Agile BI Made Easy: Two Proven Paths to Success

Gary Damiano
Hello ladies and gentlemen. I’d like to welcome you to today’s web cast, Agile BI Made Easy, Two Proven Paths to Success. Brought to you by WhereScape USA and Composite Software as part of our data virtualization leadership series of web casts.

We are pleased to have a renowned group of experts joining us today. From Forrester Research is VP and Principal Analyst, Boris Evelson. From WhereScape USA, VP and General Manager, Mark Budzinski. And from Composite Software, Robert Eve, Executive Vice President of Marketing.

Today’s topic is Agile BI Made Easy, Two Proven Paths to Success. You’ll be learning how two methodologies, consolidation and data virtualization will help you successfully implement agile business intelligence within your organization and business. You will also hear critical success factors including necessary components, next generation technology and topology options and some real world cases and examples of successful agile BI.

Please let me share some background information on our panel of speakers. Let’s start with Boris Evelson, Vice President and Principal Analyst at Forrester Research. A leading expert in business intelligence, Boris delivers strategic guidance, helps enterprises define BI strategies, governance and architectures and identifies vendors and technologies that have helped them put information to use in business processes and end user experiences. Boris has more than 25 years of experience with enterprise software and application implementation, management consulting and strategic advisory skills. Hello Boris, it’s great to have you here.

Boris Evelson
Glad to be here. Looking forward to the webinar and to some interesting and challenging questions.

Gary Damiano
Next I’d like to introduce Mark Budzinski, WhereScape USA VP and General Manager. Mark is responsible for WhereScape’s business in the North American market. In his 27-year career Mark has held senior management positions at Intel, Sequent Computer Systems, RadiSys and Applied Microsystems. He holds an MBA, a Masters Degree in Computer Science, a BS in Industrial Engineering, and is a graduate of the Buckley School of Public Speaking. Thanks for joining us today Mark.

Mark Budzinski
My pleasure as well.

Gary Damiano
And a few words about our third speaker, Robert Eve. Executive Vice President of Marketing at Composite Software. Bob Eve’s experience includes executive level marketing and business development roles at leading enterprise software companies such as Informatica, Mercury Interactive, PeopleSoft and Oracle. Bob is a prolific writer who has authored dozens of white papers and articles on BI, data integration and data virtualization. Welcome Bob.

Robert Eve
Great to be here Gary. Thank you.

Gary Damiano
I’m Gary Damiano, Vice President of field marketing for Composite Software and I will be the moderator for today’s web cast. Let’s start our discussion today with Boris.

Boris Evelson
Excellent. Good afternoon everyone. What I do want to talk about today is that there is an interesting opportunity that exists in today’s business intelligence market because from all sides, we do see that the criticality under the mission criticality or business intelligence application continues to go through the roof. And we’ll go, on the next slide, through some of the reasons why that is happening.

But on the other hand what we do see is that the complexity of business intelligence environments, platforms and applications isn’t really, well hasn’t really kept a pace with the tremendous growth of the importance and as I said, criticality of business intelligence requirements. So as a result we do see a huge opportunity. Anyone who tells me that this is a very mature, very stable market isn’t really doing the right level of research. This is a very dynamic, very actually, I don’t want to say immature, but still very vibrant market because there is no such thing as an easy plug and play business intelligence application. We do need to keep addressing this gap, this opportunity with tons and tons of best practices. And equally as importantly with some next generation technologies that are indeed available today. There is nothing that we will discuss today that is not available from the market.

So let’s jump into some details. Let’s move onto the next slide. And just spend a minute or so confirming why business intelligence market is so vibrant and is so critical. Obviously this is no news to anyone; both structured and unstructured data just keeps coming with no end in sight. Regulatory reporting requirements are not getting any easier especially in certain industries such as the financial services, manufacturing, life sciences. We see constant increase in regulatory requirements complexity.

And obviously enterprises these days are growing; they’re not sitting still, they’re expanding globally. They are acquiring companies, they’re venturing to new products and to new markets and so all of a sudden the complexity of the operations is growing as well. And probably the most important critical reason is that many leading, special leading enterprises are beginning to treat business intelligence not just as a back office application but really as a corporate asset that they use to compete and often not just compete but survive in the market because we all know that the profit margins are becoming razor thin. Everything in this world, including products and services is getting more and more commoditized, so competing on the quality of products or the differentiation of services becoming harder and harder. But if you and I sell exactly the same product, but I know what my customers are doing either better than you are or even if I have the same information about my customers but I can get to it faster, then obviously I have a very distinct advantage. 

