early eighties. But the idea is that there is process here and those red things are all the information flows from one company or one set of operations in the supply chain to another. And that implies that one process in one organization is interacting with another process in another organization. And so process flow and the data flow, both are important and both should be visible in terms of documenting what's happening and understanding what's happening and reasoning about it, because then you can come along and say, “Well, what if I applied AI to my process over here and I changed how I did this perishable management to reduce the fact that in transit or in waiting areas and distribution facilities, I have products go bad.” And so I do logistics and supply chains adjustments but it affects not just me, but upstream and downstream suppliers. It affects my processes and it has information flows that are going to be affected and so the process helps you think about how that’s going to work and who you’re going to impact and who you need to deal with. And so it applies not truly for an analyst or for a BI person or a data scientist, but it applies also to the managers who have to use this stuff.

As a more concrete example I’m just going to throw in a really straightforward thing here on marketing because I think a lot of people have a fairly intuitive grasp of the basics of online marketing. I think everybody at some time or another has probably seen the obligatory funnel diagram where there’s an audience of people out there. Marketing is not purely about advertising. It’s about a lot of things, but at the very beginning of it, it’s get the word out. Make people aware of your product or services. Advertise to that audience to generate prospects and so the audience kind of narrows down the prospects, people who may be interested in you product. And when product specs are qualified enough, they become opportunities. They become sales opportunities. So every single one of you on this webcast is a potential marketing opportunity for the people who are paying for this webcast because in fact they are trying to find people who are qualified leads. So they’re hoping that these sales opportunities turn into leads – actual people who are interested in the product or the service who want this thing, who want to have it, and of course if you buy something or donate or do whatever it is that you’re doing – this applies equally to not-for-profits raising funds. I can become a customer, a donor. And then, you know, hopefully, the hope of hopes for marketing is that you become proponents, right? So there’s always things like promoter score metrics that you can build about word-of-mouth marketing and how happy customers leave word-of-mouth to tell other people about it, which reaches out to the audience not through formal marketing channels and creates more prospects, opportunities, lead customers and so the cycle goes.

So that's a basic funnel, everybody sees that if you're doing any kind of, you know, web analytics work you see things like conversion charts, right? This is a classic BI thing, you see a conversion rate which is simply transition from one phase to the next here. So the big mass audience whom you don't really know because you just blanket advertised prospects, hopefully people you may know maybe knows something about two opportunities which are identified, prospects people, companies that you know about which then cross another boundary. And so you’ll have different campaigns. Get people to click on the banner ads and get people to attend this webcast. Get people to do something and each one of those have the conversion rate – so the number of people you reach out to and the number of people who actually take the action that you want. So a lot of conversion rates typically online will balance between, say, one and five percent depending on industry and the kind of thing you’re doing. So you’ll have a bunch of metrics.

In this case I'm showing the typical kind of analytics thing, where pages did they visit with or what was the bounce rate. But that's a singular metric and people look at those and measure things off of them, but they're really not that terribly useful. What happens is that one to five percent – and in terms of a lot of online advertising – it’s but about one to two percent if you're lucky. This is the real context, right? It’s everybody else who didn't convert at that point for that thing and that little tiny line at the bottom which gives you a much more realistic picture than this chart does. But, really what I showed you before with that funnel diagram ought to look something like this, right? The balance rate, which would be the people who show up on sale websites or on mobile sites and leave immediately, right? They just weren't really interested. Then there’s people who stuck around for a bit and then there’s people who stuck around for a bit more, maybe clicked, maybe registered, maybe did something. This is actually from retail analysis; I was doing where you have shopping carts’ rates, so the abandoned rate, filled out a form and left, started to donate money and left, started to sign a petition and left, put something in a shopping cart and left. You should really be graphing all of these things but you know what you're seeing here, you're seeing a metric for each of these things. And each of those metrics, if I go back to the funnel, is the transition from one point to another.

