You are currently browsing the tag archive for the ‘Business Intelligence’ tag.
Microsoft has been steadily pouring money into big data and business intelligence. The company of course owns the most widely used analytical tool in the world, Microsoft Excel, which our benchmark research into Spreadsheets in the Enterprise shows is not going away soon. User resistance (cited by 56% of participants) and lack of a business case (50%) are the most common reasons that spreadsheets are not being replaced in the enterprise. The challenge is ensuring the spreadsheets are not just personally used but connected and secured into the enterprise for a range of consistency and potential errors that all add up to more work and maintenance as my colleague has pointed out recently.
Along with Microsoft SQL and SharePoint, Excel is at the heart of the company’s BI strategy. In particular, PowerPivot, originally introduced as an add-on for Excel 2010 and built into Excel 2013, is a discovery tool that enables exploratory analytics and data mashups. PowerPivot uses an in-memory, column store approach similar to other tools in the market. Its ability to access multiple data sources including from third parties and government through Microsoft’s Azure Marketplace, enables a robust analytical experience.
Ultimately, information sources are more important than the tool sets used on them. With the Azure Marketplace and access to other new data sources such as Hadoop through partnership with Hortonworks as my colleague assessed, Microsoft is advancing in the big data space. Microsoft has partnered with Hortonworks to bring Hadoop data into the fold through HDInsights, which enable familiar Excel environments to access HDFS via HCatalog. This approach is similar to access methods utilized by other companies, including Teradata which I wrote about last week. Microsoft stresses the 100 percent open source nature of the Hortonworks approach as a standard alternative to the multiple, more proprietary Hadoop distributions occurring throughout the industry. An important benefit for enterprises with Microsoft deployments is that Microsoft Active Directory adds security to HDInsights.
As my colleague Mark Smith recently pointed out about data discovery methods, the analytic discovery category is broad and includes visualization approaches. On the visualization side, Microsoft markets PowerView, also part of Excel 2013, which provides visual analytics and navigation on top of the Microsoft’s BI semantic model. Users also can annotate and highlight content and then embed it directly into PowerPoint presentations. This direct export feature is valuable because PowerPoint is still a critical communication vehicle in many organizations. Another visual tool, currently in preview, is the Excel add-in GeoFlow, which uses Bing Maps to render visually impressive temporal and geographic data in three dimensions. Such a 3-D visualization technique could be useful in many industries. Our research into next generation business intelligence found that deploying geographic maps (47%) and visualizing metrics on them (41%) are becoming increasing important but Microsoft will need to further exploit location-based analytics and the need for interactivity.
Microsoft has a core advantage in being able to link its front-office tools such as Excel with its back-end systems such as SQL Server 2012 and SharePoint. In particular, having the ability to leverage a common semantic model through Microsoft Analytical Services, in what Microsoft calls its Business Intelligence Semantic Model, users can set up a dynamic exploratory environment through Excel. Once users or analysts have developed a BI work product, they can publish the work product such as a report directly or through SharePoint. This integration enables business users to share data models and solutions and manage them in common, which applies to security controls as well as giving visibility into usage statistics to see when particular applications are gaining traction with organizational users.
Usability, which our benchmark research into next-generation business intelligence identifies as the number-one evaluation criterion in nearly two-thirds (64%) of organizations, is still a challenge for Microsoft. Excel power users will appreciate the solid capabilities of PowerPivot, but more casual users of Excel – the majority of business people – do not understand how to build pivot tables or formulas. Our research shows that only 11 percent of Excel users are power users and most skill levels are simply adequate (49%) compared to above average or excellent. While PowerView does give some added capability, a number of other vendors of visual discovery products like Tableau have focused on user experience from the ground up, so it is clear that Microsoft needs to address this shortcoming in its design environment.
When we consider more advanced analytic strategies and inclusion of advanced algorithms, Microsoft’s direction is not clear. Its Data Analysis eXpressions (DAX) can help create custom measures and calculated fields, but it is a scripting language akin to MDX. This is useful for IT professionals who are familiar with such tools, but here also business-oriented users will be challenged in using it effectively.