Let’s jump to the next slide where we will see that what I just talked about is absolutely confirmed not just by tons of anecdotal evidence that I see constantly but by every single Forrester survey and the trajectory of the surveys that show constant increase in the adoption levels of business intelligence in terms of the number of respondents to the survey that are implementing, they’re in the middle of implementing, they’ve implemented something or they are upgrading or they’re considering and upgrade. That as you know is growing by leaps and bounds.

And even better news for everyone, especially the vendors or anybody in this market is that the segment of the market that previously has responded that they’re not interested about business intelligence is dwindling. I know I’m personally going to be challenged probably next year when those two bottom bars on the slide that you’re looking at drop down to zero and I probably won’t be able to show the progress. But I’ll cross that bridge when I come to it. Those were the good news, right, that was the left most arrow, you know, the left pointing arrow that you saw on the first slide in terms of why business intelligence is so critical.

But now let’s talk about the right pointing arrow that you saw on the first slide. Why is BI today typically costly, it’s complex and it takes a long time and it takes a lot of dollars and it takes a lot of skills to implement? But one of the reasons for that, there are multiple reasons, but one of the reasons is just the complexity of the end-to-end business intelligence stacks.

First of all let’s talk about what business intelligence is. There are two perfectly acceptable definitions in the market. One of them is very broad and that’s the one that I’m using here. And that basically takes you on the entire journey that raw meaningless data takes on the way to become meaningful information. So that includes all of the so-called lower layers of the business intelligence stacks such as data sourcing and data preparation and data integration and data quality and master data management and data warehousing, etcetera, etcetera, etcetera.

We can also define business intelligence in a narrower sense as just kind of the last mile to a customer. Only as information delivery or only as information usage. So in that category, a narrower definition, we include reporting, analytics, dash boarding, alerts the [inaudible 0:08:54], text analytics, advanced analytics, I mean even in that narrower segment we can still count at least a couple of dozen components.

So the bottom line on this slide is putting it all together sometimes is a complex and a lengthy process. And one of the ramifications of that complexity is that even after you’ve successfully put everything together, when you need to make a change anywhere in that 40 or 50 component end to end business intelligence stack, any change, even the smallest change can have huge implications. Because imagine if you were adding just a tiny, a tiny new or seemingly tiny, new insignificant column in your source application. Let’s say you decided to analyze customers by a new dimension. Let’s say you decide to track your customer’s profitability, right. And you now added that to your CRM application. So that is probably something, putting one column to an operational application is not a trivial task but hopefully that’s not affecting hundreds of components. So hopefully that is something that you can do within days or maybe weeks at the most.

But look at the implications to your business intelligence or analytics environment. That one tiny column can affect several of your ETL routines or data quality routines. And all of a sudden you need to understand the impact and potentially change the operational data store schemer. And all of a sudden that has an implication on your start scheme on your data warehouse and data marts which will affect some of your calculations like your performance indicators and metrics and measures and all of a sudden before you’re done, you come back to the business stakeholder and you tell them that you’ve got a touch somewhere where they’re changing or at least regression testing, hundreds of reports, dash boards, cubes, etcetera. So potentially any tiny change in business intelligence can mushroom and explode out of control.

So as a result of that complexity, what we do see in the market, and this is a result of our recent business intelligence maturity survey. We put out a survey about half a year ago to a couple of hundred Forrester clients and the survey asked them about 40 or 50 questions organized by the categories that you see here in terms of the governors and organizational structures, processes, data and technologies, etcetera, etcetera and where they rated themselves on a maturity scale of one to five. And as you can see, the averages are below; the average response numbers are below average. And as a matter of fact, you know, my opinion is that these answers are probably on a high end because this did not go out to general population. The survey went out only to Forrester clients. And not just Forrester clients, but basically our research panel. Which means that these are the people who read our reports, these are the people that talk to us. So I would say that the average maturity of this particular client segment is probably higher than the general market. So if a similar survey did go out to general market population, I would say the numbers would definitely be lower.

