QlikView 11 Advances Collaboration and Changes Course on Mobility

QlikTech recently introduced QlikView 11, the latest version of its business intelligence (BI) software, which emphasizes new collaboration features as well as enhancements to its user interface. In an about-face, though, in its approach to mobile access, the company has moved away from its native iPad application to a browser-based app using HTML5 technology.

It has been a year since QlikView 10 was launched and I assessed it. During that time, we’ve seen the popularity of social media drive a rising interest in collaborative BI. But QlikTech is not yet explicitly linking the two. While it has announced integration with Chatter, the social media tool from Salesforce.com, and is demonstrating those capabilities, they are not part of QlikView 11. The collaborative features of this release include new interpretations of well-known features such as annotations and threaded discussions. Version 11 also adds a unique collaborative BI capability to share the QlikView environment with others, much as you might share your desktop in a WebEx session. In fact, as part of my briefing about the new release, I was given a link to see the new version and interact with it even though I had no installation or QlikView license of any kind on my computer. One way in which QlikTech has reinterpreted annotations is to allow the user to save a snapshot of the system at the time the note was taken. In this way readers of the note can understand the context in which the note was written.

QlikView 11 also includes expanded comparative capabilities. Beginning with QlikView 10, users could create custom groupings of data to include in the display, but they were static. The comparative analysis capabilities of this release enable users to continue to navigate and perform selections within the groups. For example, a custom grouping to compare the northeast region with the southeast region can be further refined to see how individual product sales compare within those groups. The product also has some easy-to-understand visualizations of how the groups and their sub-selections relate to each other and to the larger group from which they were selected.

The company has added more application development capabilities in this release as well. QlikView 11 exposes application objects in a more granular fashion, making it easier to control the display and interaction of objects on the screen. The new version also supports third-party objects more fully, allowing them to execute custom code which the previous version could not do. QlikTech has also partnered with data integration vendor Informatica to build QlikView data files using Informatica’s products. This integration should help speed up the data preparation process, which our analytics research shows is where people spend much their time.

As QlikView has gained more enterprise customers, the company has been adding features to appeal to the IT organizations that must support its deployments. In this release, the product reloads faster. It has more load-balancing options, and the management consoles of the previous versions have been consolidated. More of the administrative APIs are available so these tasks can be automated. Security has been enhanced with more granularity. Now the auditing feature does not have to be applied at a system level but can be applied to individual applications, which eliminates some of the overhead associated with auditing.

While QlikTech continues to add interesting features, there are some omissions from its product. For example, it has no specific features to support Hadoop for large-scale data. In fact, QlikView’s memory-based architecture will make it challenging for the company to support Hadoop as a data source, and our recent Hadoop benchmark research shows that many enterprises are adopting Hadoop. A QlikView Community poll also shows interest in Hadoop as a data source. Predictive analytics is another area lacking in the QlikView product line. (Discussions in the QlikView Community suggest only a moderate level of interest in adding these capabilities.) Our analytics research indicates that 37% of the 2,600 organizations studied said predictive analytics are important although only 13% have deployed such capabilities so far. We are currently conducting research specifically on predictive analytics to understand this dichotomy better.

With respect to mobile capabilities, one of our five key technology trends for BI, QlikView has opted for a cross-platform approach rather than continuing to provide or expand its native applications. While some BI vendors such as Information Builders are adopting a similar strategy, others such as IBM, which recently introduced a native iPad application, are going the opposite direction. The key advantage to enterprise users in QlikTech’s choice of HTML5 is cross-platform support, which enables organizations to support a wider variety of devices. However, as I worked with the HTML5 version from the demo site, I found the browser-based version much less appealing than a native application. For example, my stretch gestures were often interpreted as selections. The browser-based drop-downs were harder to work with than native application controls, and the need to dedicate a portion of the screen to non-BI browser functionality (such as entering URLs) leaves less space for BI-specific functionality.

QlikTech deserves credit for continuing to enhance its product’s ease of use and providing new collaborative capabilities. Enterprise IT should welcome the new administrative and security enhancements. It still has more work to do on the big-data and predictive analytics fronts, which are not addressed in this release. And it will be up to users to judge whether QlikTech has taken a step backward on the mobile front.

