Ted Kummert: Business Intelligence Conference Keynote

Remarks by Ted Kummert, Senior Vice President, Business Platform Division
New Orleans, La.
June 8, 2010

ANNOUNCER: Ladies and gentlemen, please welcome from the Microsoft Corporation, Senior Vice President, Ted Kummert. (Applause.)

TED KUMMERT: Well, thank you. Well, good morning and welcome to Microsoft’s Business Intelligence Conference. It’s an honor and a pleasure to be here with you today.

I run the Business Platform Division at Microsoft. The Business Platform Division is responsible for developing Microsoft’s application platform. That includes SQL Server and our application server products and technologies. Within SQL Server, we’re responsible for developing all work loads. That’s from OLTP to data warehousing to business intelligence, including reporting and analytics.

I’m also a user of business intelligence. Every day, we’re asking ourselves questions about our business. What’s going on with this product in this channel? What’s going on in this region? How are we tracking to our release? What do these defect rates mean to where we are on the product life cycle?

We’re able to get the answers to these questions because of the business intelligence solutions that I have access to. And there’s something absolutely magical that happens when business intelligence works. When an end user with a question gains the insight they need to move forward, business moves forward and people are more effective in their jobs. That’s the thing that you are all a part of making happen every day. It’s a powerful thing to make others look good, to make people more effective at their jobs, to enable business to move forward in a new way. That’s the magical thing that business intelligence can enable.

So if you were at our last Business Intelligence Conference in Seattle, you saw something a little bit nontraditional from us. I told a fairy tale. A fairy tale of business intelligence. Now, fairy tales are great stories, good wins out over evil, people find true love, dragons are slain. And they’re wonderful stories because they’re stories we all want to believe.

Now, the fairy tale of business intelligence told the story of a future where every end user was able to use BI technologies within their job in order to move forward more effectively. So we’re going to check in today on the fairy tale. We’re just fresh off the release of SQL Server 2008 R2, Office 2010 and SharePoint 2010. In that, we delivered a new scenario we called managed self-service business intelligence, which is going to enable an order of magnitude more end users to be able to produce, consume and collaborate on BI solutions in an infrastructure managed by IT.

We’re going to take a look at that. We’re going to take a look at some customer examples, and then we’re going to shift our discussion forward. We’re not announcing our future release roadmap, we’re not committing to specific features what and when, but we do want to give you a glimpse in terms of some of the things that are on our mind and where we’re headed next.

So I imagine many of you have heard the pitch before about BI for the masses and BI for everyone. It’s been an industry conversation for a long time. I consider this slide to be somewhat of a report card in how far we’ve come. The various industry analysts disagree in the fine grain, but most of them would agree that it’s around 20 percent of the potential users of business intelligence that are able to use business intelligence today.

So what’s the problem with that? Well, frankly, the fact is there’s a lot of problems that end users can and should be able to solve for themselves. That simple modification to the report. Bringing an additional data source into the solution. A change in formatting. An additional data visualization. Many, many things that end users can and should be able to solve on their own.

And what do they do instead? Well, they come and ask you. But you’ve got other things to do. You’ve got strategic applications you need to continue to work on. So we’re very, very committed to continue to move forward the technologies and tools that we’re giving to you, the BI professionals, but we see a tremendous opportunity to enable the 80 percent, the rest, with the tools for them. Life gets better for everybody. You’ll see more impact from your solution. People will be able to build on them on your own. You’ll have less interruption. You’ll be able to spend your time on the strategic applications you’re building.

And our vision is to enable BI for everyone, to take that step to enable BI for the 100 percent. And there are a few things we think about in terms of how that’s going to be enabled. The first relates to the user experience. You need to learn too much BI technology and terminology just to do simple tasks today. I think of this as delivering the right tools for the right user in the right place. If I’m in Excel and I’ve got a task I need to get done, make it a simple extension of that Excel environment. Make it something that’s very natural and intuitive for me to do. I’ve got a task to accomplish. I don’t need to learn about star schemas and other technology in order to get that thing done. I’m a professional user — I want to live in Visual Studio. Or if I’m an end user that lives in a line-of-business application for most of my time, I want to see those visualizations and reports surface seamlessly embedded within that experience.

The second aspect is seamlessly enabling collaboration. You know, structured data, business intelligence applications, these solutions are just like any other business document. You want to be able to do everything you can do with any other document in your environment. You want to search, you want to find. You want to be able to share and collaborate. You want to be able to build on the work of others. You want to initiate work flows around these documents, to root them around. You want to apply rights management to them. Everything I can do with a document today, I want to do with structured data in my business intelligence applications as well.

And the last part is enabling IT to add value. IT adds value by managing the infrastructure. The infrastructure is always there for every end user to tap into. By managing the applications, by providing the data stewardship and the data governance, everybody’s building on data that they trust from trusted sources, quality data. BI for everyone comes together by enabling the right tool for the right user in the right place, enabling seamless collaboration in an infrastructure that’s managed by IT.

So we’re fresh off the release of a set of products we’re pretty proud of: Office 2010, SharePoint 2010 and SQL Server 2008 R2. We put a big emphasis on this self-service scenario, managed self-service business intelligence. And there’s two parts to this. There’s an analytic scenario and a reporting scenario.

So first, I want to take you through what motivated the managed self-service analytics scenario. So I’ll hearken back to the fairy tale. The fairy tale went something like this: I’m in HR. I’ve been asked to do an analysis to compare our salary data with industry trends. It’s to take data I have with data that exists in the wild. And so what do I need to do then? What do I do next? Well, I go ask IT for help. But IT’s busy, they can’t help me. But I’ve got to get it done. So what do I do? Well, I put the solution together myself in Excel. And it works out pretty good, but some problems occur down the line. Maybe I didn’t build it on the right data source. You know, maybe I left the company and others are trying to figure out how to do it, and every time you need to build the solution, it’s a set of manual efforts.

Managed self-service analytics was designed against this type of scenario. The first part of enabling this was to deliver an experience within Excel to build these type of applications. We call it PowerPivot for Excel. Every user of Office 2010 is going to be able to download and use this PowerPivot experience. What does it enable you to do? Well, in an Excel-like look and feel, it enables you to bring in data from a variety of data sources very easily, define relationships, we help you infer relationships between data and construct these type of BI applications.

It’s built on top of a very powerful piece of storage technology, a column-oriented, in-memory store, which enables very, very high volumes of data to be stored in memory on a typical desktop platform that has incredible performance, which enables a high level of interactivity, bringing high volumes of data — now you’re really able to interact with it.

Now I built this solution, now what do I want to do? Well, I want to publish it. I want to share it with others. So where do I publish that? We bring the environment into SharePoint to enable you to seamlessly publish that into the SharePoint environment. We also enable it to render symmetrically within the SharePoint environment. Using Excel Services and PowerPivot for SharePoint, you can now just go and view and interact with that application in your Web browser. No more mailing around the work group and having to keep track of where they all went. Those other users that just want to use that information don’t want to download anything can just go view it on the SharePoint site and interact with it just like on the desktop. Now, if I want to build on it, I can download the workbook, I can make modifications to it, I can upload that back to SharePoint, others will be able to see it.