These are actually process-aligned metrics. And if you, of course, want to make things a little bit more complicated, you'll find that in fact there are many channels, right? Because marketing is very complex sort of communication channels. There's the old stuff, the radio, the TV, the print, and print is not just magazines and newspapers, it’s circulars you get in your mailbox, it’s those little annoying cards that go into magazines or that they stick into your mail. They’re cards and flyers and stuff that they hand you on the street. And then of course there's mobile channel which is essentially another online channel, but it's subtly different. Games are actually a marketing channel. Movies, media are actually marketing channels. Anytime you see a brand name inside of a movie scene, somebody is paid for that. And then I just broke down the online here, you have your website, email marketing which is still very popular, interactive voice response systems – the annoying touchtone systems when you call customer support and can't get through. Many different social networks.

So each one of these in turn breaks down to many other things like the social stuff. You’ve got Facebook and Twitter and Instagram and Pinterest and 100 other things. And so each one of these has its own marketing process, its own way of choosing how to engage, how to spend, what you're spending, what you're going to do, how you're going to go about that and how you're going to measure. Each one has a process. So Facebook marketing is different from Twitter marketing is different from Instagram marketing is different from Pinterest marketing. Which means that each one of those will have similar – probably alike but slightly different – things and maybe different people dealing with them. So each one has a process. So the amount of processes underneath these metrics is actually very deep and they influence each other. By doing one thing you affect other things and that interaction is very useful and nice to see in process diagrams.

The others of the funnel concept itself is too narrow because it typically chops off at the point when people become customers. Usually that's when marketing says, “Our job ends.” Very few people realize that marketing's true job is to generate customers for sale. And so it should be measured all the way through the end point. And once the customer's acquired, the other part of marketing that people outside of marketing typically don't know about, is that it's not just acquisition, it's the management of a customer life cycle. But that's typically a different silo. As Kim was talking about earlier, we have silos and customer care and warranty support and all of these other things usually run in different departments or different departments within marketing in their own silos. But you need to see across them. You need to see the process that feeds things in, through and out. And the hot topic from – say, well five to 10 years ago but it’s still today – is all about customer 360 and user experience and customer experience management. Well customers experience the organization through many touchpoints from acquisition through support and so you can have great experiences on the marketing side and sale side and have terrible service and never come back. Or you can have a terrible sales experience, don't buy the product but decide that that's the end of it no matter how good the service is. And so it expands the view of process in the context within which you look at metrics.

And so understanding process across the horizontal, across departments, across line of business view is an important thing not purely within there. And one of the challenges, of course, as BI or data warehousing or data science practitioners is that the data is all chopped up because of those silos. The marketing automation systems handle the front end; there's online marketing systems; sales automation systems deal with the middle parts once they've translated into the bowels of SAP or Oracle OLTP systems. Then it’s different things, and of course call center biz is often detached from any of these other pieces and then you need to stitch it all back together, and so process diagrams help you understand how all the systems relate to one another, which also helps you as BI data or housing data science practitioners figure out what data goes where and how and why. So I personally use process diagrams in many different places inside of these analytics projects because they help you to map out and understand data requirements as well as do the job. As we saw earlier, there are places where process models make use of data visible. They make use of sales and marketing data and who owns what data and where that data lands visible and where those overlaps are. They also help you understand because of the location of people and departments in the process diagrams, who's doing what work and therefore who the actual process owner of that data is. So you can see who owns the financial data, who owns the health data, who’s responsible for these things. And sometimes that’s useful when you to see metrics and there's a gap between two processes and there's a data transfer between those two processes and there's a person on each side of that who’s probably responsible for either the upstream or the downstream data and you need to find them. Or you can go to the process maps and see these things.