And the unfortunate side effect of that lack of proper level maturity is that you can’t really take advantage of business intelligence as your competitive differentiator. Remember that second slide and how we talked about where we discussed that business intelligence is no longer just a back office implication. We do use it as a competitive differentiation, but how do you use it as a competitive differentiation if business intelligence is complex, is costly and as a result it’s not really pervasive in the enterprises? If the maturity or a good portion; if a significant portion of your decision makers both strategic and operational decision makers aren’t really enabled with the right business intelligence implications so how can they take advantage of this environment.

So as a result, what we do see, and this is by no means any kind of a scientific study, but we did make some pretty conservative assumptions when we came up with these numbers and I followed people since we calculated these averages and the numbers do hold. So basically what we’re telling you is that in any large enterprise, the number of people that not really have access to BI but really taking advantage, taking full advantage of business intelligence is really in a singles percentage digits number. Which is obviously not enough to enable that competitive differentiation. So at this point I think we are ready to do a poll.

Gary Damiano
The phone question is, “What percentage of your knowledge workers and decision makers are using enterprise BI applications?” The choices that you have are, below 5%, 5% to 10%, 11% to 25%, 26% to 50%, 51% to 75%, over 75%, not sure, or not using Enterprise BI applications. For example they’re using spreadsheets and desktop tools instead. That’s a pretty interesting spread of responses. Mark and Bob, what do you think?

Robert Eve
This is Bob. As you might expect, a lot below 10% and a lot not sure. And there’s some clients that are not using Enterprise.  I think very aligned with your study.

Boris Evelson
Excellent. Yeah, I mean, excellent that I guess all great minds think alike you know. Not so excellent because we doing, it appears that we do have tons of problems in enterprises with pervasive BI. But I guess good news for all of us that hopefully by the end of this webinar, we will show our audience some of the tricks and tips on how to increase those numbers. So.

Mark Budzinski
Certainly. This is Mark. Certainly tremendous opportunity isn’t there Boris?

Boris Evelson
Absolutely. Absolutely. Again, as I said at the very beginning, anyone that tells me that this market is a mature and stable and boring, obviously is not correct. Tons and tons and tons of opportunities.

So now let’s spend the next few minutes talking with, we talked about all of the complexities and all of the challenges and opportunities so let’s talk about what is it that we can actually do about it today. Not in the future, but today, basically as soon as we hang up the phone from this webinar.

And the partial answer to that question is in the next generation technologies. The reason I’m saying it’s partial answer is because kind of a much more important answer is about best practices and implementing BI and information management applications. But it’s not a topic of today’s webinar. We do, at Forrester, we do a lot of research on these best practices so anyone who is interested in more detail, please reach out to me.

But today we are talking about next generation technology and what I want to talk about here is that kind of the what’s in business intelligence in terms of all of the complexity of that stack in terms of, you know, discovering and ingesting and measuring and analyzing and delivering the results probably isn’t going to change. But the house, the house are definitely changing and we do have lots of research at Forrester on all sorts of next generation technologies that we are aggregating by the following four categories. It’s the automation of all of the business intelligence processes, it’s making business intelligence pervasive, it is unifying multiple different siloed technologies and it’s also creating business intelligence environment that does not have any borders or any limitations.

All of these can actually be tied together by what we are defining as agile methodology and agile technologies. That concept is pervasive throughout all of these four categories. And by the way, underneath each of the categories is probably a half a dozen different technologies, different approaches. We won’t go into them today, but again, anyone who’s interested in more details, feel free to reach out to me.

So let’s go to the next slide and talk about why HR is so critical in business intelligence. So in business intelligence is very different from any other enterprise application. If we talk about, let’s say, CRM or ERP or financial applications or HR applications, we can go through a traditional SDLC development lifecycle and be pretty confident that for the next six months or twelve months or very often a couple of years, we can be pretty confident that the requirements and the applications that we deliver will support those requirements with some minor, maybe some major enhancements and changes. But the basics of what we design and put out there will probably hold for a while. Nothing unfortunately can be further from the truth in business intelligence. Even if we take, not even months, but even if we take weeks to deliver something, all research shows that by the time you deliver it, the world has gone and changed on you. The competitive landscape has changed, you now have a new competitor, somebody put out a new product on the market and all of a sudden you need to analyze something that you didn’t foresee a few weeks ago. People have moved on, those people that gave you the requirements are no longer there and the new stakeholders have a different set of requirements. I’ve got tons and tons of examples of why business intelligence, any business intelligence application is old, is too old even if it was designed even a few days ago. So kind of build it and they will come is absolutely is the truth.