We bring it into SharePoint as a point of collaboration that’s also a point for IT to manage. IT manages the infrastructure. IT now has visibility of the applications. IT can apply policy; it’s a point of management as well. So we’ve got a familiar user interface within Excel for the Excel user. We bring it into SharePoint to enable collaboration, IT manages the infrastructure, adds visibility to the solution within the SharePoint environment.

The second scenario is managed self-service reporting. And we’ve been on the end user reporting journey for a while with our Report Builder tool. We’ve got a great release of that in SQL Server 2008 R2, the third release of Report Builder.

So here’s a motivating scenario for what we built there: I’m the same user in HR, I’ve been asked to build a report, and that report includes data from my department and data from another department. Now what would I then typically have to go do? Well, I’d have to go figure out where that data comes from. I’d have to go assemble the query. I’d have to go build the visualization. And I haven’t spent any time yet on actually what I want to focus on, which is the visualization around my data.

So we’ve enabled a scenario we call “grab and go” reporting. It’s built on a feature we call the component libraries. And the component libraries are now in place to save off these visualizations, these report parts, if you will. So now I can assemble these reports from these things. And those can be things you publish from sanctioned data sources, or those are things that other end users have published, or those are things I’ve just built before and want to reuse. A very, very powerful thing for enabling end users to be very productive in building very rich reports very quickly.

And reporting services is integrated with SharePoint. Again, reports are just like any other business document. You want to deal with them within that same environment, just like you do any other business document. So we’ll take a look at these scenarios in just a minute.

I also want to talk about a question we often get. The question we get is: “Hey, Microsoft, what’s your BI offer? How do I get it?” Well, the answer is up there on the screen. We really see in terms of this vision of BI for everyone, this is utility capability. Utility capability in terms of being like electricity, it’s something that’s there that every end user can tap into. That’s why we brought it to these products technically, how we integrate it, how we built the scenarios. And that’s also how you get it from us — it’s just there. It’s in Office, touching over 500 million end users. It’s in SharePoint, adopted by more than 70 percent of the enterprise organization. And it’s built on SQL Server, the most popular relational database management system in the world and the number one OLAP platform in the world. It’s just there. Utility capability for every end user to be able to tap into and use.

So now I’d like to let the code speak for a bit. And so we’re going to show you how this managed self-service BI scenario comes together. So I’d like you to join me in welcoming Michael Tejedor to the stage. Michael? (Applause.)


TED KUMMERT: What do you got?

MICHAEL TEJEDOR: Well, we’re going to step through some of those scenarios you were talking about. So the familiar experience for end users, the collaborative experience for teams, and then the managed experience for IT professionals to help manage that all.

So here what we’ve got is SharePoint up and running, and a lot of us use SharePoint at work. And one of the things we did with 2010 was we actually extended search. So if I search for sales trends, we see that the system can actually now search through the contents inside of SharePoint, and we’ve extended it to actually search through the data inside of reports, workbooks and PowerPivot.

So here we get views on what reports are in our system, and we get some data about what’s in that report as well. And this is actually something I’ve started using at work quite a lot. So we’ve got a large SharePoint deployment at Microsoft. This is a great way for me to find the reports and the data that I need.

Usually the way I refine my search is I go over here to the left. And on the left, you can see I can start to refine my search based on content type, data source. So if there’s a source that I’m looking for reports from, I can filter out the list and see all the reports from that data source.

Now, another familiar consumer experience that we’ve added is this Silverlight experience on top of report libraries. So here you can see it’s very quick to kind of scroll through and see what reports are available in my system. And I get all the great previews to be able to do that. And if I want to actually open up one of these reports now, it’s that familiar Excel environment but running in the browser. And that’s great because now I can interact with this report and I can view it and I can see the data the way I want to see it, but it’s actually maintaining one version of this file. So your end users can interact and play with the data but still maintaining one version of the file.

TED KUMMERT: This is running on SQL Server 2008 R2 in the back end.

MICHAEL TEJEDOR: That’s right. Actually, there’s quite a lot in this file itself. So if we opened it up and take a look at what’s happening sort of behind the scenes, now we can see that we are in the full version of Excel. And in the full version of Excel, we’ve installed PowerPivot. And PowerPivot is one of the new components that we’ve released in R2, and that extends Excel with this in-memory analytical engine, which allows creators of content to actually pull information from multiple different sources.

And here we’ll just go to PowerPivot, you can see what’s behind the scenes. And I can now pull information from heterogeneous sources in memory, join that data together, and then build my workbooks on top of that joined data.

So here we see I’ve got my data set all in. It’s not a huge data set, 4 million rows, not like the 100 million rows that Amir showed yesterday. But to get that data into Excel, I can pull it from a number of different sources. So here we’ve got a list of the different sources we can pull from. So here you can see we can pull information from hosts of different relational sources, from analysis services, through data feeds, through text files.

And what I actually use at work is that I pull information in from reports. So it’s very easy for me now to find the reports that I want. And through this feed, I can actually pull the information from the original source that the report is built off of. And that’s important if I’m refreshing the data that I actually pull from the original source. So that’s a great way to get sort of your data feed in here.

TED KUMMERT: That can be a great way to get data out of the line of business system as well.

MICHAEL TEJEDOR: Absolutely. Yeah, you can build a report off of any number of different sources, so it’s a great way to access the data from an IW perspective.

Now, another interesting extension here is the ability to actually build out DAX formulas. So PowerPivot is quite powerful in that it’s got this formula language where I can go ahead, and I’ve got one here, take a look at it, I can build out these expressions in DAX. And DAX, if you’re looking at the expressions, looks and feels exactly like the syntax for Excel. So if I’ve used Excel formulas before, I can quickly create these DAX expressions, start playing with the PowerPivot product.

TED KUMMERT: Yeah, that’s important to mention. We didn’t invent a whole new expression language. What we did was extend the Excel expression language for these types of operations.

MICHAEL TEJEDOR: Exactly. And they’re quite powerful. This one in particular actually is running something that if you’re trying to sort of parse out what this does, it’s running a correlated sub-query. So it’s going to a related table, it’s bringing back the summary of the sales amounts for each of these categories, so quite a powerful little calculation.

Now, as an end user, once I’ve created my data set the way I want, I can go into Excel, and I can create the workbook the way I want it to look, and then I can actually publish it back into SharePoint. And when we’re back inside SharePoint, now we can talk about another experience, which is the collaborative experience around these things.

So if we go and take a look at a team site that I’ve got created here — so this is a site inside of SharePoint that we’re using to share content and to collaborate as a team, and we’ve got that report that we were looking at right here. Now, a couple things around this for team members is they can actually apply some of the new social and collaborative capabilities to this.

So a couple of those are tagging. So tagging is a very social concept. I as a team member can go in, I can see this report, and I can actually tag it and say I like this report. And then what it will do is in my own personal My Site in the system, it’ll show my network that this is a report that I find of interest.