A wild card in Microsoft’s BI and analytics strategy is with mobile technology. Currently, Microsoft is pursuing a build-once, deploy-anywhere model based on HTML5, and is a key member of the Worldwide Web Consortium (W3C) that is defining the standard. The HTML5 standard, which has just passed a big hurdle in terms of candidate recommendation is beginning to show value in the design of new applications that can be access through web-browsers on smartphones and tablets. However, the success or failure of its Windows 8-based Surface tablet will be the real barometer since its integration with the Office franchise is a key differentiator. This approach of HTML5 could be challenging as our technology innovation research into mobile technology finds more organizations (39%) prefer native mobile applications from the vendors specific application stores compared to 33 percent through web-browser based method and a fifth with no preference. Early adoption of the tablet has not been strong, but Microsoft is said to be doubling down with a new version to be announced shortly. Success would put Office into the hands of the mobile workforce on a widespread basis via Microsoft devices, which could have far-reaching impacts for the mobile BI market.
As it stands now, however, Microsoft faces an uphill battle in establishing its mobile platform in a market dominated by Android and Apple iOS devices like the iPhone and iPad. If the Surface ultimately fails, Microsoft will likely have to open up Office to run on Android and iOS or risk losing its dominant position. My colleague is quite pessimistic about Microsoft overall mobile technology efforts and its ability to overcome the reality of the existing market. Our technology innovation research into mobile technology finds that over half of organizations have a preference for their smartphone and tablet technology platform, and the first ranked smartphone priorities has Apple (50%), Android (27%) and RIM (17%) as top smartphone platforms with Microsoft a distant fourth (5%); for tablets is Apple (66%), Android (19%) and then Microsoft (8%). Based on these finding, Microsoft faces challenges on both the platform front and if they adapt their technology to support others that are more preferred and used in business today.
Ultimately, Microsoft is trying to pull together different initiatives across multiple internal business units that are known for being very siloed and not organized well for customers. Ultimately, Microsoft has relied on its channel partners and customers to figure out how to not just make them work together but also think about what is possible since they are not always given clear guidance from Redmond. Recent efforts find that Microsoft is trying to come together to address the big data and business analytics challenge and the massive opportunity it represents. One area in which this is coming together is Microsoft’s cloud initiatives. Last year’s announcements of Azure virtual machines enables an infrastructure-as-a-service (IaaS) play for Microsoft and positions Windows Azure SQL Database as a service. This could make the back end systems I’ve discussed available through a cloud-based. Ironically, the cloud-based Office365 suite does not include core productivity applications such as Excel and PowerPoint, so front-end access will still come through the client version of the software.
For organizations that already have installed Microsoft as their primary BI platform and are looking for tight integration with an Excel-based discovery environment, the decision to move forward is relatively simple. The trade-off is that this package is still a bit IT-centric and may not attract as many in the larger body of business users as a more user-friendly discovery product might do and address the failings of business intelligence. Furthermore, since Microsoft is not as engaged in direct support and service as other players in this market, it will need to move the traditionally technology focused channel to help their customers become more business savvy. For marketing and other business departments, especially in high-velocity industries where usability and time-to-value is at a premium and back-end integration is secondary, other tools will be worth a look. Microsoft has great potential and with analytics being the top ranked technology innovation priority among its customers I hope that the many divisions inside the global software giant can finally come together to deliver a comprehensive approach.
VP and Research Director
Information technology for business is changing rapidly as organizations demand innovation to help them discover insights and facts. Our research into business technology innovation found analytics to be the top priority in 39 percent of organizations. Businesses feel pressure to be better, faster and smarter in operating processes, and understanding their various types of information is a key to success. Businesses are looking to capture value from all types of information both within the enterprise and on the Internet. In this context technology providers are now using the term “discovery” to capture potential buyers’ attention; it became an area for technology spending in 2012 and likely will be for years to come. In fact my colleague Tony Cosentino has identified discovery as one of the four pillars of big data analytics.