The unfortunate challenge in business intelligence applications of being agile practicing agile approaches in business intelligence are infinitely more critical than in any other software development environment so let’s dive into more detail. So Forrester defines agile as having two components. Number one, it’s the development methodology but number two it’s the agile BI technology and architecture because unlike any other enterprise applications or software, you can’t really practice agile methodology unless you have the right architecture. Because there’s so many components and business intelligence that are loosely or tightly coupled that just saying that you’re doing agile product development isn’t going to help you if you don’t have the right supporting architecture.

In terms of the methodology we talk about for our areas, it is you have to react to change as opposed to planning. Remember the previous slide where I talked about, you know, you can’t really plan because plans change and new requirements spring up on a daily basis. So you have to react as opposed to planning. And this is very difficult by the way, you know, including somebody like myself who’s been a hands on IT practitioner, you know, things like planning and specification and traditional SDOC, that’s something that I lived and breathed for and I was trained in and indoctrinated in for many years. So this is almost like rubbing against the grain for professionals like me. But never the less this is the right approach for business intelligence.

So reactions versus planning, it’s interactions as opposed to a planned processes. It’s building rapid prototypes. It’s infinitely more important in BI than in any other application because as we discussed, by the time you finish with specifications, they may be outdated. And face to face involvement, not even involvement but face to face kind of business ownership and involvement, personal involvement of key business stakeholders as opposed to working through liaisons and middlemen are absolutely critical for an HR business intelligence.

But if we look at the next slide, unless you have some of the critical components that technical components, architectural components that support agile, you can practice the methodology all you want but if you don’t have the right underlying architecture, it’s not really going to get you very far. So things like today specifically where we are going to talk about the first two components necessary, critical components for agile BI, which is integrated metadata and integrated end to end business intelligence component management. There are plenty of others such as using some of the new databases that are built specifically for business intelligence, not for transaction processing. End user self-service is absolutely important, critical, very critical for agile and successful and pervasive business intelligence. We won’t touch on that today but we do have lots of research on that subject so I encourage you once again to reach out to me.

So let’s talk about what benefits do we draw if we indeed have integrated metadata and integrated end-to-end component management in business intelligence. So number one, the obvious advantage is that you can understand the impact analysis. The impact analysis of everything that kind of goes from the bottom of this endorsed pyramid to the top understanding how any change in any element affects any other one is very important. And you can only achieve that with integrated metadata. And kind of the reverse of that process is integrated metadata gives you data lineage because just a simple process of debugging in the old fashioned source code doesn’t work in business intelligence because you have, you know, there is no single piece of source code. You’ve got, as we talked about, 40 to 50 components. So how do you know that kind of improperly behaving metric on a dashboard is a rendering problem on the dashboard, is it calculation problem in the dashboard or a calculation problem in your KPI that there was populated using an ETL routine or whether there was something changed in a data cleansing process as we talked about 40 or 50 components where that particular bug may occur. So without integrating metadata and understanding what the data lineage is of any data element or attribute. So this is extremely critical.

So the other benefits of this new generation of metadata driven or integrated BI applications are very obvious. It can do as we talked about, it can do rapid prototypes, which is extremely important. It can generate documentation rather than kind of building documentation as a separate process. You can push a button and a document, your application’s very important benefit here is reverse engineering BI application. So we’ve seen some interesting cases out there where clients rather than, let’s say converting from one BI platform to another or rather than migrating to a new version, that they kind of take that opportunity to reverse engineer the BI applications into these central metadata driven BI environments so that not only to speed up the conversion of upgrades from this particular project but for all of the future projects.

The organizational and process benefits are obviously huge. So obviously this type of an approach gives you much better governance and change management process. You can build in some of the best practices into your metadata so that when your data marks or your queries are generated they already conform to some of the best practices in terms of tuning and optimization. And because some of these applications and tools that we’ll talk about later do indeed give you much faster time to market, much that they enable you to roll out your applications much faster, it does enable your IT organization to have this kind of a more relaxed and more even pleasant “Yes” culture. Because as we unfortunately know, we on the IT are often challenged with having to say, “No” too many times. But when we know that we have a tool, we have a platform that makes our job easier, then we can roll out a new BI application with less effort, we can definitely say, “Yes” to most, well not most but more often.