TED KUMMERT: And these are all just features of SharePoint 2010. We just brought BI into that environment.

MICHAEL TEJEDOR: That’s actually — yeah, that’s an excellent point. Any of the BI content we create is treated just like a file inside of SharePoint, so you can apply all the functionality to it inside SharePoint.

Here we’ve got the ability to actually add commentary to it as well. So we’ve got the noteboard. If I want, I can add a comment like — “this is a super, duper report.” And my team members when they come back into this, they can see the commentary that’s been left for the report, and we can actually start to see, you know, the comments that the team are leaving for this particular report.

Now, one of the last social features I’m going to show you is this ranking capability. So I can actually rank and let other people know if I find this report super useful or not useful at all. And that gives other people on my team a heads up as to whether or not this is going to be useful for them.

Now, we’ve got these additional capabilities here on the dropdown list, so I’ll just zoom in for that. What you can see here is all the host of other capabilities that you can apply to this file. So as you were mentioning, this is just another file inside of SharePoint, but it’s a little different. It’s also a file with a lot of content in it, a lot of data. So what I can do is I can set the refresh policy for that data right from here. I can check out the report, I can check the permissions for the report and manage the permissions for the particular report.

And one really interesting feature is the work flow feature. The work flow, the way that works is that I set up a work flow with a number of different checks and milestones. And then I can apply the file to that work flow. And the system will track and route the document appropriately through those different milestones.

Now, that’s interesting from an IW perspective, like an information perspective because it’s quite easy for me to send the report out for approval or for feedback from peers. And from an IT perspective, it’s quite useful because if you’re in an industry where you’re a little sort of constrained when it comes to regulatory compliance and that type of thing.

Then what you can do is you can set up the system so that it’ll track the workflow and any document that’s posted to SharePoint goes to you for approval before it’s made more widely available. So if you want to check the workbook before other people can see what data is in it, you can set that up as well.

Now, if we go into the managed experiences you were talking about, let’s take a look at the new dashboards that we’ve provided inside SharePoint for better IT pro management. If, one, you want to see how your system is performing, we provide a number of different metrics to provide you with that information so you can see the query response times and that type of thing. And also, if you’re interested, to just see what people are doing in the system. We allow you to actually see the workbooks in the system, who’s creating them, who’s accessing them, what queries they’re running, all that sort of information.

So we provide you with an interesting little chart here. It also tracks lineage of time. So we can run this report, and each one of these bubbles is an individual report in this system, and we can see the popularity of the report over time. And some of the ones over here are actually becoming quite popular. They’re running a lot of queries. So now you might make the decision, “Well, maybe I want to take that and I want to add it to a more dedicated system somewhere so it has the resources it needs.”

Now, one last thing I’m going to actually want to show you here is what if you’re interested in the sources of data being pulled from all these different reports? What if there’s some copy of the data somewhere in your system that people are building reports off of and you’re a little worried of where that source is? So here we’ve got a report that shows you all the reports, all the queries that are running, and what data sources those reports are actually coming from. So you have the information to see those rogue sort of data copies out inside your organization.

So with that, those are sort of some of the experiences we’re offering now around familiar, collaborative and managed experiences with the new product.

TED KUMMERT: Great, thanks, Michael.


TED KUMMERT: Thanks. (Applause.)

So one of the things we showed there is a component we’ve just recently released, which is the BI indexing connector. And this basically brings BI content into the SharePoint search experience. It basically enables you to index data in reports, Excel models, PowerPivot models, and it brings that seamlessly into the SharePoint search experience. It’s just another good example of how we’re bringing BI into those other experiences. We’re not creating standalone, we’re bringing BI into these environments, and we’re able to leverage the power of that environment and to provide that in a familiar way to end users.

So now I’d like to show you some of what a customer has done. And we’ve been working with CareGroup Healthcare for many years. They’ve been a long-time SQL Server customer. They have not been using our BI or much business intelligence in their environment. So I thought it would be great to talk to them about some of what they’ve done. With 2008 R2, they found some things they could do with managed self-service business intelligence.

So I’d like you to join me in welcoming Ayad Shammout. He’s the lead DBA from CareGroup Healthcare to the stage. Ayad? (Applause.)


TED KUMMERT: Well, thanks for joining us today.


TED KUMMERT: Can you tell us a little bit first about CareGroup Healthcare?

AYAD SHAMMOUT: Yeah, CareGroup are four hospitals based in the Boston area. We are a teaching hospital of Harvard Medical School. We are a long-time SQL Server customer. We maintain about 3.5 million patient medical records. And we have ICU and OR run on SQL cluster servers.

TED KUMMERT: Great. Well, can you show us some of the applications you’ve built with SQL Server 2008 R2?

AYAD SHAMMOUT: Yeah. All right. Actually, in the past, the key challenge we have is the gap between information worker and IT pro. Now, both of us, we have to work together in developing reports and staging data. So it’s time-consuming for both teams.

With PowerPivot, it redraws the line between the IT pro and the information worker so it empowers the information worker to produce their report, analysis report, some of them are really complex reports, without IT dependence, and also the IT staff can monitor the back-end servers, monitor the resources, compliance without user obstruction.

So today, actually I’m going to share with you a couple of scenarios. Again, as a healthcare provider, regulation is a big requirement, like HIPAA, Privacy Act, high-tech. So we have to audit and monitor all access to all patient medical records.

So a couple of years ago, we developed, working with the SQL CAD team, we developed a centralized audit solution based on 2008 Audit Feature. And having that massive amount of data, we’d like to do some analysis and understand the audit events happening in our environment, where are the people coming from, what kind of classes or actions they are doing on the backend databases.

So this report actually will show you — it looks like a regular pivot table. The only thing with a part of it allows us to add slicers, so we can filter through and look at specific class type and go through specific class type and look at the events by department, number of events, and also we can do charting, which is good. But the cool thing about this, the data behind these reports, we’re talking millions of records.

So let me just show you the workbook behind this report. So this is one of the tables we are querying. And as you can see here, we’re working with 44 million records, and trust me, this is a true number.

So let me flip back, and what’s cool about this technology is actually it’s so quick, so fast. And once we click, and this is basically — it’s querying that 44 million back-end table.

The other thing also we can do from the dashboard. So this dashboard is a presentation of the data, viewing data different charting. So we can also look at years, we can go back to different years, we can look at the whole event period, and you see all these charts updated at the same time. So it gives us the ability to view the data and analyze the data from different angles.

TED KUMMERT: How long did it take you to build this application?

AYAD SHAMMOUT: This report took us about six hours. And the cool thing about it — the other great feature is once it’s designed and published to SharePoint, it’s actually we schedule a refresh, and that takes a few minutes. So we don’t have to head back and update the data and spend more time on it.

The other thing actually that we have, like, inside this data, we have map information. So we can view data by state. So do you like to see the data by, you know, presented as a map and, you know, which basically show you, you know, a different view of the data. So here’s the — this took a few minutes to build this map using the PowerPivot workbook as a data source. So we can now show the data and present the data in a different way and quickly find out where are the users coming from. So we can show the states like, you know, Washington and Massachusetts, kind of the big two states, and then we can analyze where the people coming accessing the patient’s medical records.

