In this episode we talk with Eric Lundberg, Senior Product Manager for Camunda Optimize.
Optimize is a "Business Intelligence tool for BPMN", providing you with actionable insights into your processes running on Camunda Platform or Camunda Cloud.
Eric gives us the low-down on the value that Camunda Optimize provides to business and technical users, how it can be leveraged across the whole of business - including fully manual processes and parts of the business that are digitised but not in Camunda - and what to look forward to in Optimize's development.
Josh Wulf (00:10):
Welcome to this episode of the Camunda Nation Podcast. I'm your host Josh Wulf. I'm a developer advocate at Camunda, and I apologize for my croaky voice - I'm just coming out of an intense flu that I've had for the last week or so. Now, in this episode, which I recorded before I got a croaky voice, I'm joined by Eric Lundberg, who is Camunda's product manager for Optimize. Optimize is a business intelligence tool for BPMN. It's available for Camunda platform, and now also for Camunda Cloud. Eric explains about the motivation for Optimize and the opportunities that it unlocks as part of the Camunda tech stack, from both the business and technical perspectives. So, without further ado, here is the conversation with Eric, Camunda product manager for Optimize.
Eric Lundberg (01:03):
Hey everybody, I'm Eric Lundberg, I'm the senior product manager for Camunda Optimize, and I've been building things to automate and basically make processes better since grade school. So, I originally started in the military where I was driven crazy by repetitive bureaucratic processes that do not scale, and so I decided to get into a place that does scale. So, I decided to help solve problems using software. I spent about three years at Citrine Informatics, which is a Silicon Valley AI startup that optimizes manufacturing processes by providing new recipes to make stronger, better, more sustainable materials. And that all brings me to Camunda. So, now I can bring all of my backgrounds together for analysis, building things and process improvement into one.
Josh Wulf (01:52):
Awesome. Thanks for coming onto the podcast, Eric. Really appreciate it. And I did not know that you had a military background. That's a big area of things that need to be, well, not automated, but they need to be done systematically. So, I can see how that would expose you to the need to do things systematically, processes and automation. And you are the senior product manager for Camunda Optimize. So, for our listeners who aren't familiar with it, what is Camunda Optimize?
Eric Lundberg (02:25):
So, what I can do is I can walk you through my vision for Optimize, do the whole spiel first, and then I'll kind of break it down to cover the main points. So, for process owners who want to perfect their business processes, Camunda Optimize is a process intelligence tool that leverages process execution data to continuously provide actionable insights. Unlike standard business intelligence tools, Optimize specializes in BPMN-based analysis, and can show you exactly what to change to succeed. So, that's a lot of words. That's a lot of, I would say, business speak because it's a lot of information coming into one. So, I'm going to take a few minutes to unpack what's going on with that. So, the kind of people that benefit from this most are the process owners. These are the people that are responsible for the metrics: the director of customer onboarding….
Eric Lundberg (03:15):
It's an IT manager, it's a director of loan application. Anybody who's responsible for running a business process. And it's not just any business process. These are the business processes that make or break your company. If you are a banking company and you don't send out any loans for a month, how bad is that for your organization? Does it go up to the CEO if something breaks like this? And so these are the areas that you really need to spend the time to make sure that they run very, very smoothly. And you could throw a whole bunch of manpower into it, you could throw a lot of energy or you could figure out what that one point of friction is and remove it, and Optimize helps you find that. And so that's the “about it being a process intelligence tool”. So, it learns about your processes and it tries to give you some actionable insights or shows you the data in a way that you didn't think of before. Instead of “things are taking too long”, you can answer why.
Eric Lundberg (04:15):
And that “why” is really important, because once you can ask “why” long enough, and you can get down to the root cause, there's usually not that much additional work that you need to do in order to fix something. This is borrowed from the “Five Whys” lean manufacturing and just the general manufacturing process improvement. And the thing that allows us to do this better than anybody else is that we have really detailed process execution data. And that's based on all the information that's gathered or related with the process instance. So, if you can put all the details about, let's say, a loan application, all in one area, how long did the user take in order to submit the application? How many errors were there in the application? How many times did it go back for corrections? How long did the system take to process?