We do have a research paper that we will mention at the end of this webinar that does describe the different approaches in a certain level of detail because some of these metadata driven BI applications are not created equal. So as you evaluate them we will point you to some architectural differences in terms of what kind of platforms do they run on, do they support top down versus bottom up modeling and on the other hand in terms of in this whole end to end business intelligence stack, how many components can these metadata driven BI applications either eliminate or generate automatically. So can they automatically generate your ETL scripts and can they automatically generate from the same metadata report, not from two different ones, but from the same one. Can they generate your all app cubes and your metadata layers etcetera.

So I encourage you to look at that report and understand all of the differences between the technologies and platforms that you have here. And I think actually we do have time for another poll at this point.

Gary Damiano
This is a good point to survey the audience with another polling question. Let’s present that. Question is, “How often do your BI requirements change?” And the choices you have are, hourly, daily, weekly, monthly, semi-annual or annual.

Robert Eve
And this question really gets at the heart of why we need agility. I mean we start right from the start with all the business reasons and that sort of thing which is just like a really, that really clarifies the fact that, you know, activities happen, things are changing and therefore BI needs to respond.

Mark Budzinski
I think it’s pretty illustrative that in a project delivering BI solutions to the business community, you’re never done. I mean if you look at this data, this sort of weekly even monthly evolution of what the business thinks they need just reeks of an agile approach. I don’t know how else you would successfully go at it with that kind of dynamic in the requirement sense.

Boris Evelson
I would say it definitely supports the challenge that I talked about earlier with 40 or 50 components that potentially can be affected in any typical average BI application even when the requirements do change on a weekly basis as close to 40% of the respondents indicated. That is very, very challenging.

Gary Damiano
This begs the question of flexibility and responsiveness. Let’s poll our audience once more with a question regarding that. How flexible and responsive is your BI environment? Your request is satisfied in less than one week. Your request is satisfied in weeks, months, the backlog is endless.

Robert Eve
It looks like months is the winner.

Boris Evelson
Okay wow, so I’m not seeing this report, I’m not seeing the results yet but if we do have a significant portion of the respondents responding in months, right. So over 40% so that and then 16% of the people are saying that the backlog is endless. And you know what, I would venture to say based on my experience, we probably have mostly IT people on the phone today. I would say that if we had a significant percentage of participants from the business side, they probably would think that even a gloomier picture. But I think the results, even here the results speak for themselves that the frequency of the changing business requirements for BI and the speed with which they can be satisfied absolutely creates that gap and that opportunity that I talked about at the beginning of the presentation. So, you know, I hate to say I’m glad that it is confirmed because obviously there’s lots of people on the phone that live and breathe these challenges but at least today we do have technologies and approaches to address this.

So let’s move on. And so I will start wrapping up by introducing two [inaudible 0:31:55] these are definitely not the only ones. As I talked about earlier there are at least a couple of dozen of next generation agile BI technologies and architectures that we are forced to really think are necessary. But these two are the ones that you see today.

One is the data consolidation approach. And it’s not normally the consolidation; it’s automating of multiple stacks in this complex 40 plus component BI stack. And the other approach is a federation approach and it’s not only federation approach but it- the way it works it actually can help you eliminate some of the complexity of this end to end BI stack by eliminating some of the components. And what we actually see in practice, most of the more advanced clients in very large enterprises that have extremely, not tough challenges, actually implementing both of these approaches in a complimentary manner. You know, there is definitely no single approach that solves everything out there. So you have to mix and match.

So let me wrap up with my last slide by saying that agile BI is here today. You absolutely start looking at this so the very first step is identify opportunities. And you’ve got plenty of applications; front office versus back office, local versus global applications, mission critical versus kind of a nice to have application. So kind of a loop through the mix and understand which ones have the best fit for agile and then conduct a proof of concepts. Obviously we never recommend jumping full speed ahead into the use.

And last but not least, work with some experiences systems integrators. Experience and agile, they are the last thing that you would want to do is pick a systems integrator that does not believe in these next generation technologies and practices the old fashioned approach that takes months and years and they just want to bill hundreds of billable hours to you. So look for the right partners. With that, let me turn it over to my friends from WhereScape and Composite.

Robert Eve
Thanks Boris. Hi Mark. Can you tell us about WhereScape’s consolidation approach to agile BI?