TED KUMMERT: This uses the mapping control that we just released as part of SQL Server 2008 R2?

AYAD SHAMMOUT: Yes. This is actually designed by Report Builder 3.0. So that was the one scenario.

The other scenario is the analysis report. This is actually a readmission analysis report. Understanding this report allows us to reduce the number of cases of preventable patient readmission. So we look at, by department, at number of discharges and how many patients come back within 30 days. So that will help us to save money, time, and also improve our healthcare. So by viewing some departments like medicine, it will show us, like, number of discharges, we’re not doing that great there, so we have to find out why some of the elderly patients, they rush to go home quickly.

So part of this report helps our nurses to follow up with some phone calls so they understand the medication. What’s cool about this, actually, our information worker used to spend usually two days every month just to update the data behind this report. We have a heterogeneous system. The back-end data coming — this report is querying different data sources. We have Oracle, we have SQL, some of the data is local in an Access table. With the PowerPivot for Excel, it allows our information workers to import the data from these different sources and have them on the work station so they can join them together and build this report.

So now this report has been developed in six hours — four hours — and that saves two days every single month from our information worker so they can work on developing additional reports.

The other thing also is the storage saving as well. The backend database for some of these reports is, like, many gigs, like one of them 30 gigs. When we import the data into the PowerPivot, it cached it, and it compressed it as well. So the file size of this report is 5 megs, so really, it’s a huge compression. So 5 megs published to SharePoint, and everybody can view, based on their permission, view these reports. And that’s actually the benefit — and this actually helps us to get into the BI quickly and start building our BI solution.

TED KUMMERT: Well, that’s great. Thanks for sharing with us.

AYAD SHAMMOUT: Thanks. You’re welcome.

TED KUMMERT: Thanks. (Applause.)

I love the second story he told, you know, in taking what had been a manual process where they were doing things manually in Excel and bringing the data together, spending a couple of days every month, and now they shifted to something where they’ve built an application, and now it’s just the data’s refreshed as it need be, and now it’s just an asset that’s there for them to use and they’re saving time month over month. It’s a great example of the type of thing we’re trying to enable.

Now, with Excel 2010 and PowerPivot 2010, we’ve got a pretty powerful platform for building applications. We’ve looked at some typical kind of business intelligence applications, but we wanted to throw it out there to the community and see what would come back. So we did this Alpha Geek challenge where we challenged the community to show us the kind of things that could be built, and we were amazed by what came back. It came back from this community, it came back from the broader community. We saw some amazing things, things that made us wonder, “Is this really Excel?”

And we wanted to show you some of that, to potentially spark some imagination around the kind of things you can build with this platform and the kind of applications you’re going to be able to build.

So I’d like you to join me in welcoming our kind of hall of fame Alpha Geek, Mr. Donald Farmer, to the stage. Donald? (Applause.)

DONALD FARMER: Hey, good morning, Ted.

TED KUMMERT: Good morning.

DONALD FARMER: Thanks very much.

TED KUMMERT: Always a pleasure.

DONALD FARMER: Yeah, great. So we were going to talk about the Alpha Geek Challenge.


DONALD FARMER: Which was a competition we ran over six weeks in February and March last year where we invited people to send in their workbooks to us that they built using PowerPivot, and then we put them up for people to vote on. People could vote on Facebook and Twitter and so on. And then we took the three winners and then put them together and chose an ultimate winner. So we can we review those entries?


DONALD FARMER: Could I have the clicker?

TED KUMMERT: You can have the clicker.

DONALD FARMER: Thank you very much.

TED KUMMERT: I’ll take my water.

DONALD FARMER: You take the water, I’ll take the clicker.

TED KUMMERT: You are the Alpha Geek, after all.

DONALD FARMER: I am the Alpha Geek, there we go. So the very first one, we asked for just the most interesting analysis in general, kind of very general thing to get the competition started. The winner there was Dan Commongore, who made this amazing kind of analysis of employee morale based on the AdventureWorks data set. But you look at this, it looks more like an application than Excel, and people seem to like this a lot. It got a lot of votes on Facebook.

The next one that we looked at, well, the winner again was called Dan. I think there’s a little kind of pattern going through — I should do some data mining on that sometime. And so Dan English, what we asked for there was the most interesting data set. And what Dan English did was he analyzed the flight information of all the flights to and from the last BI conference in Seattle, which I thought was a very interesting twist on the data set. And, again, he analyzed that, created some analytics for it in PowerPivot, did a nice visualization of it, and he won the second round.


DONALD FARMER: The third one was about visualization, and this was Brian Fosse, and this is a remarkable visualization. This is Excel. It’s quite extraordinary what he did here. And we’ve got a lot of votes for that. So there were three really good Alpha Geek Challenge winners.


DONALD FARMER: As it happens, we had a late-breaking entry. We had three, and the fourth entry was somebody I didn’t recognize at first, Ted K from Microsoft.com. And —

TED KUMMERT: This slide wasn’t in the rehearsal last night, just so people are kind of clear.

DONALD FARMER: Well, I couldn’t quite understand the analysis. But as I went through it, I realized that what it was saying was that if you’re doing a demo in a keynote, your chances of promotion vary with the amount of applause you get. (Laughter, Applause.) Well, stop, stop, stop -—


DONALD FARMER: — because —

TED KUMMERT: It is the beginning of performance review season.

DONALD FARMER: It is, exactly.

TED KUMMERT: We’re going to do a little one of these at the end and just see where you all land.

DONALD FARMER: If you reach a certain tipping point, you can see on the analysis that your chances actually turn negative. If I get more applause than you, I’m in trouble. (Laughter.)

TED KUMMERT: Because I feel upstaged?

DONALD FARMER: You feel upstaged.


DONALD FARMER: I would never upstage you, Ted.

TED KUMMERT: Oh, yeah, never, never. And especially Amir, who is coming next.

DONALD FARMER: Right, exactly. So let’s see who won the overall Alpha Geek Challenge.

TED KUMMERT: I bet it wasn’t me.

DONALD FARMER: I’m pretty sure it wasn’t you.

TED KUMMERT: I’m positive. (Laughter.)

DONALD FARMER: So our overall winner was Brian Fosse, who is sitting here today. (Applause.) Do you want to stand up for us, Brian? So well done, Brian, that was a fantastic entry. And what we’d like to do is just switch through and show Brian’s work in action, if you like. So this is Excel. Brian coded this in Excel. And it’s an analysis of the kind of psychological attributes that help you work well with a team. There’s an analysis down here, a little radar chart of the team’s sort of psychology. This team is particularly benevolent and high achievement. Must be SQL Server. And here you can see that this person, John Smith, he is kind of — his openness to change, he’s interested in stimulation and self-direction. And you can see that kind of arranged on this chart.

