QlikView is a BI platform based on in-memory associative technology. It has enjoyed some success with customers looking for affordable rapidly deployed BI solution with an attractive user interface. While QlikView’s approach is novel, it is ultimately limiting. QlikView prevents you from enjoying the decades of rich technical development put into true business intelligence tools already, or for that matter, taking quick advantage of new business intelligence developments.
When is associative not associative?
The limitations of QlikView are numerous. QlikView uses an in-memory data storage, storing data in a tabular format. In order to overcome the limitations of traditional on premise business intelligence from such vendors as Cognos, Business Objects and Microstrategy, QlikView stores data in memory in a compressed form with associations defined between data items rather than joins as used in traditional databases. These associations are derived automatically by QlikView during the data load process based on matching column names across tables. Although QlikView touts its “unique patented in-memory associative technology,” the underlying data structures are not associative at all but rather the data is stored in a regular tabular format.
Sacrificing power for speed – and dropping best practices along the way
QlikView’s associative data model is designed for deployment speed and for providing users with an environment that is designed to encourage data exploration. In comparison, the products of leading BI vendors like Oracle, Microstrategy and Business Objects use logical dimensional models that are core to BI best practices that have been proven out over the last 10 years. The reason that QlikView abandons this approach is because traditionally logical modeling requires much more time to set up and is technically complex. However, when done properly, it allows more analysis flexibility and power than associative data models. The Birst approach is to combine the benefits of logical modeling with the quick time to value realized in a SaaS environment.
Scalability and Performance Limitations
This means that QlikView Applications are constrained by how much RAM can be addressed in a single hardware box. As a result, QlikView is designed for gigabytes-scale BI applications. Given a typical 10x compression rate, QlikView performs well on data volumes up to 10-20 gigabytes of compressed data. Since QlikView has to bulk load all data to the system RAM, reloading data is a considerable workload and can take a substantial amount of time. This limits the practical frequency of data updates in case of larger data volumes.
QlikView often grows inside of a larger organization via small and isolated departmental deployments. At some point in time, corporate IT will step in and seek to take control. Consolidating and managing separate QlikView deployments is very difficult and requires considerable rework.
Lastly, QlikView stores its data in proprietary files. If the need should arise, the proprietary storage format will complicate the migration of data and reports from QlikView to other BI products.
No OLAP-style analysis
QlikView does not offer the richness of MDX and other logical query languages provided by traditional on premise BI products. These languages enable end users to create far more powerful calculations (e.g. via addressing cells positionally in either a materialized or logical cube) than are possible in QlikView. QlikView provides a user interface for users to interactively explore and analyze information based on associations between different data sources. This is a unique approach but because it is not a logical query language, it has limitations.
No Pixel-perfect report writing
QlikView lacks some of the advanced capabilities required for creating highly formatted reports. QlikView is designed for interactive analysis and not for report writing. To create a formatted report in QlikView requires the use of macros and the duplication and maintenance of QlikView objects. Previously created QlikView objects need to be cloned and placed into a “Report Section” in QlikView.
QlikView takes shortcuts that can result in more work
QlikView makes no distinction between columns that are facts and columns that are attributes. While this is designed to encourage data exploration, it makes the creation of well-defined formatted reports more complicated and time consuming.
No real-time analysis and reporting
The latest release of QlikView provides a feature that allows organizations to run a script for continually updating the in-memory data of a QlikView deployment. Previously, this was not possible but even with this feature, only scenarios where updates to the source system are infrequent and limited in volume would benefit from this. For fully scalable real-time analysis and reporting purposes, a BI platform needs to be able to query the source system or database directly and on demand.
Limited Data preparation
While QlikView provides a scripting language for integrating and loading data from multiple sources into memory, it does not provide advanced ETL capabilities that are needed for more complicated data integration use cases, like parsing of web log data for sequence analysis and hierarchy tree navigation.
Higher Total cost of ownership than SaaS BI
Compared to SaaS BI solutions, the cost of QlikView is higher because of the procurement and maintenance of the hardware needed for development, QA and production environments, and costs associated with downtimes. Only when QlikView is compared to on premise BI products does QlikView have a lower TCO.
About Birst
Birst™ is the leading provider of on-demand business intelligence solutions. Birst brings the benefits of fact-based decision-making to a much broader audience by making it affordable, fast, and easy to use. Birst is designed to support users of all sizes -- from individuals to groups and entire companies, so that everyone can benefit from greater insight into their business.
For more information on Birst and the superior benefits of SaaS BI, click here.