Discovery is one of many business analytics methods that can be used realize value from current and future investments into big data. Discovery, of course is the act of finding something, whether it’s truly new or just overlooked. Wikipedia adds, “With reference to science and academic disciplines, discovery is the observation of new phenomena, new actions or new events and providing new reasoning to explain the knowledge gathered through such observations with previously acquired knowledge from abstract thought and everyday experiences. Visual discoveries are often called sightings.”
In business the knowledge gathered by individuals engaged in discovery is critical to provide context; typically those people are analysts responsible for the organization’s analytics or increasingly are business professionals competent to delve into the discovery process; for either, analytic technology should provide meaningful information in dynamic fashion. Done right, discovery produces intelligence, and analytic tools have improved the usability that enables more people to discover insights using this class of technology. I have already pointed out why conventional business intelligence is failing business; improving on these failings should also be a guide to what we need from discovery technology. With the wide adoption of big data technologies in varying approaches, organizations need to find the right tools to take advantage of it, but adequate data and visual discovery are not currently available in almost one-fifth (19%) of organizations participating in our technology innovation research.
There are four main types of discovery for business analytics: in no particular order, they are event, data, information and visual. Let’s consider each of them and the potential they hold for realizing full value from business analytics and big data investments.
Event Discovery: Enterprise networks now must handle extreme velocity in the streams of events that pass into and through them. If they are processed effectively, discovery through analytics could reveal current bottlenecks and opportunities for improvement or if trended and projected could indicate patterns developing in a negative direction. Processing events in a real-time or right-time manner has evolved from complex event processing into a category of its own, operational intelligence. Our research in this area found that nearly half (45%) organizations consider it very important to analyze business and IT events, and another 44% indicated it is somewhat important. The process of discovery applied to events can take many directions; for example, discovery analytics can immediately notify someone to take action, or the results can be displayed visually to make it easier to identify outliers or trends that should be further analyzed for review. As well, discovering relationships between events is very important to 53 percent of organizations. But the right tools are necessary for success. Our research shows that the large majority (91%) of organizations that use specialized tools for this are satisfied with them, compared to 76 percent that use general-purpose BI tools. The role of event discovery, now being called big data in motion, is changing rapidly as Tony Cosentino has pointed out.
Data Discovery: Enterprise databases contain ever larger volumes of structured data that describe transactions and interactions with their customers, their products and employees, the locations where their business operates and relations with their partners and suppliers. This kind of data is significant and can be sourced from in-house business applications and data warehouses and from software rented in the cloud computing environments, as well as through new investments in big data. From whatever source, having more interactive and data-centric discovery is critical to empower analysts and even data scientists. Data discovery is not new, but it has evolved greatly. Intuitive and flexible a new tools have advanced from the foundation of OLAP to perform data discovery on large volumes of source data and place it into a business model or analyze it while still in its almost original formats. The ability to combine and relate data from Internet sources expands the realm of what is possible to know and act on while expending less time and resources. Many business intelligence suppliers are just beginning to see what is required to meet the needs of today’s analysts compared to the past needs of IT or BI professionals needs to publish reports and dashboards. Our research finds that exploring data underlying analytics in a discovery manner is a critical business analytics need in 37 percent of organizations. Some new big data-oriented analytic tools can do more comprehensive data discovery, and IT departments will have accommodate them to provide what analysts need to do their jobs.
Information Discovery: Organizations now must handle a broader variety of information than ever before, including documents and semi-structured content whose data is not contained in a database. This wealth of information provides an opportunity to increase business understanding, but users need to access and integrate it for a range of tasks such as fraud, risk and compliance; process improvement; and reporting and analytics. This information can also be combined with data from databases to provide a comprehensive foundation for applying discovery processes. Our research shows that content is the second-most important type of data in 59 percent of organizations, right after customer data (71%). We believe that information optimization is a key benefit of big data and information management investments, but as I have pointed out it requires flexible technology to utilize all this information. Information discovery was once left to very expensive technology and specialized resources, and hence was beyond the reach of many organizations. Now many vendors offer tools that can perform content analytics to discover key information in proprietary formats and harvest it for business operations. The new generation of tools also provides the ability to define templates and placement for information to be integrated without the work of developers; it can analyze the layouts of documents and process their contents on an automated basis.