Eric Lundberg (05:05):
How long did the loan application officer take to process it? How long did it wait before hitting the loan application officer? All of that information is automatically tracked. You don't have to lift a finger to get it. And we want to take that data and give it to you when it makes sense to you. So, whether you're modeling or whether you're implementing, or you're monitoring your process, these are all times where having some information in front of you takes the guesswork out of what you're doing. So, that's, I guess, the first part. I'll stop here to breathe for a sec.
Josh Wulf (05:36):
From the description you've given there, optimize literally sounds like the perfect name.
Eric Lundberg (05:41):
That's what we're trying to do.
Josh Wulf (05:43):
Yeah. Business intelligence for automated processes, Camunda platform stack, we have Camunda cloud. And so where does it fit into Camunda's kind of tech stack and business model?
Eric Lundberg (05:54):
So, Optimize is a premium add-on for Camunda that fits the business intelligence role. So, that means that yes, it's paid, it doesn't come unlimited for the community version or Camunda Cloud self managed, which also has a free option for non-production usage. Optimize helps your business processes make, I would say, 50% better, 100% better. And in those cases it's definitely worth spending the time, spending the money in order to perform that analysis and get it. And it's different because it's made for BPMN. Instead of a lot of our competitors - think Tableau, Microsoft Power BI, Excel - we really leveraged the information that you pull already from the engine, task list, et cetera. You have lots of different available information like user task data, flow node data, decision data, process instance data, variable data and incident data that are all rolled into one.
Eric Lundberg (07:02):
And so it automatically pulls the data from your engine, from task lists. It uses the model that you've been working on in Modeler, and it takes all that information together and it packages it in a way that makes sense and it's readable. And we've spent a lot of time trying to get the most value out of the data that we've pulled together along this journey. You'd probably put a team of data engineers on it, so they could spin up their own elastic search database and set up their own queries, and it still wouldn't be able to be as powerful as Optimize.
Josh Wulf (07:37):
And how does it compare with the Camunda platform? Is there a comparable component in that tech stack?
Eric Lundberg (07:46):
I would say that Optimize is available and it's actually the same code across both Camunda Cloud and Camunda platform. There might be a little bit of a delta in terms of the features that are available. There's some certain things that you can do in the cloud that are really cool like scalability, reliability, providing a clear and smooth user experience as you flow between the different applications. And that's in terms of how it goes for the cloud SaaS function. In terms of Camunda Cloud self-managed, I might be going out on a limb here to say this, but it's basically like Camunda platform in your own self-managed cloud. So, if you have a Google Cloud or an AWS, and you want to run Optimize there and run it off of the other cloud components like Operate, which is the cloud analog to cockpit, then you could run that. But I guess in short, optimize is the same application between platform and cloud. If you like them in platform, you can port it easily.
Josh Wulf (08:55):
And are they exactly the same between, or are there differences between the one... Like if I've been using it in platform, looking at going to cloud, is it just one-to-one? Or are there any kind of differences that I should be aware of?
Eric Lundberg (09:10):
I think there's a little bit of a difference in terms of the setup, and basically everything around the core user experience. But the core user experience is setting up reports, setting up dashboards, getting those insights. Everything on the periphery, the setup is going to be a little bit different. It should be a little bit easier on the cloud, I would hope, because there's a different degree of flexibility. There's also, in some cases, a little bit less functionality right now on the Camunda Cloud version. We're still building it out and that's going to be the successor, but we add a lot of functionality to the Camunda platform version. And we're still catching up a little bit. You can expect the setup experience for Camunda Cloud to be, I would say, dramatically different. So, Optimize automatically connects to the engine, imports your data.