Now, this is all coded in Excel, and we can have a look and see this person is also kind of promoted by stimulation, this person is basically more traditional, is very motivated by achievements and so on. And then Brian did this using a fairly basic data set in PowerPivot that he put together. It’s anonymized because we don’t want to expose people’s psychological attributes in a keynote. (Laughter.)


DONALD FARMER: Well, we have John Smith here because Pocahontas gave us permission to show his, but apart from that. But this is all coded in PowerPivot and Excel, and I think it’s really remarkable.

We were chatting with Brian backstage, as you remember —


DONALD FARMER: — and he said to us, you know, “I’m embarrassed to win this competition because I’m not a developer, I’m just on the business side.” And we said, you know, that’s the entire point of PowerPivot.

TED KUMMERT: That’s the idea.

DONALD FARMER: That businesspeople can do this kind of work and build this kind of analysis. So we’re thrilled to see this, and I think the overall competition was a high standard, and Brian’s a very worthy winner.

TED KUMMERT: Well, great.

DONALD FARMER: But good try, Ted.

TED KUMMERT: Thank you, thank you.

DONALD FARMER: Thanks very much.

TED KUMMERT: All right. (Applause.)

DONALD FARMER: Not too much, not too much.

TED KUMMERT: Yeah. Only 55 or 60 percent applause, whatever that means. Well, you never know what’s going to happen with Donald. I remember when I did the fairy tale and he was demoing right after the fairy tale and out he trots wearing wings and with a wand. So this is what you get. So back to our regularly scheduled program.

Now I’d like to shift our discussion with a look ahead to the future. And we’re not committing to specific features in specific releases, I’m not announcing what our future release roadmap is, but we did want to give you an idea of some of the things that are on our mind as we think about the future. And we’re going to frame it in terms of the industry-level conversations we’re all having, and then we’ll focus that in on the implications for business intelligence.

So the first conversation, the cloud, the cloud, everybody’s talking about the cloud. A lot of the discussions on the cloud center on the infrastructure-level benefits, and that’s correct because the benefits there are transformational. But there’s also implications for the user experience. We’ve seen today with this self-service user experience, which puts the end user at the center, they’re managing their applications and data and someone else is managing the infrastructure, and they never see that. Whether that infrastructure is hosted by you and your data center, but one of our partners, or by us and our public cloud offering. Cloud computing brings the need for these kind of rich, self-service end-user experiences.

The second part of this is, well, where are you headed in terms of providing business intelligence as part of your cloud offering at Microsoft? We’re in market today with Business Productivity Online Services. We’re in market with the Windows Azure platform, which includes SQL Azure. If you want to think about kind of the guiding principle for how we’re moving forward on these offerings, think of the principle of consistency. Think about consistent capabilities, a consistent platform, consistent application platform, consistent identity model, consistent tools. Everybody’s going to be able to use their skills, their assets, their knowledge in using our cloud services.

So SQL Azure became commercially available earlier this year. It’s a set of relational features oriented toward application developer scenarios. But our intent is to offer all the capabilities of SQL Server in SQL Azure and that includes reporting and analytics. We’re hard at work on that now. I’m not announcing the specific release timeframe for that, but this is something you’re going to hear from us shortly in the future.

The second conversation, the consumerization of IT, and I really think of this as a discussion about how the user experience evolved. A lot of things happen first on the Web, they happen first in the consumer scenarios, but then it becomes very natural for us, you know, as that becomes the way we interact with things, for that to be the way we interact with things in our work lives as well.

Consider search. Years ago, we were just using search to access information on the public Internet. Now we use it everywhere to access information within our own enterprise. And now we’re bringing BI seamlessly into that same search experience.

Consider social media. We’re bringing BI into the SharePoint environment, enabling sharing and collaboration, but we saw tagging and ranking and that kind of social media experience, which is — really enriches the experience for end users. We think generally about investing in more immersive user experiences as well. One example Ayad showed, you know, how much more compelling is it to have seen that data laid out in a map using the mapping control in SQL Server 2008 R2 than it is to look at that data in a list by region or by state? It’s much more compelling. You’re more compelled to use it, and it enables you to gain more insight.

A third discussion is the compliance discussion. One of the ways that centers in our conversations about BI is on data quality. One of the key data quality problems is that your systems, your applications, your BI infrastructure all has various opinions about what the reference content is, what the key reference data is for the enterprise. You need a place to bring that data together and resolve conflicts and manage that. That’s the area of master data management. In SQL Server 2008 R2, we delivered the first release of master data services, which is our solution for those master data management challenges.

Now, we’re also investing in our data cleansing technology. We think machine learning can be applied here, and we also see an opportunity to make the experience around this an integrated thing. The data steward has one experience as they think about master data, as they think about kind of the data cleansing and data quality test, to bring that stuff together in one experience. And we think we can make this a lot more easy and accessible thing for people to use. So you can start small, and you can grow the solution over time.

Another compliance discussion is the lineage and the impact analysis topic. You know, lineage is the question of how did I get to this answer? What led to the calculation that I’m looking at? What data sets when led to this thing that I’m looking at here? Impact analysis says if I make this change, what’s the dependency tree so I know what I’m going to break? Whose report isn’t going to work tomorrow because of the change I’m making?

This requires thinking deeply about a metadata strategy across the platform, including an extensibility strategy. And this is a thing that we’re very, very focused on and spending a lot of time on.

The last conversation is about the increasing data volume. And we all know where this is coming from. This is from higher transactional rates, this is from different types and shapes of data that we want to include. It’s different sources of data, data from sensor networks, RFID. You need platforms that can store that high amount of scale and allow you to analyze those types and shapes of data on that scale.

Later this year, we’ll be releasing the first version of our parallel data warehouse edition of SQL Server. This will add scale-out data warehousing to the SQL Server product family. This will take SQL Server to the hundreds of terabytes in data warehousing. You’ll have one solution now that you can use from the large kind of enterprise single version of the truth, mission-critical data warehouse to the surrounding data warehouses and data marts to the surrounding reporting and analytics work loads. It’s one platform for IT; it’s one platform for end users.

We’ve just finished our second technology preview. We’ve got a lot of great feedback from customers about query performance, as well as data load performance. So we’re well on our way. And we’ve been working closely with our partners in the hardware community to build reference configurations. So this is going to be a very easy experience to bring out of the box and get up and running. And with choice of industry standard, hardware platforms from these vendors, we’re going to have very low total solutions cost as well relative to other competitors in the industry.

Another thing we think about is enabling you to bring together data easily from other sources. We announced at our last professional developers’ conference a data marketplace service code-named Dallas. And this data marketplace is going to be a place where you’re going to be able to go and find both public data sets, as well as commercial data sets. And we’re going to integrate the Dallas consumption experience into our experience, so it’s going to be easy to go get these data sets, find and purchase what you want, then you’ll be able to bring that type of reference content easily into the solutions you’re going to build.

So we want to show you some code now, some stuff that’s kind of hot off the developer machines. And we’re not standing here today saying all of this is going to appear in exactly the next release, but we wanted to give you a glimpse of a couple of the things that we’re working on.