So process model can help make this visible and so you can leverage these things in your projects. And you know, as we look forward, a lot of what I talked about at the beginning around BI and analytics and even some of the data science, the aspects of things at a superficial level, they’re all about analyzing basic process and metrics. But the other thing that you can do, aside from embedding analytics into processes or analyzing processes and changing them, is building simulations. The old way of building simulators, the way we used to do it a long time ago is you got smart, math-y, people, they built models that would simulate the system, typically by understanding the processes within that system. But there's another way to do that, which is to take some of that understanding and then feed data into it. You built a simulator, it says it works this way, you have all of this data. You should be able to map that data into that simulation and see if your simulation is crap or if it's good. And so you can begin to build simulations of process or interacting process, which is a very hard thing to do.

By analyzing and feeding data into sort of black boxes – there’s black box and white box simulation models you can construct and so you can validate simulations – you can use the data to construct simulations; you can do more interesting things and that's really a big part of where the future is going. That and something which has been around for a good decade or so which is decision automation itself – which is to take the very routine things that people do that are rote, that you just spend time, you know, pressing buttons for – and begin to do decision automation, and some schools call it “complex event processing.” But you know those are another angle of injecting item decision making and analytics into process, which means you need to diagram those processes to see how and where that practice can be applied.

And then finally, we’ve almost never turned process modeling around onto the thing that we do, which is make decisions using information. And that is one of the areas that decision automation and CEP actually do a little bit of. But I've done it a little bit myself in terms of research around decision making and that is, what is the process that a human being goes through to make decisions about a particular thing? So it might be merchandising, it might be marketing, it might be something in logistics, but there's a human being making decisions and if you model the decisions and that they make, you have a better understanding of the data and the metrics that are needed for them. And so you can use that decision process model as an actual mechanism for constructing better dashboards for figuring out what analytic functions can be applied in use to make that or to enable that person to make better decisions. And so it's one of those things that’s still kind of out there to be explored.

And so with that I'm going to terminate here so that we have time for questions.

Eric Kavanagh: Yeah, that was a lot of really, really good stuff and Kim, I have to say, between you and Mark, I think you both just laid out quite an impressive array of situations and scenarios where process modeling will really pay dividends. I guess I'll just throw it out to you, Kim, first. How do you get the business to appreciate this and to realize how much time can be saved, money can be saved, profit can be increased and so forth by really focusing on distilling those processes down to a set of diagrams and then analyzing them?

Kim Brushaber: Yeah, I think that the first thing you have to do is identify a champion in the organization that wants to see their processes mapped out. And once that— and have that be a key stakeholder in the organization. And then identify a small group to start building out the processes and again focusing on what's the business goal and what's the business trying to accomplish, not just what's going on within a department. And take that one goal and map it out within the champ and take the champion and then show the rewards that you get from the process and that will then allow other parts of the organization to go and start to build those processes as well until you can build the entire organization because most people can't just bring a consultancy in that'll just diagram out all of their processes all at once. So they have to do it in bite-sized chunks and picking the most strategic places to look or the places where you expect the most process issues to exist. And start to kind of untangle the Christmas lights and see how that comes together.

Eric Kavanagh: Yeah, that’s actually a great metaphor – untangle the Christmas lights, because underneath it, you are going to find a lot of complexity and a lot of workarounds. Really, I think that's where a lot of problems usually emanate, are either through a merger – as you suggested earlier – or just workarounds that have gotten baked into the process over a period of years that no one ever took the time to untangle, right?

Kim Brushaber: Right, or somebody just started doing something and it was never discussed in the first place.

Eric Kavanagh: Right, that’s interesting. Here is a— and this is a good one. I guess I’ll throw this over to you, Mark, and then Kim, if you want to comment on it. One of the attendees writes, “Given the ever-changing and growing omni-channel environment, how is attribution best managed or allocated?” I think that’s an ongoing question, but Mark, what do you think?

Mark Madsen: Yeah. The whole attribution problem in marketing is huge. If you don't know what attribution is, that's just taking, say, a sale of something – like the online example, if you go to Amazon and you buy a book. Well, how did you get there? Was it search engine optimization that led you to that place by just getting the rankings of that book at that particular spot so he went to that particular place to buy it? Was it an online ad, was it a social media campaign? And you know the problem is that the idea of attribution modeling is that there's this sort of main cause, but there's obviously multiple things. Maybe you saw the book on the book stand and you saw a banner ad for it and then you decided to search for it later because you were looking for something to read and then he went up there.