Eric Lundberg (10:07):
Self-managed, we're trying to make that a little bit easier, but I think the main difference from the user's perspective is going to be in trying to have a coherent experience across the entire Camunda application suite. So, let's say you're viewing a dashboard in Optimize, you get an idea on how to improve that process and then you jump directly into the modeler. You commit those changes, you hit execute, you run a test process instance, and then you can validate almost immediately that your idea was able to make an impact on whatever business metric you're trying to improve. And so the goal here is to try to make the development of Camunda projects a little bit more agile, so you can rely on these metrics that you care about.
Eric Lundberg (10:56):
I care about how fast my process is, I care about the incident percentage or the percentage of process instances that have an incident. And you can use that to decide: “yes, my project is done”, or “no, my project is not done”. And, hopefully, that can help lower the barrier to get the first version out into production. So, you can test to gather data and make sure that you're performing well. But also later down the line, you can take your data that you already have and quickly iterate and push out new versions that allow you to get to a stable performance version of the process in as little time as possible.
Josh Wulf (11:37):
I could hear in the background like your military experience really has informed your appreciation to process automation and analytics, and the whole kind of space.
Eric Lundberg (11:49):
Yeah. There's definitely a strong process culture in the military, because you need everybody to work the same way. And that's not because you need to be able to look at people marching in uniform and you need compliance, but it's for a reason and that's trust. If you know that every time you push a button on your phone and it does the same thing, you can rely on that phone to do what you expect it to do. In the same way, if you have an organization of 1,000 people in a life-or-death situation, you need to be able to know that you can rely on somebody. And so you drill really hard in order to get very consistent with working with these processes. But that's a really, really expensive way to do that. You could just automate things a little bit better, and computers are very, very reliable.
Eric Lundberg (12:42):
They're very consistent. They do exactly what you tell them, even the bugs that you program in. And so since you have to think about everything working exactly just so, you can spend a lot of time to standardize everything that must be standardized, and then only use people's judgment for where it matters. And so for basic or simple or mindless tasks, you can and you should automate as much as you can. There's still always a little bit of human judgment. And those are the edge cases. That's instant handling.
Eric Lundberg (13:16):
Those are the process instances of like medium-risk applications where you need to go to a manual review. And what you should be able to do with Camunda is try to reduce that number as much as possible. Optimize tells you naturally, because I'm always going to go back to Optimize. Optimize tells you what percentage of instances that is. And you can take a look at the data that's captured in Optimize, and you can refine your DMN table so that the mindless tasks never see humans’ eyes. They can be completely performed by a robot or by software robot, but really it's done by the system. But there's no reason that you need to put human eyes on a simple task.
Josh Wulf (13:59):
You were saying about being an AWACS air battle controller, about dealing with like headstrong personalities. And that's like dealing with developers. You want to revisit that?
Eric Lundberg (14:10):
Yeah. So, I think there's a maxim in software development that “either you manage what you do not understand, or you understand what you do not manage”. And in that same way, and then that's specifically around coding. So, as a product manager, I don't know how to code production-level Java. I don't understand a lot of the technical intricacies of what's going on and I have to rely and trust the engineers that they know what they're doing. But by that way, I'm kind of the business person in that case. I care about the business metrics that optimize the user experience, what the person interacts with. But a lot of things that are under the surface, I don't know, nor do I necessarily need to know in order to be effective. As an air battle manager, that's one of the skills that you have to develop.
Eric Lundberg (14:54):
I can't tell an F16 fighter pilot to, let's say, turn left by 10 degrees, such that he can avoid a missile that's being shot at him. I just tell him, "Hey bro, there's a missile coming at you. You should avoid that." And a lot more technical language than that. But there's this inherent trust that you have to develop between the person with the big picture of what's going on and the person, let's say, that's looking through a soda straw and has really, really strong technical skills, who knows exactly what's going on, but doesn't necessarily understand the big picture. And that communication and trust is critical. And that's hopefully something that Optimize can help build for the developers and business people that are working with Camunda.
