BI disruptions Q/A - from reporting to advanced analytics
- Summary:
- Are BI disruptions being felt in the big enterprise shops? I recently got a field view with BI architect Derek Loranca.
During our talk, Loranca shared how his company is addressing these changes and how he has been personally driven to evolve his own skills. Loranca has been to enough trade shows to acquire a distaste for sexy BI buzzwords like 'predictive' and 'big data'.
Plus, his company already has a deep history with predictive. So what's changed? Why is predictive suddenly in the spotlight? Are their industry applications of these catch phrases that actually make sense? That was the fodder for our discussion, which I've condensed into these Q/A highlights. Note: Loranca's company is an SAP BusinessObjects customer, but does not run SAP ERP.
Jon: Tell us about your day job.
Derek: I work in insurance and managed care for one of the oldest in the business. The Affordable Care Act is big for us - our CEO is out there talking a lot about it.
Jon: In other words, changes afoot.
Derek: We're in the midst of a transition internally because of the external pressures, and then, as an older company, you've got a lot of legacy systems and applications, as well as merger and acquisition apps all coming together. Then you have to figure out how to make the business work on top of all the external changes, and the changing dynamics of BI into analytics, for lack of a better term.
You've got that conundrum of: how do you keep the current going, along with new features, new applications and new data? So it's definitely a crucial atmosphere right now, figuring out where things are going and what horse to tie your cart to.
Jon: About those external pressures – are some of them unique to insurance and some of them broader?
Derek: You've got the pressure of where health care is going. You've got one camp pushing single payer, you've got another camp pushing the continuation of the managed care model. Then you've got the Affordable Care Act in the middle. You've got that pressure of: how do you stay relevant in business with all these other external pressures coming down? Which then shakes up how people handle their information.
It's one thing for companies to say, ‘I'm going to move my business to the cloud.’ That's fine - you can have your data in the cloud. But in a company such as mine, the cloud carries its own set of risks. It's not so much that risk of the data loss and the government scrutiny of the personal health information, it's the public face of it.
You might say: 'You put my lab test in the cloud?' But nowadays, is there really a difference from having your lab result in my system or the AWS system that I own, too? Technically, there's not a whole lot of difference, but publicly, there's a lot of difference. It’s those pressures, and they boil down to how people inside the company running the businesses are trying to access that data. It circles back to a fundamental change of how we're looking at BI in general.
Jon: Do you think BI is helping to solve these problems? Or is the whole point that BI needs to be retooled in a way that helps business users and is more agile, if you will, so it can help solve these problems?
Derek: Currently it’s helping on the reporting side. But I think data logic and data analysis is eventually going into every day work in my industry. Everyone's going to need to figure out how to do progressions and cost runs to accurately forecast things. Obviously, the call taker who's answering your question about something you called about isn't going to need to run that. But the analyst for the plan sponsor is going to need to be able to get out of the mode of, ‘I'm going to run a report’ to ‘I've got access to the data; I'm going to run my own analysis.’ I think we're at that shift now.
Jon: At trade shows, we hear a lot about what I call the sexy side of BI, everything from self-service to big data to predictive to mobile BI, which gets into ‘bring your own device.’ Does any of that stick to the wall?
Derek: As I move out of an operational role into an architecture role, they all relate. The question is - and I think any large enterprise has the same problem - how do you shift that paradigm without a clutch, as Dilbert put it? How do you go from first gear to fifth without pressing the clutch?
Mobility's great. But how's mobility really working for the data analyst who has a desktop computer? Or the executive who says, ‘I'd like to see this report on the mobile,’ but they don't have a corporate mobile device, or their personal mobile device isn't provisioned with a VPN or something else to see the corporate data? In 30 years, we're going to be wondering what took us so long to allow for seamless desktop-to-mobile-to-cloud consumption of that data.
Jon: One example is visualization – visual data is all the rage, but you see a limitation.
Derek: Problem one is: how do you access the data? As the data grows on the relational systems, it gets slower and slower and runtimes get longer, and you can only speed that stuff up so much. Then you have the sexy buzzwords of visualization, the tag clouds and so. People are like, ‘I want that! I'm going to put it on this 50 billion row table.' Then they say: ‘Why is it taking so long?’ You roll your head and go, ‘Well, that's because you're doing an advanced analytic on a 50 billion row table.’
That's where the SAP HANAs and IBM BLUs of the world come along, and say, ‘We'll just take your data and stick it in-memory and it’s going to be faster.’ That solves part of the problem, but if you're not going to change the way you're analyzing that data, or the tools you're using to analyze it, are you really solving a problem, or making your current problem faster? It's like putting better tires on a sports car.
Jon: Do you have an approach to make sure that when you pursue one of these newfangled BI products, it is relevant to users? Are they involved in the process in some way?
Derek: A qualified yes. We're slowly getting there. Just like most large enterprises, we’re got certifications to live up to and a full waterfall organization around that. But the business is saying, ‘That worked great for the last 15 years, but now we want to give you something small to start.’ Or they want to get involved in a design thinking kind of approach.