And then the question is, “How do you apportion the media spend or the value of that sale and customer across various campaigns?” And it's an enormously complex task and you need to do it because obviously you’re trying to budget your most effective campaigns. But also because a lot of times there's a cost like an affiliate fee or something or click-through that gets charged to you for this. And then you have to decide who gets paid. Does Google get paid, do these guys get paid, do those guys get paid? Because typical attribution schemes are “first guy gets paid.”

And so I think the bottom line is that it's an enormously complex problem and it's a multivariate kind of statistical analysis problem that has no clear answers. And that means that, you know, you need to track metrics and see what you can try to tease out and there are things like conjoint analysis and other weird stuff that used to be popular that might become popular again for those types of purposes. But that in turn means that you kind of have to understand the process metrics, at least at the level of “I have five different types of marketing campaigns, I need to know what the inputs to that campaign is, know how much money I'm spending to process metrics, like how many emails or how many ads did I show?” And outcome metrics corresponding to the timing or a link or a tracker on this thing, this transaction occurred. So that you can begin to build that picture – and again that's another good example of where kind of mapping out at least the basic process interactions can help you reason about it. Bottom line, I don't think there's any clear answer to attribution, though.

Eric Kavanagh: Yeah, I think you're exactly right it. And you're never going to know, it seems to me. You can know for the main at least, you can have a good idea where most of the things came from, but to presume that you can know it all or could ever know it all, I think is just a mistake at the outset.

Mark Madsen: I think Heisenberg already wrote about this.

Eric Kavanagh: What’s that?

Mark Madsen: The Heisenberg uncertainty principle rules it.

Eric Kavanagh: That’s nice, that’s a good one. Let me throw this over to you, Kim, because as I'm looking at this and I'm listening to this presentation, what you mapped out with a lot of these different scenarios and then what Mark did as well, you know what pops into my mind is this whole concept of digital transformation that everyone keeps talking about. And to me, that's a great entree for this kind of discussion, because if you look at the new winners in terms of major innovation like Uber, irrespective of their cultural issues, and Airbnb and some of these other companies, what they did was distill key processes down to this level, to the diagrammatic level, and they really focused on building out bulletproof infrastructure to serve these serious services in the marketplace. And they did so at scale, right? Well digital transformation is all about leveraging the new power of cloud computing, of machine learning, of analytics, of whatever the case may be. So to me, anyone who's talking about digital transformation needs to do process modeling. What do you think?

Kim Brushaber: Yeah, and I think that another term that's frequently floating about right now is “process automation,” which you in first need to build out your business processes and understand what they are before you can start to automate them. And then you can put your plans in motion. But absolutely when you're dealing with the digital transformation age, you know you need to be looking at what is the information that I'm gathering and really getting into agreement within your organization on what of that information is important. Because you know, like the slide that Mark shared where you’ve got all the different TV screens with all the different information, we have the capacity to gather so much data now that you really need to define as an organization and get on board with everybody, all of the key stakeholders, and say via business processes, “This is the critical information and these are the critical steps,” and also be able to understand where your pivot points are. So, you know, “This is a process that isn't really working well for us. Let's go in into a fine-tune detail and figure out how we can do that differently,” and talk to the different touch points and see their inputs in the conversation as well.

Eric Kavanagh: Yeah that's a really good point and I thought this slide also did a good job of communicating the importance of dependency. You know, anytime you change one of these components, you change them all, and trying to wrap your head around how that can impact business processes frankly takes some time and effort. But again, it's the kind of thing where if you're talking about engaging in any kind of digital transformation, you need to realize where processes can be collapsed, where they can be eradicated. I think that's usually one of the sort of unsung heroes of successful implementations is when you realize you no longer need X, Y or Z processes if you re-architect the overall plan.