Mark Budzinski
Yeah. I’d be thrilled to have that opportunity. And thanks Boris, that was just a fabulous set up here in terms of the agile approach. And I think just to summarize you mentioned our two approaches. One of which is a consolidation of data and the other is more of a federated view. And I think my comment would be both are completely valid. They’re both very active approaches that you should take on and see which one is a better fit depending on your specific environment.

WhereScape is about one approach and really only one approach because we think it’s a very important one that a lot of customers pursue. And that is the building and managing of a proper data warehouse. And if you’ve done this before or been involved with a team that’s done this before, you know there’s a lot of scary monsters that are sort of in the closet that emerge sort of somewhere in the middle of the project. These probably can vary the global budget. Gartner says 50% of them actually fail or go over budget. They’re very difficult to manage. You end up spending more time, more money, more resources. Then at the end of the day by the time you get back to the business and say, “You know, ta dah we’re done.” Based on the polling results we got a few minutes ago, many times the business requirements have moved on and you know you’re not done. You’ve got to go back to the drawing board and sort of keep at it.

So WhereScape brings a very fresh, certainly agile approach to this problem by approaching it from a much more dynamic, flexible and largely automated way. And the way to think of it is if you take the traditional ETL world, the classic extract, transform and load sort of part of the solution and then you combine it with what’s often done in the data warehouse team itself. So design the schema doing the indexes, creating your surrogate keys, you know, doing all that stuff including the skilled procedures that actually load the tables, creating the aggregates, creating keys and all that. If you combine that world which is very labor intensive with the classic ETL world, glue those together and then automate that process. That’s essentially what WhereScape has done.

All right so before I tell you more about WhereScape, let me just quickly look at this what I call traditional approach. Because I do think there’s a little bit of looking in the mirror that we can all learn from to a certain extent and enjoy. You know, if you think about traditional processes to building and managing a data warehouse, we start this notion of go get the business requirements from the business. And there’s a lot of specificity there, right? Let’s get them not just in generic terms but specifically, what business questions are we trying to answer. And from that knowledge we go off and we design models. From those models we analyze our search data, we create our ETLs facts, we build the code, we load the data warehouse, we build our sort of reporting design front end and we actually get on with implementation and again this process tends to be flawed in the sense that by the time it’s done, you know, it’s not exactly what the businesses wanted.

So we believe that it’s not a matter of, you know, well meaning people didn’t try hard enough or something along those lines. It’s literally a matter of approach in terms of why so many of these projects fail.

Now WhereScape takes an approach that again is agile, it’s flexible. We want to sit down with a business user and get the jest of what they need as opposed to every last detail and we want to go connect to the source data and start building. Now we can do that because the tool set and the development environment that we have is completely conducive to that kind of flexibility. It’s very, very fast because so much of it is automated. Everything that you see in this red box is essentially the nuggets of our data warehouse development environment. So we connect the source data and that can be live source data or it can be an existing data warehouse if you already have archive data. But invariably you’re going to want to integrate that with maybe stuff that’s in somebody’s laptop, perhaps a CFOs special spreadsheets or what have you. And that is all integrated within the environment. So this is where the ETL part comes in but look at all the other stuff that actually happens in an automated way. Database modeling is actually a consequence of getting on with the environment. So rather than getting models right and you know quote unquote perfect before you go off and build your data warehouse, let’s build the source to target next based on real data and real results as we have this very fast efficient way of building.

You see over on the right side, the BI front end, this is the presentation layer that is essentially your choice to integrate our sort of plumbing solution with. So if you like MicroStrategy or Hyperion or Yellowfin or whatever your favorite front end is, you know, we essentially gruel our environment to that. And when I say gruel, I mean we present the relational structures or we present the cubes depending on how you want to see it.

The databases that we support here within this red environment are pretty pervasive. So if you’re building a Sequel Server, you’re building in Teradata, you’re building an Oracle for example, we literally build, automated, the store procedures that you would otherwise write by hand in that native sequel; so the P sequel, the PL sequel, the Teradata sequel. And as you can see up in the upper left there, we actually create a metadata repository within the local environment. So within Teradata, within Sequel Server, what have you, that actually is very profound because with pushes of buttons you can literally document everything that you’ve done for users or for technical staff that’s track forward, where did this data come from. You know, if somebody streams the datas wrong, “No, I can show you authoritatively that this is the source that it came from as we built our load tables and built our dimensions and built our facts” and so on.