One of the areas is just kind of connecting the dots between this self-service capability that we’ve delivered and our professional BI platform and tools, how we’re unifying the models, and how we’re unifying the experiences spanning from the end user in Excel through to Visual Studio and BI Development Studio. I want to show you that.

The second thing we want to show you is some new data visualization solutions we’ve been working on. We’ve got — this will be rolling out in about the next 30 days for you to play with, but it’s a great example of kind of how one of these more immersive and compelling user experiences really enables you to gain more insight.

So I’d like you to join me in welcoming Amir Netz, distinguished engineer from the business intelligence team to the stage. Amir? (Applause.)

AMIR NETZ: Hi there.


AMIR NETZ: Hey, guys. So I hope most of you saw the keynote yesterday.

TED KUMMERT: Yesterday. Feels like a long time ago.

AMIR NETZ: You know, I got a lot of great feedback on the keynote, and people came back to me with lots of questions and said, you know, mostly two questions we had. One of them was, “Amir, you know, this people tier that you showed, how do I build my own applications on it?” And the other one was, “Amir, how do you guys come up with such a confusing name?” So I am here to answer the first question. And to do that, we are going to go and take a look at the real application we actually use here at Microsoft.

And this is an application that was authored by our own VP, he’s running the American group for EPG. The EPG group is selling software to the largest enterprises. You know, he has a lot of account managers, account managers have quotas, and his job is to meet the quota. And he built this application. In the center of the PowerPivot application, we have here something called the waterfall analysis.

We have your quotas, you know, everybody gets what a quota is. But then there is ITB. ITB stands for “in the bank” — means that these are deals that we closed and we’re going to get paid for or already got paid for, you know, things that we are sure of.

And then we have something called cold deals. Cold deals means deals that we are pretty sure that they’re going to happen, not really fully signed yet, but we are pretty sure about them. Runways are deals that are small deals that are just happening, nobody notices, they might just be coming in. Then the quota, then the outside. The outside means deals that are still morphing, and we’re not sure what will happen with them, and maybe we’ll get some money, just gravy.

OK, so Chris will go and look at this and say, “I’m going to slice for the two, the eastern region, look at this team, look at Jacqueline, who is an account team unit manager, we can look at the specific account managers.” This is a real application we use at Microsoft, and we change the numbers, we changed the names of the people to protect the innocent, but it’s the real thing, right?

So Chris loves it. He built it, he loves it, he uses it. But he tells me, you know, “Amir, this would be perfect for me if I was an accountant. I just care about the numbers, that’s the only thing I care about, it’s perfect. But I’m also a people manager. And I care about the people as well, and this is all about the numbers. I want something that focuses on the people, puts the people at the center.”

So we thought about this people-view experience. And we thought, you know, we can go and build one where we’re going to have these people in the center of it. And how would we go and do that? Well, we start with our reporting services report. We’re going to create a scorecard here for each account manager. So let’s look at such a report we have here. And we built one. This is kind of a little bit untraditional scorecard. We see here at the bottom, you know, we see here at the top the name of the account manager we’re going to show in this scorecard. It’s going to be a prioritized report, we’re going to get the account manager name.

We see here at the bottom this waterfall chart that we have created again here in this report, and then we need two more things that are kind of special for this kind of application. First of all, we put here something, which has an image on it. And this image here is going to be the image, the photo of that account manager.

The second thing that we’ve done is we have used the background color of the report, we put here as a conditional expression. I’m going to show you the conditional expression. And this expression basically says look at the gap from the quarter for that account manager. If the gap is big, put a red background, otherwise a very small gap or no gap at all, put a green background, otherwise yellow. And with all this, we can take a look how one account manager looks like, once we run it. See, this is kind of how one scorecard looks, see the background color around. You can see how important this background color is going to be.

We’re going to see here the photo of the account manager, and we’re seeing the waterfall, right? We have this. And now we can — you can see parameterize, I can put other names of other account managers. But what I’m going to do is instead of looking one account manager at a time, I’m going to start generating those -— I want to generate those report cards or scorecards for each of the account managers. And to do that, we’re using a new utility that we have created. It’s called People Viewer Extension for Reporting Services that essentially automates the whole process for us.

So we are going to look here. We have a collection of those people-viewer applications, we have four of them here, we’re going to look at the fourth one, and we look at how we create one of those applications. Very simple. You have to have basically two components. One of them is the report itself that we’ve just seen, the template for each one of those scorecards, and the other one is a query, in this case, the query is coming from our own PowerPivot application through the XLS file, and you can see, and this query is going to read all the information about the account manager, and just, we’re going to use it to iterate through those reports and generate those scorecards for each one of those.

And when we have those two to marry together, we can just go in and start crawling and start producing it. So I’m going to go in and start the production and say “crawl.” And what it does, it executes that query against the people-viewer workbook, actually against analysis services that holds that people-viewer workbook, it starts generating all these scorecards for all the people. We have hundreds of people. I’m going to stop it at this point. I’m just going to show you how it looks like at the end.

So this is the application at the end. And you can see that we have here all our hundreds of account managers, these are small thumbnails. And now you can see how it looks like, right? The people who are doing really well have this thumbnail with this green frame. And you can see here the people with the red frame, you know, not so good.

TED KUMMERT: Not so good.

MIR NETZ: And you can start playing with it. You seen maybe yesterday similar thing, let’s go, and let’s break them by regions and see how they split. Drill down into the eastern region, look at them by district, maybe look at those that are more challenged. So let’s go that you guys need some help. So go with the gap ratio and look at those that have a gap that is positive, which means a negative thing. And we do that. Now maybe focus on one of those people. So let’s look around. Here’s Eugene. You can actually look how well Eugene is shaving. (Laughter.)

Eugene is not happy. You can see that, right? (Laughter.)

TED KUMMERT: He appreciates you doing this.

AMIR NETZ: Yeah. Eugene is not happy because he’s doing really, really, really not good. Right? He has a quota of about 14 million, he’s in the bank only 1.8 million, core deals are only 4.4, this is a gap, the guy is in deep, deep trouble, right? (Laughter.)

Now, you can see how Eugene is doing compared to his group. So one of the things we have here is we have all the information about Eugene. And you see his quota, you see everything about him. You see one of the things is that his manager is Destiny. And if I click on Destiny, I’m going to find all the other account managers that are reporting to Destiny, share the same manager. So I’m clicking on Destiny, and we’re going to just go find the rest of the account managers. I can see the team is really not doing well.


AMIR NETZ: Maybe. And let’s look at this guy, right? We can go and take a look, this is Ian. We call Ian the raging bull, you can figure out why. (Laughter.) You know, he’s not looking very happy. We caught him at a bad time. You can figure out maybe by yourself what kind of time was it. And what we have here, Ian is actually doing really well. Look at this. He’s actually over the quota, he should be happy.

Now, these are just some of the insights we can get. You can go and take a look and see whether experience matters, for example. So we can go and look at experience, the breakdown here, it’s very obvious. Look, one to three years people, lots of red. If you have been with the company for many years, lots of green. Stay with us, right?

