 Live from Orlando, Florida, extracting a signal from the noise. It's the Cube covering Pentaho World 2015. Now your host, Dave Vellante and George Gilbert. Welcome back to Orlando everybody. This is the Cube we're live here at Pentaho World. Chris Zekin is here. Chris gave a keynote yesterday. Chris, it was awesome keynote. I mean it was like drinking from a fire hose listening to you. It was absolutely fantastic. He had a couple of epic statements, tentpole statements. We are not an out-of-the-box BI tool. Strategy should be consistent. It shouldn't waver. So I really enjoyed it. I learned a lot. It underscored the depth of the Pentaho platform. So first of all, congratulations on managing such an impressive set of capabilities. Thank you. How do you feel? I feel terrific. This you know, we had a great Pentaho World last year and I really feel both our partners, our customers, and ourselves really stepped up another notch this year. It was a tremendous event. You're absolutely right about a couple of those key points that I tried to hit home at the keynote address. Number one, we're in the enterprise. This is now enterprise-grade big data and it's real in our customers. And so we stepped up to make sure version six continue to make sure it was enterprise-grade from lineage to life cycle to the way that the system runs and a big commitment to upgrades because a lot of our existing customers want to take advantage of these new features. But it can't be hard. You've got to be able to get them to the new version really easily. And we took that to heart and made version six really a nice smooth upgrade path. You mentioned the strategy. Yeah, the strategy didn't waver, right? Last year we talked about govern data delivery. How do you balance that? An end user can have self-service with the right IT governance and not pendulum swing between those two extremes like the market has done over many, many years and govern data delivery is that vision. Streamline data refinery is one of the capabilities that serves that mission and that's what you heard earlier in some of your interviews for example with FINRA and how does an analyst say I need some data today and not wait for a long time to get it and that data needs to be blended. It needs to be merged. It needs to be cleansed. It needs to be enriched with predicted predictive algorithms. So how does that user ask for that kind of insight quickly and have it delivered at their point of impact? So embedded into their experience. So that's been the mission. We didn't change the strategy. We simply executed. Focused on building that product line to that enterprise grade ability and you're spot-on about the the comment of out-of-the-box BI. You know there's an absolute need in the market for reporting and analysis and dashboards and we do that. But every customer wants to twist it and turn it and customize it and instead of being in your day job and say well I need a report and I go out and I open up a report writer the analytics should be right in front of me It's part of that call center app or it's part of that medical application or that marketing application And so we focus on embedding the analytics into the context of whatever that person's life looks like on a daily basis And that's a paradigm shift from maybe the old way of doing it with a standalone report writer So we err on embedding We err and put more effort on extensibility So a customer can tune it and customize and bring some of the unique elements of their world into the solution And the hardest part is all about data blending. How do we connect big data? Plus traditional data and do it at scale and do it in that balance of governed data delivery That's been the real kind of nut to crack and really pleased that version six moves the dial aggressively in that direction I want to follow up on a couple of those themes and what I want to pick up on I love thinking about the history of the technology business and if you go back to the early days of big data When people started to realize you guys started, you know around the same time as a dupe when people start to realize The momentum behind a dupe in 2010 the big legacy vendors most of the people in the organizations What's a dupe? They didn't even know when they started to realize the potential of big data The big legacy guys came out and said we're going to harden big data We're going to make it enterprise grade and I remember at the time talking to John Furrier my partner saying well, isn't like Cloud cloud era and Pentaho aren't they in the best position to do this because they started the whole thing Yeah, and that's exactly what happened. I wonder if you could talk about that. Yeah, absolutely It is what happened and you know, we're great partners with cloud era and Mike was here doing part of the keynote address And we partnered for a long time with not only cloud era But the entire ecosystem because we do realize that customers have choices and they're going to choose different distributions Sometimes it's cloud era. Sometimes it's Horton. Sometimes it's EMR on the cloud And so our solution has to be heterogeneous It has to be future-proofed because customers may choose different paths in the future And that that maturation that's happening is accelerating There's always new keywords and buzzwords and new technology, you know flying at us from left and right Part of our strategy is to make sure that we have enough capabilities to protect our future users Start investing here, but if you decide to change in another direction of technology Can our technology keep up? Is there a future-proofing or what we call an adaptive big data layer Insulating our users from those kind of innovations that are quickly happening around us Some of those innovations materialize into beautiful things and others are short-lived and we need to be able to handle both of those But cloud era is enterprise hardening. They're in the enterprise absolutely scaling up every day securing more every day and Tahoe's doing the same and we're seeing it where enterprise clients are running 24 by 7 operations with big data and It's not in one department. It is an enterprise wide initiative. It's