So if you’re building or want to build in a dimensional model in a classic sort of Ross Kimball sort of environment, all of that knowledge is embedded into the environment and that’s what makes it so fast. So rather than having to worry about writing the code and how do I deal with a solo changing dimension and how do I deal with error handling and what we affectionately call the donkey work at WhereScape. That is just off the table, it’s automated and you can deal with the tough issues of actually getting on with it.

So just to align it, I know some of you on the call work for larger companies and we have over 350 customers in the world, many of which are larger companies that would sort of mirror your environment so the Wells Fargo’s and Sirius Satellite Radios of the world. And the usage models there are typically the data warehouses in place, an ETL process is in place, but we’re very much used productively on the access layers, semantic layers side of the problem where by we can validate real data, we can build metrically, we can work in a sandbox environment. So before you go off and waste a lot of time putting things back through the classic production ETL sausage grinder, you have a chance to validate what you’re doing and make sure that you’re productive. If you’re a smaller company or a midsize company that’s moving into data warehouse for the first time, you can see on our chart here that we’ve got a number of customers that would sort of mirror that as well.

To close, we’ve got a couple of customers that I think maybe speak to the value proposition better than I can. In the case of United Rentals, this is a company that you may have seen along a street corner near you. They have these stores where they rent big cranes and diggers and dump trucks and stuff for construction projects. They have essentially built their entire data warehouse environment in Teradata using WhereScape. In the original budget that Dan Mosher had was two days to build a table. He multiplied that out times the number of tables that he expected to build for the environment. Turns out it’s only two hours per table. So literally a factor of, you know, like 8 to 1 improvement, he was able to save 80% of the budgeted plan time and ultimately money associated with building up this environment. And that is essentially what WhereScape brings to the table and why we believe in the agile environment. It’s a good play.

IPC Subway, a very similar story in the sense that they’ve automated their whole environment to build this thing out. So with that, let me close and pass the baton back to, back to the floor. Thank you.

Gary Damiano
Thanks Mark. Hi Bob, there’s been a lot of industry discussion about the growing use of federation or data virtualization. Can you tell us about Composite Software’s data virtualization approach to implementing agile BI?

Robert Eve
Yeah, great, thank you. I figure it’s very complimentary that Mark talked about this because a lot of times you not only are requiring to do consolidation, you’re going to talk about a more agile BI approach using data virtualization or data federation which is either an extension to your existing data warehouse strategy or there’s an alternative.

Basically the core of how it works is around a very simple idea and that is when you build a view for federated data access. So the first thing you need to do is discover the data, what’s the data going to virtualize and display and deliver to the business intelligent solution. We have some automated tools that help you understand the data and the relationships across multiple sources. Then you kind of do your classic model laying and build your view and it’s, all you guys know it from the database environment. Building views is quite a simple activity. And then step three is you get some online validation because you see the actual data and you’re ready to go. So it’s really just that quick.

The benefits really here are the timed solution. You’re not really having to build the ETL scripts and actually set up the scheme and all of those activities. You’re really jumping right into the chronicle schematic layer and I think you also could iterate much more quickly which I think speaks to the fact that the amount of change that was already talked about in the poll and just the dynamics of the business.  And building views, everyone knows how to do that so it’s kind of a nice benefit as well.

The next development in production, what tends to happen is you run on a more of an ad hock query approach where the report actually calls the data, composites them, pulls the data together in real time using an optimized data access and retrieval approach and then deliver that to the report as required. The benefits of its approach, of course it’s up to the minute data and you’re getting a high performance query of some real time information. But from an agile point of view, this much replication with fewer consolidated sources, etcetera, it’s much easier to change as things go along. So that’s the big idea.

Let me tell a customer example because sometimes the customer can tell the stories better than the marketing folks. This is a customer that happens to be Pfizer and in their R&D organization they have a ten year development cycle from a concept of a new drug to the- by the time it actually gets out to market and gets through all the approvals, etcetera. As you can imagine, in that sort of environment lots of things can change. As Boris said, “The world keeps changing.” And they have clinical trials that gets feedback and that might have to- or other research that they might have to respond to. They have regulatory changes, etcetera. So you can imagine the kinds of things that can happen over a ten year process and you can also imagine the breadth of data required to really analyze and stay ahead of all these changes and react to these changes.