Regards,

David Menninger – VP & Research Director

Qlik Here for My View of QlikView 10

My colleague recently wrote about QlikView, noting its rapid ascent to providing a very robust support of mobile technology platforms among BI vendors and integration with SAP. On the occasion of its release of a major product revision, QlikView 10, I’d like to add my perspective on the company and its most recent release. I first learned of QlikView about five years ago while working on the TM1 product line which, like QlikView, is also a 64-bit, in-memory analytic technology supporting business intelligence needs across business and IT.

Since July 16, 2010, QlikView has been a publicly traded company (NASDAQ: QLIK) so you can review its audited financial reports and see its growth to being a $100 million company. And because it is focused on analytics and BI, the results are not buried in an entire portfolio of other products and services where it is difficult to determine the financial and customer health of the business.

QlikView emphasizes ease of use, and our latest benchmark research in business intelligence found that to be the number-one consideration in evaluating a vendor and product, especially on the business side. QlikView 10 builds on this attribute. For example, it has expanded search capabilities. As you search for items to pick from a list or display in a report, you can view all the data, not just the attribute(s) displayed in the list box. For example, suppose you are looking for the new customer deal that just closed last week, but you can’t recall the customer’s name. What you do remember is the sales rep who closed the deal. In QlikView 10 you can search by the sales rep’s name in the same way you would do a literal string-search in most other products. Just type what you know (sales rep name, product sold, region or any other attribute) and the relevant matching items appear in the list. However, the search capability focuses on structured data or attributes, while our research finds that integrated analysis of structured and unstructured content is a growing demand and the top usage capability that organizations plan to deploy and desire to have.

Enhancements to the user interface in QlikView 10 include two new objects: Mekko charts and a shared display area. Mekko charts are a nice addition because the area of the bar quickly communicates its relative importance. The shared display area is similar to a “tabbed” workspace. It allows end users to switch between several objects easily and quickly, repainting just the portion of the screen that needs to be refreshed.

Customers can now create their own visualizations. These custom objects have all the attributes and properties of built-in QlikView objects and can be shared among users within an organization. QlikView also has created and published an API for high-performance data integration when users don’t have ODBC access to data. Version 10 also provides new APIs for management tasks such as adding users, which can be done programmatically from outside of QlikView.

In the area of performance, reloading data is now multithreaded, so the speed of reloads should improve by approximately the number of CPU cores in your server. In addition, the asynchronous nature of QlikView has been extended to its Web version, so while the total elapsed time to complete a task won’t change, the first display will be available as soon as it is ready. Since control returns to the user, you can cancel the rest of the task or navigate to a different display if you have already seen what you need to see. Other performance enhancements take advantage of machines with more than 32 cores.

Version 10 has manageability improvements that include metadata tagging, centralized user management and enhanced audit logging. All these features will help QlikView win over IT departments in enterprise deals.

In my view this release has important usability, manageability and scalability enhancements, but I would have expected more in a major release since it has been almost a year and a half since Version 9. Again it’s noteworthy that QlikView has invested heavily in native support for mobility across popular platforms including Apple, RIM and Android that is no easy task and larger providers are just beginning to investigate and develop. But perhaps the recent IPO and the velocity of QlikView’s growth have kept it from making larger, more visible advances. And I’m sure the threading improvements related to reloading data and producing performance advancements, while not very visible, were a particularly thorny development issue.

There are a couple of things I’d like to see in future versions, not only from QlikView but most of the other BI providers as well:
– A better mechanism for sharing QlikView objects externally. It seems there are limitations that make it difficult to do this outside an organization. There could be a thriving marketplace of third-party objects and extensions to QlikView.

– A massively parallel processing (MPP) version. While QlikView scales across servers to handle more users, it does not have MPP capabilities, which could limit scalability for very large user deployments.

All in all, QlikView 10 offers some incremental improvements in important areas such as ease of use and performance. The new version is available immediately so  you can download and try it yourself and get access to documentation and videos of demos and how to use it. In the future, it would be nice if you didn’t even need to download it – maybe the next version we can simply run in the cloud. Now QlikView has to continue its growth, expand its presence in the lines of business and use its new release as an opportunity to grow its market share.

Let me know your thoughts or come and collaborate with me on  Facebook, LinkedIn and  Twitter .

Regards,

David Menninger – VP & Research Director