treated like any other strategic asset in their organization And it's allowing them to break through Just because they can get a customer 360 degree view or a patient 360 view or they can get unstructured and structured data together Or they can ask for data on demand that dynamic blending that we do and not always have to wait For the perfect warehouse to be formed. It's creating that agility for a business And that's really the end of the point, right? How can you get insight not data? How can you get insight from that data into the user's hands quickly and with trust though and with confidence and I'm glad to Mention others because it wasn't just you know cloud era the Hortonworks map bar participated many many many others I don't want to want to leave them out Yeah, the other point I wanted to ask you about help me square the circle on you've got the capability to customize Which is unique and we'll talk more about that and the different parts that you do the areas that you play in But then there's the scale piece you mentioned that yesterday and a touch is part of that so When you said that the customization is like, okay, how do you scale? Yeah, how touch is one piece of that obviously, but talk about the challenges of scaling Yeah, and accommodating high degrees of customization. That's a great great question And I don't know if you remember on my keynote There was an interesting little diagram of the the weighted scale of the big rock the tough stuff to move and the smaller Rocks that are agile and I started with that exactly Striking that balance is super critical If all you do is small innovative little things, but you never enterprise grade them It won't run 24-7. It can't be trustworthy so you need to move the bigger rocks of things like security and scale and Performance and administration, you know, they may not be the coolest features, but they're essential features for an enterprise So you have to invest appropriate there at the same time How do you strike the innovation cord so it continues to drive fast and furiously and in the case of Pentaho and Hattachi, there's this unique blend of our data labs that we have big data labs And those are labs that are allowed to explore new technology. They're allowed to innovate fast and furiously Some projects will come to life and some won't they're allowed to fail and they're allowed to fail often But fail quickly So there's an innovation incubation chamber that can happen that way and then in the product group where I am I can adopt those Technologies and product ties the ones that matter the second kind of innovation is because our platform has an open source heritage quite unique in the marketplace we have extension points and Partners and customers can add to our platform because we can only shoulder so much of the resource and the development experience So I'll give you an example Melissa data a great partner of ours did an announcement a couple of days ago They they created a set of data quality rules and data quality snap-ins that show up in Pentaho data integration as Just part of our experience if you want to do a data quality or a data profile Capability snap it right into the interface. We didn't have to build it our partner build it And those kind of extension points are all over the board So another example we have great charts visualizations in the box But I guarantee you the next day someone will say well, I have a special chart Can I use a special chart from let's say D3 the answer is yes You can integrate your own bring your own chart and customize it and put it inside our experience Bring your own predictive engine because I like are or Python or Weka or MLM Bring your own predictive engine but plug it into our environment so we can orchestrate it And so that allows us a lot of flexibility so that customers can both customize But also accelerate innovation at the pace that they require it at without waiting for us to do all of the you know The development work, but we have to own the main development team the main core that enterprise grade platform that orchestrates blending data Enriching the data whether it's predictive or other kind of enrichment quality and then Visualizing that data and again our visualization strategy is to make sure that visualizations at the point of impact Embedded into your portal embedded into your application that you run every day Not some foreign object in another standalone interface. That's where we you know strike that balance and to us the sweet spot Data analytics or data integration blending and connecting data With embedded analytics when those two fuse together The value comes to life and that's what we've been focused on for the last two years and version six is just a great leap forward of Manifesting that visual well, but that's what's interesting about to me about the strategy because you could have gone after visualization or you could have just gone after data visualization and Just focused like a laser like every VC one and others have and and and and grow faster Yeah, but you chose to solve a harder problem, right via platform play Yeah, it takes a lot of discipline your spot on there And some of the marketplace just does great visualization work and good for them And some of them do great self-service where an end user can start slicing and dicing and discovering good for them Equally ETL vendors focus sometimes just on ETL and good for them. We said when the two combine You need to bring date integration closest to the business user and that point of impact And so we chose to go in that tougher part combining both date integration and embedding analytics on an enterprise grade platform Which means tough investments to really make an enterprise grade. You're right that it was the road less traveled And I think it's going to pay off and I think we see that from our customer base here at Pintaha world and beyond that The ROI is showing the benefits are showing the successes there And I've been a lot of user conventions over my 25 years in analytics and some of them are not so friendly There's sometimes customers upset and frustrated this is year number two for Pintaha world and There is a