In using our products they’ve been able to use that development time for new reports opting with same day service, really a good agility to take on one-time activities and make a decision and then maybe never have to get that information again. And then it had some huge business benefits by taking cost out just in running the operation by being able to make the decisions more quickly. But the key aspect is being able, you know, the customer says that bottom, it’s really stay on track and take time out of that process because the earlier they get the drug to market, the faster they get return on their investments and bring in revenue.

If you’re going to use this approach, this virtualized federated approach to gain the BI agility, there’s really no one better to work with than our company here at Composite. And I say that not because I say it but because our customers say it. We have a large customer installed base, very focused in some of the major industries such as financial services, pharmacy, pharmaceuticals, energy, you can see they’re kind of the top large companies and the enterprises that we work with. It’s very proven technology over the years so I think you can learn it quickly, get up and get benefits fast and really achieve some of these agile benefits that everyone’s seeking.

Gary Damiano
Thanks Bob. We’re running a little bit long so lets- this really begs the point of throwing out another poll question which I’d like to do but also let’s do the poll question, let’s talk a little bit about how we’re going to address all the questions that have been coming in over the chat line. And then I want to wrap up with talking a little bit about the white papers that we’re making available to everybody.

So why don’t we go to our polling question. “What approaches to implementing agile I would be useful in your organization?” Data consolidation approach, that Mark was talking about with WhereScape USA where the data virtualization, Composites approach, the combination of both or neither, do you need more information before deciding or my favorite option, I’m still confused and need help, or none of the above.

Robert Eve
I’m watching the polls come in and it seems like to be a combination of both is proving to be the top input as well as, “I need more information.” So Boris, how would you interpret those two as the top?

Boris Evelson
Well the later one is obvious because this is a new topic. Even though, as we said, these technologies have been around for a while but to most of our clients these are new technologies. So I think between Forrester Research and both WhereScape and Composite, I’m sure we’ll all be happy to provide more information. I know Gary will let everyone know how to get a copy of my recent kind of a deep study into this, the white paper that describes lots of details. So hopefully that will be helpful.

And in terms of a combination of both, that’s precisely what I’ve been seeing in some of the leading companies that really came to realize that no single approach works for everything. There’s a huge difference between, let’s say, a back office application such as, let’s say, regulatory reporting that does require a lot of governance and a lot of controls versus some kind of a front office application which really needs to be much more agile. And if in my sales analytics application I bypass some of the controls for the sake of the speed of delivery and for the sake of agility or being able to analyze, let’s say, competitive threat, you know that we have to pick our battles and we have to really understand where control and risk management is more important than a quick answer to a key stakeholder.

So I think these leading enterprises indeed are realizing that in some cases you’ve got to use one approach, in other cases another.

Gary Damiano
Thank you. We’ve had so many questions that have come in and we’re at the top of the hour and really can’t address them. So what I’d like to propose is that we will take the time to go through those questions and respond to them. And what we’ll do is we’ll send out an email to all the attendees with our responses to those questions because we do want to- you took the time to put them in and they’re really, really good questions and we want to address them. But we’ll send out a follow up email with the responses to your questions.

I’d like to finish up and talk a little bit about some white papers that we have available. As I mentioned earlier in the web cast there’s some really informative and useful resources that we’d like to share with you. If you’re a Forrester client, you can access Forrester’s white paper, Agile BI Out of the Box, at Forrester’s website. If you’re not a Forrester client, and you can send Boris an email and ask if he can share that with you. That’s I’m going to put Boris a little bit on the spot here but, because I know not everybody on the web cast is a Forrester client, but I’m sure that Boris would love to hear from you. The second one is WhereScape USA is offering a white paper covering best practices for BI that can be accessed on their website.

Robert Eve
That was authored by Claudia Emhoff, right?

Gary Damiano
Yeah by Claudia Emhoff. Composite Software recently published a white paper entitled Data Extraction, Best Practices, that discusses agile BI and can be found on our website as well.

That’s all the time that we have here today. Thank you for participating in the web cast. If you have any additional question or you’d like to speak to any of our speakers, please feel free to contact them. We have more data virtualization leadership web casts coming your way. Please go to www.compositesw.com to register for future events. You can also follow us on Twitter. Our handle is Composites. Does WhereScape have a Twitter handle? It’s just WhereScape for following WhereScape on Twitter. This concludes today’s web cast, Agile I Made Easy, Two Proven Paths to Success. I’d like to extend a thank you to our presenters Boris, Mark and Bob for spending the time with us today. And a special thanks to you, our audience, for attending our web cast. Thank you for your participation.