Kim, I guess I throw that back over to you. What do you find is some of the key success factors when this stuff goes very well? What are some of the characteristics of those success stories?

Kim Brushaber: I think, I mean, obviously collaboration is essential and that's why I decided to focus the slide deck that I've got on silos so much, because collaborating between the different organizations and figuring out where those redundancies are, that's a huge way to streamline and make your processes more lean and to have these conversations about, “OK, so this is the way that I'm doing it,” like the one with the merger slide, when you're talking to multiple different departments or you're talking to companies that are coming together and really figuring out the best practices. And designing what are the best steps to take and getting everyone in alignment with those steps definitely makes all of that information go a whole lot smoother.

Eric Kavanagh: Yeah and I'm glad you threw in the word “collaboration” too. Mark, I’ll just throw it over you for comment. Collaboration is such a game-changing component of the new world of business, even with simple stuff like Google Docs, for example. Instead of passing one document through five different people by email, you can have all those five people looking at the document in real time and making adjustments and see what each other comments on. That's a big deal; that’s a major change in process. And that same component can be applied, of course, to business intelligence, to process modeling, really any of these disciplines that we use to optimize the business. Collaboration should be first and foremost anytime it makes sense, right?

Mark Madsen: Yeah, I think so. I mean, this idea of the lone decision-maker is sort of like, you know, that the lone analyst who's magically going out there to do their analysis and turn up that negative gold. And the lone decision-maker sitting at their desk is kind of an old-school, 1990s-era view of how people and organizations make decisions, you know? You sit behind a desk and you look at this thing and then you make a decision, but that's all captured in process and applications now. The real decisions are typically made across departments or with other people, and that requires broader understanding and communication of what's going on. Otherwise, you just dig in your heels and everybody fights and nobody wants to own anything, which is why I don't work at a number of companies anymore.

Eric Kavanagh: Well, you know, that’s a very good point and Kim, I'm really glad that you brought up this concept of things being lost in translation. I often think people do not appreciate nearly enough the importance of context in any discussion, anywhere. Context is so important in terms of helping people understand that the range of issues being discussed and whatever the decision points are. And if you can use process modeling as a mechanism, again to distill what can be fairly hairy complex organisms down to relatively simple – and if not downright elegant – diagrams, to me that's very useful for: A) communicating what is essential, but B) not overlooking things that are critical but could get lost in conversation, and C) finally crystallizing something visually that, frankly, words in the dialogue would have difficulty nailing down. What do you think?

Kim Brushaber: Well it's really interesting that you keep bringing up this term “conversation.” And I included the slide that was on the conversation diagram where there were the multiple different pools that were talking to each other and interacting with each other. That's why the BPMN organization decided to create that diagram, because they understood that the conversations that take place between different departments is complex and there needed to be a way to be able to showcase all of the pieces that were involved in a process and all the different players and all the different aspects so that no balls got dropped and everybody knew where responsibilities were outlined. So in business process when you were talking about, you know, having the right sense of context, business process diagrams are really great because they are visual and pictures are worth 1,000 words, and when you can see these things in a very visual context, it allows people to be able to understand a lot better than, say, if you wrote your process was out in a paragraph format and you wrote them, you know, physically or even if you numbered them with bullets. The pictorial representation allows you to be able to gather that context and that understanding much faster than if you were you know trying to read it or understand it.

Eric Kavanagh: Well, you could also depersonalize things to a point, right? Where people won't take things so personally and you’d have a much more objective view then of what the business is actually doing and certainly for the more complex processes, I think that would help both business and IT audiences better understand what the big picture is, because at the end of the day the big picture is the business and wanting the business to succeed in, let’s face it, it’s fairly tumultuous times. That's why I think the time is right, and it always has been, but it seems even more so these days as we do see certain processes optimized or even eradicated. For example, going to the cloud, just offloading a whole component of your service offering to the cloud or to some partner or whatever the case may be. But having that depersonalized, clear diagrammatic model of the business is a very useful thing for redesigning and for staying on top of things, right?