passion in the room. There's a success in the room by all of our clients and customers So I think the road less traveled is paying off building an easy button I called it enough for analytics, which is really really hard. Go ahead George. So One of the things we hear from customers and in terms of the use cases where they're trying to push boundaries With Hadoop. It's this wonderfully innovative ecosystem But that's it's also this ecosystem. That's wonderfully complex. Yes, but you put now this end-to-end platform above it So the raw Hadoop user has a customer journey, right? How much accelerated is yours and Where are you seeing them? Hitting the boundaries now and where are you pushing those out? It's a good question the Setting up the environment Needs to continually become easier, right? Will it ever get to the big red push big easy button right the end of the day these are complex environments, right? But it has to become easier We have to do our part in that story too because it is an end-to-end right and running Hadoop for its own sake doesn't bring value Putting data into Hadoop doesn't bring value until you add analytics and draw the insight out of that environment So this is a full top-to-bottom experience And so we've been doing our parts especially in version 6 to continue to simplify that so when a cluster is defined It should permeate across the Pentaho platform. I shouldn't have to define it twice or three times It should just be a simple graphical interface Versus maybe the original days when it was more of a command line in a script Which is perfectly fine in the beginning era when developers are exploring and playing but this is now commercial software And so the graphical interface is required so things like that push the easy button in the right direction But the other part I think that's really unique for us and critical is this Path we call streamlined data refinery or this blueprint we call streamlined data refinery We've got a series of these blueprints that are true to real customer stories Customer 360 monetize my data and make it a business for me Internet of things, but streamlined data refinery is one of those and the blueprint was great on paper and you could Implement it at a customer site the hard way a lot of people time a lot of effort time We've been codifying that into the product So the streamlined data refinery capability is more and more automated every release that we do so as I give an example This is where you said let's chip away And is this related to the the modeling the inline modeling exactly related to the inline modeling and more So if we think of an end user case, I'm a user and I need some data It's not well formed in a warehouse. It's not that I'm simply filtering a report. That's traditional I don't have the data blended the way I need it. It's spread out in different data warehouses data marks It's in Hadoop. It's in no sequel How do I request that? How does the Pentaho system connected and blend it dynamically? How does the auto model say, you know, this is a date dimension and this is a hierarchy and auto model it for me and Publish it to the user without calling it Now it has to be responsible for the guardrails and setting up the environment making sure there's the right kind of parameters So, you know the end users don't run with scissors and hurt themselves There has to be good controls, but you want to enable that self-service So that's an example of making it easier by taking a use case instead of making the whole technology easy Which is impossible we take it by use case and streamline data refinery needs to ingest data blend data Auto model auto publish and land it at the point of impact for the user without calling it Makes it simpler, but it is tough under the cover. Don't get me wrong complicated But that's where the Pentaho technologies allow that Intellectual know-how and that smarts to be applied by both our technology and then the user So I'm glad you clarified that Chris on the easy button because it's not easy But making it easier, but there's a parallel that George and I have been talking about with the cloud guys They're trying to develop an end-to-end data pipeline and deliver it as services Yeah, you guys have the end-to-end data pipeline. You're good partners with Amazon. Absolutely customers love it, right? Yeah, is there an eventual collision course there or is it so far off? No, I think far off where we're there's a big Ecosystem and even if you take the next wave of the Internet of Things the trillion dollar marketplace There's lots of room for lots of us to play and there's lots of You know more innovation to happen in many areas So I don't think there's collision courses like in that particular area. You're right about the cloud Many are embracing the cloud Full in two feet down, you know jumping in you heard from FINRA You heard from NASDAQ as examples of that where they want the elasticity They want the storage and processing on demand And they're using Pentaho for that streamlined a refinery or for analytics but they deploy it and process it and Hadoop on the cloud and It's it's a beautiful story that again We know all the savings of cost of you know savings of cloud computing and capex versus op-ex and all those No kind of things but it really is in the mainstream now It's not a hypothetical on a PowerPoint. It is in production in massive organizations now and Massive is big data is now massive and eventually we'll get accustomed to that and it'll just be data And then we'll reach the next level above it, but cases like FINRA 75 billion Transactions per day through Pentaho and into the Amazon cloud and then imagine trying to find the needle in the haystack Of all that data. That's what they're doing today with combination of the cloud Pentaho data mining and predictive algorithms and it means value to them It's turning that data finally into some insight because everybody knows the little phrase that we've been data rich But information poor We're now getting to the point where insight and information is actually seeing the light of day for these companies And it's I think it creates that atmosphere. We talked about at the beginning