Josh Wulf (15:41):
Okay. So, Optimize kind of gives the big picture of the business and you can direct developers to the different parts of the business and the process to work on sections of...
Eric Lundberg (15:55):
Yeah, I'd say it works both ways. So, that data is the uniter between the business and the technical folks. So, for the business, that data is the detail about what exactly is happening. So, you might say: “person applies for a loan, seven days later it hasn't been approved and they're unhappy about it”. That's a story that's really, I would imagine, nerve wracking for somebody who's responsible for that application process. But if they knew that along that process that it was sitting for five days waiting for the customer to provide additional documentation, and that was swallowed by spam, that's something you can you use. And so Optimize provides that information, that really detailed information about what's happening in order to help the business person understand. And on the flip side, you also have information that can help the developers make better choices. So, if as a process owner, I say, I need this process to run twice as fast.
Eric Lundberg (16:54):
Most developers would just say, okay, cool. I'll do what I do best, which is to automate, make my scripts run twice as fast. But it makes no difference if you make a script that runs in one minute 30 seconds, if the entire process takes a week. You need to find the point where things take days, not minutes or seconds and focus on that. And that provides you with an opportunity to focus on what the root cause of the issue is and tell the developers. Instead of improving what we already have, maybe addressing tech data, improving some scripts, we need to automate this new step and here's why it's valuable. And this is why you have, I guess, the resources and the time that you need in order to fix it properly.
Josh Wulf (17:38):
Yeah. I can see the link between your experience in the military and this kind of big picture thing. Yeah, there's no kind of high stakes business to be in and where that kind of process-
Eric Lundberg (17:51):
Josh Wulf (17:51):
... improvement would make a difference. A user in the community looking at going from that brief production use to the paid offering asked me, is it possible at the moment to - because Optimize is business intelligence, the processes - to fuse the data coming from the process engine with data from other data sources in their business. Is that possible?
Eric Lundberg (18:16):
The short answer is “Yes”. The longer answer is: “How?” And there are two main tools that we have for that. The first tool that we have is “external variable ingest”. So, let's say, I'll go back to loan application. You have some information about the customer, let's say the loan amount. The platform really doesn't care about that. But if you're trying to make sure that for your highest performing or highest tier of customers, let's say people are taking out a loan of a million dollars, use American currency for sake of argument, that they have a really good experience and you have an SLA for them of like a 15 minute application period. You can actually take that loan amount and you can import it into Optimize. It can use that to perform filtering or perform visualizations off of, and really leverage the information that's stored elsewhere.
Eric Lundberg (19:17):
That's what we call variable data. The second option that you have are embed based processes. So, Optimize is designed for end-to-end business monitoring. And that end to end monitoring doesn't matter or doesn't care if you're automating through Camunda or a different tool. And so you can take the entire process, and if you're using a legacy tool, if you're using a different tool, you can pipe those events in and you can visualize exactly what's happening without having done the multi month process to code, test and ship a Camunda process yourself. So, there's a lot of different ways that you can pipe in data to optimize. They require a little bit of work, but if it's valuable to you, then it's definitely worth the time.
Josh Wulf (20:06):
Yeah. Everything takes some work, right? And what I'm hearing is “it's possible”.
Eric Lundberg (20:11):
Josh Wulf (20:12):
That's awesome. It reminds me of those heat maps that you have for websites and tracking people's flow through that. So, you can track people's flow through your entire business process, both the processes that are automated with Camunda and parts of your business that aren't necessarily automated with Camunda, but are involved or touch your whole business. So, you’ve a view of your whole business there with that setup.