Visual Discovery: This is the latest technology craze as analysts clamor to add it to their analytic tool sets. Our research finds that presenting data visually is the second-most demanded critical business analytics capability for 48 percent of organizations. This type of discovery helps visualize large volumes of data to add new focus to finding areas of opportunity and challenges. Visualization may seem deceptively easy, but it is actually quite difficult to design technology that a range of nontechnical roles find easy to understand and present. I have seen vendors just attempt to lay more sophisticated visualization on top of their existing products, but doing this does not produce the usability and interactivity users insist on; lacking these qualities has severely hindered adoption. Being able to use visualization as a selection method for further discovery dramatically reduces the time it takes users to analyze data and find new insights. At some point in the visual discovery process, users want to share a visualization or a chart for further collaboration, to identify potential places for root-cause analysis or to make recommendations for resolution. At this point in development, however, most of the technology in visual discovery is not able to easily associate comments or bullets to a specific view and then enable sharing or collaboration on it electronically.
Not all discovery technology is created equal, as my discussion of the four big types of discovery shows. Some tool providers excel at one type, but few do all of them well. Thus your organization will have to create a discovery strategy as part of your business analytics efforts and choose and budget appropriately as you identify your most critical needs. You certainly will need to do something: Our business analytics research finds that analysts in 42 percent of organizations spend the majority of their time on data-related activities instead of actual analysis, so it is vital to reduce these chores and that ensure discovery methods do not increase that time; have IT staff automate these tasks through data integration and virtualization. For the CIO, it is important to ensure you are not just investing into evolution of your business intelligence tools as Tony points out; remember that the priority now is meeting the needs of the business and especially the analysts who are held responsible for analytics. Just serving reports and dashboards faster and in prettier form in many cases is not going to provide the critical insights. As I have already pointed out, putting more charts in dashboards or embedding key performance indicators are not smart strategies for guiding business improvement.
There is work to do here in convincing skeptics of the need for investment. According to our business technology innovation research building the business case (43%) is the second-greatest t barrier to adopting innovative technology like discovery, following lack of resources (51%). It is not easy to use a traditional business case that projects value and benefits from a specific investment in the case of discovery, where the technology is about finding the unknown and providing benefits depends on the competency and skill of your analysts and business professionals as well as the effectiveness of the technology. Because usability is critical (the evaluation category criteria most often classified as important by 64% of organizations in our research) you must look beyond just the capabilities in the technology to how easily a variety of roles can use it and collaborate on the insights found from it. Most tools are only suitable for certain roles and maybe not for every analyst, let alone directors or vice presidents. These innovations by many technology vendors are just coming to market. Many organizations decide the potential value of an investment on the time to utilization of the tool and the projected time to value on potential insights from it; for analytics through discovery that is not always easy today.
To evaluate the benefits of the new generation of business analytics that utilize the discovery methods discussed here, start with a conversation internally to identify your needs and look at some online demos to see what is possible. Our next-generation business intelligence research finds discovery to be a critical consideration for 69 percent of organizations, which indicates the strong interest in discovery technology that will work for business and IT. From another perspective, our spreadsheets in the enterprise research finds that 74 percent of organizations are still using personal spreadsheets with BI to meet many of their data and visual discovery needs, at the same time as 56 percent find it time-consuming to combine spreadsheets through copying and pasting and more than one-third (35%) find data errors from their use of spreadsheets. As my colleague Robert Kugel puts it, it is time to end denial about the use of spreadsheets and focus on buying tools designed for complex tasks like discovery. For IT organizations, it is essential to address data and integration needs for analytics and analysts so business is not spending so much time on data-related tasks; they also should understand that while there will be more tools and vendors in use, the architecture and support of them should be simplified. For business, it is critical to understand what the types of discovery are all about and where technology innovation in 2013 can help make your organization’s processes run better and faster. It will be well worth your time to investigate why more organizations – and maybe your competitors among them – are getting benefits from making their business to smarter by using discovery.
CEO & Chief Research Officer