The question becomes: where does it make sense to pick apart the projects? What we're pushing in our team is that with BI and analytics, it makes sense to be agile. There's no need for a full waterfall methodology when you're working with the business to design a semantic layer, or to create a new data set off an existing dataset.
That's where an agile or design thinking methodology works much better, because you're able to sit one on one or team and team, figure out where your gaps are and what you need to build. Otherwise, it really becomes bogged down into ‘We need to change the font,’ and ‘Well, you’ve got to submit a change request.’ Next thing you know, you need to have a product inspection and three QA guys, - all this for changing the font or a logo on a report. Or changing a minor dataset on the data that’s already there.
Jon: You described your shop as going through a journey from a Crystal Reporting shop into an advanced analytics company. How does one make that journey?
Derek: Very slowly. Given the size of a company at an enterprise level, you've got your corporate methodologies and bureaucracies and approvals, but it's more than that. It's the business driving things.
Jon: But that means you have to turn this into a series of smaller wins. rather than a massive multi-year transformation, right?
Derek: I think that's how you have to do it, to be honest - that's how you get buy in. Those smaller successes drive the bigger enterprise project. We're not an ERP shop in the classic sense; we’re not making widgets. We have our own enterprise processes, but they're all home grown. We’re still running big iron.
What you've got now is business units coming to you and saying, ‘I think this is going to increase my productivity’, or ‘Analyzing the data this way is going to make us money.’ That's where the grassroots of people saying ‘We need this change,’ is coming from.
Derek: As a BI professional, we're all experts in certain tools. For me, it was BusinessObjects, and still is, to a point. But if you're not staying current with the overall market, you're going to get pigeonholed, and next thing you know, the new product comes in, and the product you know is down to three users. They're going to wonder: ‘why are you still there?’
Jon: Let’s dig into predictive. You’ve been an SAS shop for many years. Why do you think predictive is suddenly gaining so much market attention and momentum? Have we turned a corner with what we can do, and if so, why?
Derek: I think we have turned a corner. Believe it or not, I think the corner is the whole big data movement. I think it's the fact that, I can now have millions and billions and trillions or rows. Whether it's Hadoop, HANA, or BLU, I no longer have to worry about reports running for six and a half hours, I can now do those predictions and those algorithms in one quarter of the time of I used to do.
Jon: Does that have potential in your industry?
Derek: Absolutely. In our industry, it’s more cost effective for people to be healthy as opposed to not healthy. We've been doing disease management. But now it’s about analyzing health data to help predict and work with people to have a healthier lifestyle. That healthier lifestyle obviously saves you and your employer and the insurer and the entire system money. Or you could mine the data to predict whether or not someone might possibly be non-compliant with their meds, like a Type I diabetic who hasn't filled their insulin script in six months.
Either they're going somewhere off the record for it, or they don't have the money for it, and maybe an outreach person can reach out and figure out if they need help in paying for their medicine. Because you'd rather pay out the cost of the medicine that keeps them in check rather than cover the hospital stays and the inevitable downfall of somebody's health when it comes to a disease like diabetes.
Jon: Right.
Derek: As analytics speeds up and the data speeds up, what you've got is the chance to fundamentally shift where things like that are predicted and where they're going. But at some point, there's going to be a level of trust. We have to overcome that, and I'm still not there yet. Do I trust the entire ecosystem with that type of detailed data?
Five years ago, when you talked about visualization and analysis, you were talking dashboards, but now we've moved beyond dashboards into data science and full blown, large data set analysis. That's completely different than we were five years ago. People are starting to want that now. So the question becomes - to people like me and other BI professionals - how do we actually implement this with our current stack and with our current dataset, and if it's not going to work with the current stack and current dataset, how do we either make it work and tie it together, or do we build it out new? That's a conversation in and of itself.
Jon: What about your own career transition to BI Architect?
Derek: I've been doing a dual role for the last three or four years where I've had a small team and my own individual BI application for halt the time. For the other half, I've been doing this architecture center of excellence type role.
It was a matter of timing, job security, and seeing where things were going. Where I want to be career wise is helping to pick the products, become an expert in the product, and help implement the products, as opposed to actually building the products. I think that's where I wanted to be, so it kind of worked out.
Jon: Do you find yourself mentoring younger BI professionals?
Derek: I do, and I really like it. There was somebody in my role before who did that for me. We became good friends, and unfortunately, he passed away of a brain tumor a few years ago. For someone to do that for me even when he was sick, the least I could do is help the next person who wants to learn. I have that with some of the younger folks. I don't think I'm that old, but these guys can code circles around me nowadays. I wish I could do that 15 years ago.
With DevOps coming on strong now, pushing developers into doing operational roles, you've can’t be heads down anymore. You've got to be able to think clearly and think broadly. You've got to conduct your own career, and if you're not the conductor of your own career, you're just in the back playing triangle.
Links to my full conversation with Derek Loranca: BI disruptions video, BI disruptions audio-only
Image credits: Decision Choose Change or Same Old Street © Michael Brown - Fotolia.com, Derek Loranca photo in gray shirt, Greg Myers,
Disclosure: SAP is a diginomica premier partner as of this writing.