Eric Lundberg (20:37):
Yeah. And what you're describing are user tasks. In my mind, the easiest way to automate a business process is to start with a completely manual process that you just pluck off the shelf. And the first thing that you need to do is figure out what's going wrong. And quite often people spend lots of time in order to automate a process when the existing process wasn't really that much broken. You could take a series of manual tasks and perform some zero code offline improvements that still achieve your business goals. So, in that case, Optimize isn't just: “how do you make your automated processes better”, but “how can you make any process better?” Specifically, if you wanted to improve this, or improve a given process, just model the entire thing with nothing but user tasks. And that way you can see how quickly things flow through the workflow and you can actually document what's going on. Once you realize the highest ROI, return on investment, nodes to automate, just automate that node. And that's a really easy, low barrier way to make a process better.
Josh Wulf (21:55):
Yeah. That's incredible. Because I was thinking of Optimize as “you automate everything, get it all set up, and then at the very end you now optimize the process”. You're saying you can do it the other way round: do everything manually, and then just effectively log what's happening.
Eric Lundberg (22:11):
Josh Wulf (22:11):
And then look at that to see “which parts do we automate, which ones are going to be the ones that are going to give us our biggest bang for buck?”
Eric Lundberg (22:18):
Yeah. Because the stakeholders - all IT managers that all the developers are working for - they don't really care if a task was performed by a human, or if it was by a robot or if it was reform by RPI. They don't care about the methodology. They care about the results. And the results are happy customers, low cost. And there's a lot of different ways to get there. So, where Optimize fits into that role is it uses hard data to change the relationship of business and IT. It's not about shipping code, it's about moving metrics and you can use that to create a data-driven culture of continuous process improvement that helps you actually get to those business goals.
Josh Wulf (23:00):
And what about the future plans or roadmap for Optimize? Given that you're the senior product manager of Optimize, what kind of insight can you give us into that?
Eric Lundberg (23:11):
Yeah. We've been spending some time this last few months talking about what the future looks like, and I've already shared my vision. That's like a two to three year, it's aspirational, it's directional. It tells you which way that I'm taking the product. In terms of something that's a little bit more tactical, I can share the themes for 2022. And the tagline that I share is, “do less work, let Optimize handle the load”. And there are a couple different loads that I'm targeting here. The first load is a mental load. You don't have to be a business analyst in order to get insights about your business processes. You don't have to be a data engineer in order to get your data together for analysis. And so it's taking a lot of the burden off of those personas, such that a smaller team, a more general team, a casual user can still get valuable insights out of it.
Eric Lundberg (24:03):
And I think of this in four ways. The first is through immediate insights, and that's the ability to generate actionable insights from your process data within minutes of getting started with Optimize. This involves two things. First, getting the data in quickly, and we're making sure that the setup process is as smooth as possible, but also making the learning curve smoother or shallower - so you don't have to spend as much time learning how to use the tool. The second one is passive monitoring or in other words, set it and forget it. Where Optimize tells you when something important happens. So, if you're trying to track a metric, there is absolutely no reason for you to put up a television and then assign someone to stare at that number the entire day. That's a waste of their time. You can do that. The dashboards look pretty.
Eric Lundberg (24:55):
You can view them in full screen and dark mode, because everybody likes it that way, but it doesn't necessarily mean that you have to. And so we're working on the ability to have metrics of alerts and notifications when something interesting happens. And that definition of “what's interesting” differs from person to person. So, we're going to have to get smart in terms of figuring out what is interesting and what's worth sharing with the user. The basic thing is: let you track more metrics and get you emails when something interesting changes. That brings us to the third theme for 2022, which is process intel. And that's to receive actionable recommendations that help you achieve your goals. I'll break them into two parts. The first part is actionable recommendations. Let's say you're in your car and the gas gauge is low and you get a gas low indicator.