Kim Brushaber: Yeah and the ER Studio products, we have a lot of searching and filtering capabilities as well. So if you wanted to go and designate that something was cloud behavior, you could go and fine-tune it and do a search to see what are the pieces that are interacting in the cloud once you’ve diagrammed out all of your processes. Or, for example, let's say you're looking at marketing and you just want to fine-tune into marketing – and I certainly don't mean to pick on marketing – it's just the first one that came to mind that most organizations that have. But, you know, to go and be able to say, “OK, so I'm thinking of changing my marketing department. These are all the behaviors,” and so you can look at all the processes and say, “OK, I’m going to put these tactics that we use to do this way up in the cloud and do this and that's going to affect these pieces and that’s going to affect these people.” And if you have that process diagrammed out, then you can see very visually – it’s like looking at a giant puzzle, right? You've got all of these different puzzle pieces that all play together and you can figure out, “OK, do I need to rearrange these puzzle pieces in order for everything to fit in one piece?”

Eric Kavanagh: Yeah and you know I’ll shove one last question to you. And folks, I’m about to post a link to the slides from today’s presentation; look at your chat window to see that. But, of course, process modeling and data modeling terms for data information that's going through systems is critically important because systems either work or they don't, where the business it can be a bit more loosey-goosey. You can have workarounds – let's say in the old days at the end of the process or the beginning of the process or anywhere in between – you can have the workaround that someone just figured out one day when something broke that nobody knew about. Well with data, you'll know for sure because the data doesn't show up in the field where it's required and the transaction doesn't get done. But do you see now that A) we're going towards a more digital economy, but B) we have all these different mergers and things happening. Do you see that companies are starting to greater appreciate the value of business process modeling as well as data modeling? Is that kind of carried over? Because I know certainly for data modeling, data modelers have been very passionate about that for years and years. Does the business get it these days? Are we getting closer to where there is the necessary appreciation of what the stuff does?

Kim Brushaber: Well, I mean, that's exactly what we are trying to accomplish at IDERA. We have the ER Studio Suite involves both the data modeling suite and the business architect suite, so thank you for queuing me up so nicely.

Eric Kavanagh: There you go.

Kim Brushaber: But we do— obviously the data modeling piece is absolutely essential for anyone in information architecture, solutions architecture, anyone who's responsible for the data within the organization. And the way that we have built our product allows the business and the data to kind of work kind of hand-in-hand using our enterprise team additions suite so that you can push all of the objects that are available to the business process and the data process together and be able to bring those two worlds together. And certainly I don't have enough time to go into the details on that, but anyone is welcome to go and look at IDERA and see how we do that.

But the question being, the world of data is going to continue to become complex. Storage has become cheaper and cheaper and cheaper and so it means that we are going to acquire more and more and more data and that's where the items like Mark was discussing on, “OK, so now that I have the data, how do I analyze it? How do I understand it? How do I extrapolate it and how do I use it for my business?” And so being able to overlay that information onto the business process and say, you know, “I need to make a decision on a manufacturing decision and I need to know how many times are my trucks getting delayed due to snow in the wintertime? Do I need to open up a business in Costa Rica to be able to ship things from there instead of shipping them from the north?” And being able to look at all those aspects, but you don't even know that you need to look at those aspects until you can start some to map out that process, and in this case it's a transportation process, but every business has complexities in their process that they can throw down in a business process model and start to understand where those pieces can move.

Eric Kavanagh: I love it. I especially like the part about opening a business in Costa Rica.

Kim Brushaber: Why not?

Eric Kavanagh: If you need a PR guy or a moderator down there, let me know. I did post the link of the slides in the chat window, so check that chat window. Of course, if you did not see that or you want to share this with your colleagues, we do archive all these webcasts for later viewing. And you can email Kim right there, she’s got her address on the screen. Feel free to send her an email directly.

And with that we're going to bid you farewell. Thank you for a fantastic presentation; this has been great. We’ll catch up to you next time, folks. Take care. Bye-bye.