Or we can go in and say, you know, let’s see who brings the most money, people in the bank, most money in the bank. Let’s look at the tail end, who our heroes are. And here we go, the two heroes. And look at Jonathan here, you know, look at that. The guy had a $63 million quota, already have in the bank 46, the guy is doing well. That’s a moneymaker, right?

So a lot of this kind of great — it’s just fun. You can play with it for hours, it’s just lots of fun. And what’s really interesting is we always kept the people in the middle. And whenever you care about doing analysis where you don’t want to lose the individual items, whether it’s people or stock or projects or big customers, and you never want to lose those, not just aggregate matters, but also the individuals matter, it’s a fantastic experience.



TED KUMMERT: So this is going to be available I think in the next 30 days.

AMIR NETZ: Next 30 days.

TED KUMMERT: For people to download and play with.

AMIR NETZ: Yes. You can download it, the whole system — you know, just the things that take you from a report to a complete people-viewer system all available in the tech preview in the next 30 years.


AMIR NETZ: You guys can ask more questions in the booth, OK?

TED KUMMERT: Absolutely.

AMIR NETZ: But enough with fun and games, let’s talk serious stuff.

TED KUMMERT: Yes, let’s talk about the professional BI experience.

AMIR NETZ: Professional BI. Right? So we got a lot of great feedback about PowerPivot. You know, lots of you play with it — how many people played with it? Oh, wow, look, it’s fantastic. Got some great feedback on it, and people realize how quick it is to get up and running, how simple and easy to learn, it’s very easy to learn. It’s really performing wonderfully. But when it comes to the BI professionals, they are telling us that it’s not enough. You know, they want to do more.

So I’m showing you here a real application that we have — this is a loan-risk analysis based on real data from a real lender. And you know lenders, you know, lenders give money to people and what they really hope is the people will pay them back the money. They really, really care about getting paid back.

And you see the thing that they care about is, you know, some people don’t pay back, and they care about the percent of losses that they’re going to have. So we have here risk analysis, you can analyze your loans by different dimensions. You have the percent loss, which is really a very, very important metric for any lender.

And when we have a very important metric, we like to put it as a KPI. And there are no KPIs in PowerPivot. There are KPIs in analysis services, there are KPIs in the BI Development Studio, everybody knows how to use KPIs if you’re a BI professional, but there are no KPIs in PowerPivot.

And people said, “OK, so can we get these kind of features, the great features like KPIs that you have in BI Development Studio? Can we get them also in PowerPivot?” Now, the good news is the PowerPivot engine is Analysis Services, so we actually have the capability of the engine, we just didn’t have the user experience for it when we shipped PowerPivot. So we’re adding these kind of features in.

And let’s take a look, when we add in this KPI features, we’re going to do it in a different way. You will not need to know MDX, it’s very much the PowerPivot way of doing KPIs. So I’m opening up the KPI dialogue, and this is how it looks like, it’s a very visual thing, right? So you go in, and say I’m going to set here the ranges, let’s say, you know, up to let’s say 4 percent is great and beyond 4 percent to let’s say 12 percent is OK, and otherwise — and you can see every time I make a change, it immediately shows up. Of course, the rates should probably be green, so I can switch the colors and immediately, you can see how it switched the color in the image.

I can select other kind of indicators. I don’t have to stay with three states. I can go with things like five-state indicators and so forth, maybe even go with something that looks like this, but in this case, I think three states is great. And it’s just very easy, very visual. You don’t need to know MDX, you don’t need to know coding, you don’t have to open Visual Studio for this, it’s just in your face here in PowerPivot. This is the way we’re trying to bring this classic great features for Analysis Services to the PowerPivot experience. (Applause.)


AMIR NETZ: Now, the other thing that we hear from the BI professional is as much as they love the way we are doing the visual development of applications in PowerPivot, when the applications become big, you know, complicated, enterprise BI applications, the experiences that we have here are kind of challenging. Because look at this, this is the loans table. The loan table is the primary table we’re going to have in this analysis. You know, this is everything we know about the loans, and you can see it’s kind of a fairly wide table.

And if you go and start — and we have to build a lot of calculated columns and it’s becoming pretty challenging to scroll left and right — we have such big tables and such a lot of calculations, and that’s how big, complicated BI applications look like.

So we thought, you know, maybe we can do better. And we have something that is kind of experimental I’m going to show you; it’s also not very polished yet, but I hope you understand what we’re trying to do here.

We’re going to do something that says, “Can we instead of just limiting ourselves working with a loan being just one row, build something that allows you to work on a two-dimensional grid?” Now, what you see here is the same table. The same 15,000 rows. And I can scroll here. All the loans that I have available here, instead of just giving one line per loan, I’m giving you a whole grid per loan. And now it’s all the fields, all the columns are laid out, and you have titles. You know, you put them in groups so you can see this is general performance parameters and all the fields that are related. And we see here the calculated column in blue, and I can click on one and see exactly the expression that it is having here.

So everything is laid out. I can go and operate here. So I’m going to say, you know, let’s create a column here. So I’m going to say this is debt to income ratio, you know, very, very common calculation for lenders. And I say, let’s put here a calculation, it’s just a point and click. So I need to take how much the borrower already owns, I’m going to add to that how much he’s asking to request to borrow from us, I’m going to divide it by the monthly income that the borrower has times 12 to get the annual income, and I’m getting how many years it will take the borrower to return all his debts. OK, that’s kind of what the measure is.

When I do that, I just created a calculated column in the table. But it’s a much, much better experience to do that. Now, we’re still working on it, and actually, people cannot see this, so we’re still working on it. But if you tried this and you work here, you will never want to create columns in the flat table again — this is a much, much better environment.

So we do that, and then we continue to get more feedback. And the other feedback we get is it’s a great environment if you are a single developer, but if you have a big project with multiple developers that build a lot of different parts of the project, then they really, really miss the kind of feature that we have in the BI Development Studio, you know, check in and check out, versioning, team development, the notion of development, test, production deployment, all of that stuff is something that is really valuable, and we don’t get that with Excel. Excel is not designed as a professional development environment.

OK, we’ll have to take that into account. So we said, maybe we can just get the same experience that we have in Excel, get it into the BI Development Studio. So I’m going up here, and I’m opening BI Development Studio, and I’m going to open and import from PowerPivot. And once we get here, the credit analysis workbook we’ve just seen, and I’m going to import it in. And I’m going to get everything here now running inside Visual Studio. It’s the same kind of experiences, you know, you can see here instead of having big ribbons on the top, we get property panels, and we get the solution explorer and you get source control and everything else. But except for that, it is the same kind of experiences that you have in Excel. So you get the ability to go and look at the loans table, and you go and you can enter the formulas on the top. It’s the same thing. You work visually, but now it’s in the professional development environment.

Now, the other thing that we’ve done is we added even more professional features in here. For example, these applications are pretty big and complex, and it’s hard to track the calculation, how they relate to each other, we don’t have here scripting windows. So we thought maybe we can help.