Eric Lundberg (25:51):
That's actionable. You know what to do. I see a light, I go to the gas station, I fill it up. If you see something where the process is not performing, okay, cool. What do I do about that? And so trying to get to, what is something that you can do in order to help you achieve your goal, and then package that in such a way that is low touch or doesn't take much time and then hand that off to the user. That's the actionability. And the focus is on getting you to achieve your goals. There are typically, I would say, three types of goals. There's time, cost and quality. So, time, you want your processes to run fast. The number one reason why people automate processes is so that those processes run faster. Cost is the second one. There are a couple different ways you can measure cost. Quite often, it's in labor.
Eric Lundberg (26:42):
So, go back to time. Another one is if you do an outsourcing of process, or if you use a different type of labor. You don't necessarily need a manager reviewing everything if you can empower the people on the front line to make decisions using rules. And the last one is quality. And quality can be defined by how often something is rejected or goes back in the workflow. Usually things are supposed to go from left to right, but sometimes they go right to left and we want to minimize the amount of rework that's done. Another one is the amount of processes or the percentage of processes that end up in a good state. “How often is a loan rejected?” There's probably a target rate, and you want to reject loans for the right reasons, not for something wrong, or sometimes there's a happy path and there's a less happy path.
Eric Lundberg (27:34):
And the last theme is to have an integrated experience. And that's to experience a single coherent interface that takes advantage of the upstream Camunda applications. There's no reason that we have to rework to also implement a feature that's available on let's say cockpit/operate or the modeler. You can just send you with an optimized information and a link back to that section, and have you perform that task. Rather than thinking of these applications as stand alone, think of that as a tightly knit Camunda experience.
Josh Wulf (28:09):
Thank you for giving us not just the bigger picture of the vision, but also drilling down into some of the specific initiatives that are coming down the pipeline. In terms of using Optimize, kind of experiencing the value that it might offer to my business, for example, how do I go about exploring that to see what kind of difference it could make and what kind of licensing is there? How does that all work with Optimize?
Eric Lundberg (28:37):
There are a couple different ways to take this. So, the first one is if you just google “Camunda Cloud”, we've recently revamped our marketing content. So, you can get a clear picture as to what Camunda platform, Camunda Cloud Optimize are, such that you can understand if it meets your use case, if it's something that your company would be interested in. In terms of the trial options, the general gist that you should take away from Optimize is that you have enough time and a free trial to test it out, but not so much that you can use it for production processes for an extended period of time. So, Camunda platform comes with a 30 day free trial, as well as Camunda Cloud SaaS also has a 30 day free trial. Camunda Cloud self managed is a little bit different. Instead, the limitation is not in time, but in scope of use. So, you cannot use Optimize or the premium features in a production version.
Josh Wulf (29:34):
Okay. But I can use them on a test environment?
Eric Lundberg (29:37):
For a test environment. Yes.
Josh Wulf (29:39):
Okay. So, I still have access to them, but I don't have a license to use them on my production data set.
Eric Lundberg (29:45):
Yeah. You can play around with it, but you just can't make money off of it, is the short version.
Josh Wulf (29:51):
And then for Camunda Cloud, I can use it on my production data set, but for 30 days. So, anything else that you want to share with our listeners?
Eric Lundberg (30:00):
Yeah. I'll add my standard plug here. One thing that we're trying to do is getting a lot more iterative and data driven in terms of how we develop this product as well. And so we're looking for people who want to volunteer for user testing of our products and to share our thoughts. And then on top of that, I'm always interested in talking about process analytics. So, if you're interested, if you're an Optimize user, if you're interested in getting into Optimized and evaluating, then you can check out my Calendly at calendly.com/eric-lundberg.
Josh Wulf (30:34):
I'll put that in the show notes for sure so people can reach out to you, book a time to talk with you about Optimize. Awesome. Well, thank you for coming on Camunda Nation Podcast.
Eric Lundberg (30:44):
Thanks for having me.
Josh Wulf (30:48):
Roll the credits. Audio engineering provided by Inclusion Audio Production. Check them out at inclusionaudio.com. And the soundtrack is provided by the legendary Japanese Progressive House DJ Shingo Nakamura courtesy of monstercat.com.