of the success. We're seeing in this room frankly So quick question on Because what we've been wondering about is you know as I was asking about earlier with Hadoop You know having a fair amount of complexity because it's a bunch of independent projects Like but taking that streamlined data refinery concept Layering it layering it on some of the Cloud native services That's still great value add on top of them, and I would imagine Pushing out further would be more advanced analytics. Yes and making that more self-service Or maybe even to predictive or prescriptive. Yeah, the The orchestration of data is right in front of us, right? We're blending and we're orchestrating that data that enablement of a balance of governed data delivery that self-service with IT Absolutely alive in front of us and you're spot on that Looking in the rear view mirror is not enough. You need foresight. You need to have predictive capabilities to prevent churn predicts that something will happen in maintenance of equipment before it breaks down right so that predictive or advanced capability is Another part of the equation a big part of the equation We have a flexible way to do that today. We call it the data science pack and it's a like your cell phone. Bring your own We're not going to prescribe that you must use our must use weather or Python bring your own and how do we orchestrate it? Because we got to get the data. We got to blend it and mash it up We've got to enrich it with those predictive algorithms, and then we have to publish it to the point of impact That's the full pipeline so predictive is an important part of that enrichment phase But again, we live in an open heterogeneous ecosystem So leveraging the best of all these kind of technologies and can they be extension points or snap-ins to our platform? And the answer is yes, but more and more of these data-driven applications that we're seeing are Predictive in nature. It's to predict and prevent the big equipment failure It's to predict and prevent the cost of that. It's to predict fraud not after it happens when you chase it down It's before it happens Student failures before it happens likelihood of taking your medication before you don't so there is a really large Predictive element to all of these data-driven applications the next sort of frontier It's the frontier and I'd say it's active very active now And I think if I was to project a frontier ahead of it or behind it Whichever order we're looking at the scope from the Internet of Things is what's that next super super kind of charged? Exponential wave on top of the tsunami of big data. It's another bigger one and The Internet of Thing open oh Internet of Things opens up a new avenue of all the edge data Did this light just turn on yes? It did but does it matter well? It does it depends if you're trying to save money It depends on whether the room is occupied a car just stopped in a sensor said that the brakes were engaged Does it matter depends is this a prone area with accidents? Is this a school district is the weather turning bad when you start blending data from your traditional? from the big data and from the edge where the sensors are Imagine what kind of applications the world is going to see as we simplify and automate that so that big next wave Absolutely includes predictive, but that's happening now and the Internet of Things is staring at us right in the face as the huge wave Enable by infinite compute and storage so we're going to be popping up against the keynote soon Two questions one is if you could summarize 6.0 at a high level and then talk roadmap great So 6.0 highlight reel date integration and embedded analytics We've always had that in the portfolio This is an enterprise grade release all of those features that the enterprise requires and loves from administration To data lineage to know where the data came from to monitoring There's a series of all those enterprise grade capabilities Sticking to the same vision because what visions don't wobble right we should stay to a course and we've done that but enterprise grade The platform itself all those administration and security and those kind of things the embedded analytics more Apis more openness so you can embed it even easier Into the into your own portals or your own applications, but it has that enterprise grade capability set in terms of the roadmap Strategy stays the same Governed data delivery Continuing to chip away at making that balance easier between end user self-service and that it good governance And so the roadmap has everything kind of underneath that just continuation of more Simplicity every step of the way can it get even simpler to connect to data Can it get even simpler to model the data can it get even simpler to embed that data? So the next wave of our releases just continue to be on that agenda of enterprise grade because Big data and Hadoop is now mature. It is real It is in the market and you have to treat it like a corporate asset not just some science project on the side So we're doing our part to make sure that the data integration and analytics is as mature and As self-serve and simple as we can make it Behind that and underneath it IoT is going to open up a whole new set of pathways for us And the fact that Hitachi is part of the Pentaho journey is another Exponential driver in our roadmap. You mean Hitachi is now part of the Pentaho family. Absolutely It may be true because every Hitachi application to get that absolutely every Hitachi application every industry Application needs data turned into insight I think we have some great goodies to bring to them to help all of those solutions become enabled with insight. All right We're up against the keynotes Love Chris. I love the energy vision and the substance really came through this week So thanks very much. Thank you for having us and thanks for being at Pentaho world. All right So we're going to stay tuned for the keynotes. We're done here. I'm done as you can tell by my voice But thanks for watching everybody stay tuned for the keynotes and check out silicon angle calm and wiki bond calm for all the research and news Thanks for watching. We'll see you next time