So we’re working here on something that is giving you — you talk about lineage and impact analysis, let’s talk about lineage and impact analysis. We’re now looking at the view of the flow of the calculation in that workbook. For example, we looked at the percent loss, we can look here and see this is the expression of the percent loss. It made out a number of loans divided by the number of loans lost divided by the number of loan count, and we can go in and, you know, we can see here this is the number of loans lost and the expression for it, number of loan count, and you can see exactly all the components that are making up this calculation. And I can do it for all the calculations that I have in this workbook.

So I can do it for everything else and look at it, or I can go and do the opposite. Impact analysis. What will happen if I change the issue date? And I can see the flow of the calculations that are depending on issue date here. So you can see both ways and track it. It’s a great way to understand what’s happening within your calculation to do impact analysis, the lineage, or to just document what’s going on in the application.

The last thing that we are hearing is that, you know, the whole, you know, all these wonderful demos of 100 million rows and so forth in Excel are great, but the big applications, the really big applications that we have are much bigger than that. 100 million rows are not enough.

So, OK, so let’s look at another project that we have here already loaded into Visual Studio here, the BI Development Studio. And one thing that happens when you work in the BI Development Studio, you’re not working anymore against a local engine in Excel, you’re working against analysis services. You’re connecting to a real analysis services, a real server, and because of that, you can expect that you’ll be a stronger machine with more memory, and you see here, we have a much bigger server.


AMIR NETZ: With more memory. And because of that, we actually can have more data. So let’s look a bit at this. You know, this is a moving data set that we saw yesterday in the keynote with just a bit more data. Let’s look at how much data we have here. This is a purchase table, all the cell transactions. And we’re going to look. This one has over 2 billion rows. We made it 2 billion and seven so I can say over 2 billion rows. (Laughter.)

And, you know, we’ve seen yesterday the kind of experiences, the performance we had with 100 million. Let’s take a look at how it performs with 2 billion. So let’s do the same thing. Sort the retail price from the smallest to the largest. Click. Done. Two billion rows just sorted. Let’s do it the other way around, from the largest to the smallest, done.

Let’s go and do some filtering. Let’s go see just the sales in euros, let’s see, so we have two billion rows being filtered down to 26 million. OK, this is the performance, working with 2 billion rows. You know, with talk about wicked fast yesterday, this is beyond wicked fast, this is the engine of the devil, right? (Laughter, applause.)

TED KUMMERT: You said that, I didn’t. (Laughter.) How about your performance review?

AMIR NETZ: Remember the applause. (Laughter.)

TED KUMMERT: It’s the same storage technology —

AMIR NETZ: It is the same storage technology, it is — actually, what we have here, what you’re seeing here is something that — it’s the same engine that powers PowerPivot, there is no aggregation, no caching, no indexing, no tuning. So we just uploaded the data and let it run. That’s it. It just happened. That’s how it performs natively.

Now, let’s look at it in a different context. I’m going to upload here another workbook, it’s the movie analysis. And we’re going to analyze some movies here. And this one is not connected to PowerPivot; let me just enable the content here and close this so we can see it all together. This one is connected to that 2-billion workbook, 2-billion-row workbook that we had before —

TED KUMMERT: So we’ve connected the PowerPivot experience in Excel to Analysis Services.

AMIR NETZ: Yeah. No, I want you all to understand what’s happening here. Right? We have 10 widgets on the screen, six slicers and four charts. And each one of those sends two queries to the data source in order to render itself. So we have about 20 queries being sent. Every one of those queries is a full table-scan of the 2 billion rows. So all together whenever I click, we are scanning 40 billion rows, and it takes about two and a half seconds to do that. So if you just do the math in your head, we are seeing here a scan rate of a trillion rows per minute. That’s kind of what we’re talking about. Performance review.

TED KUMMERT: Performance review.

AMIR NETZ: Right? (Laughter.) So this is the kind of experience that we are seeing here.

So I hope you’re kind of getting an understanding of what we’re working on and where this is heading. We are taking the PowerPivot experiences, and we are expanding them to bring all these great classic features of Analysis Services to that experience. We are enhancing the authoring capabilities both in tools and environment putting it back into the BI Development Studio for the professional, you know, professional environment for BI professionals. And we are bringing you the scale that you need in order to handle the largest possible applications of BI.

And we’re doing it all by maintaining a common platform of technology, of model that can stand self-service BI, corporate BI, analysis and reporting all unified under the same platform. That’s what we’re working on.

TED KUMMERT: That’s great. Thank you.

AMIR NETZ: Thank you. (Applause.)

TED KUMMERT: Life is interesting working with these guys. That’s all I have to say. It’s great to have people, his enthusiasm is contagious.

Well, so we spent a little time today talking about the present, where we are with managed self-service BI and talking a little bit about the future. One of the things we really wanted to emphasize by showing you some of what Amir showed you just now is our deep commitment to continue to evolve the experience, the tools, the capability for you, the professional BI users. We talk a lot about enabling all the end users and really have made a big investment in the other 80 percent, but we’re very, very committed to continue to invest to make best-in-class tools for you.

Now, I’m not sure of the fairy tale, did we slay any dragons? Did people find true love? I don’t know, the keynote was kind of long, who knows? Not sure, nothing that serious, but I think we are making progress in enabling this vision of BI for everyone. Managed self-service BI is a pretty transformational step forward for the BI industry. BI technologies can be used by more end users than there are today. They can get a lot of things done that they ask you to do.

There’s a tremendous amount of power in that. There’s a tremendous amount of power for you. Your solutions will be more widely used. They’ll be built upon. You’ll see less interruption, and we’re continuing to move forward to enhance the tools and the capabilities for you. So you’re going to be more effective in building the strategic applications that you’re responsible for.

Business intelligence is so empowering. I mean, it truly is magical. You make people look better. They now know something they didn’t know before. They have some insight they didn’t have. They’re going to be more effective at their jobs tomorrow. Business is going to be better. That’s the work that you do. It’s tremendously empowering. And with the last couple of years of this economic climate we’ve been in, the priority on this type of solution has only increased. The work you do is tremendously important.

We’re really proud of our partnership with you. We’re really proud that we’re able to work with you to build tools so you’re able to do the work that you’re able to do. It’s a great thing for us. I love seeing all the customer solutions we’ve looked at today — what came back from the community is just tremendous.

We’ve got an exciting future ahead of us. We talked a little bit about where we’re headed with SQL Server, how we’re taking it in scale-out data warehousing. We showed you some new end-user experiences that we think are going to be more compelling, really enrich the experience, enable more insight. We talked about how we’re going to move forward and offer our business intelligence solutions in terms of cloud services and how we’ve already got this user experience in terms of the self-service user experience to enable that and make that happen.

We talked about how we think about really bringing together these two worlds, kind of the self-service capabilities all the way through to the capabilities for you, the professional BI practitioner.

Well, I want to thank you for your time today. I hope all of you have a terrific week at TechEd, at the Business Intelligence Conference. In New Orleans, it’s been terrific to spend this time with you, so have a great week and thanks very much. (Applause.)


Related Posts