Jason Long and John Trigg from Qliktech suggested to me and to any person who cares about the future of Qlikview to place their new ideas for Qlikview rather in Ideas area of QlikCommunity and I took a 6 of my ideas which I originally suggested on LinkedIn to where Qliktech wants me to place it (and yes I have more ideas then Qliktech can handle, but I hate to waste my time and be rejected, see about the rejection below).
Qliktech approved 4 of them, they are active now and collected more then 179+ votes and 28+ comments already and still open for voting!
Qliktech rejected my 5th Idea (Free Qlikview Public Cloud Service with limited data sources) on the ground it is rather marketing then technical idea and because mentioning competitors on Qlikcommunity forum is not appropriate. Qliktech also rejected my 6th idea (Qlikview as Data Collector, especially Qlikview Personal Edition as data collector into QVD and QVX files), I disagree with Qliktech: these are the best ideas and rejection of them is a strategic mistake. I think Qliktech will pay price for mistakes, but everybody makes mistakes, so it is expected.
Qlikview is a good foundation for Visual Analytics, but does not have enough an analytical power. Competitors like Spotfire (and soon Omniscope) have an excellent integration with analytical tools, like R library and R scripts; Spotfire even has own commercial version of R called S-Plus. Qlikview should have the same or better functionality. FYI, R community has more then million of active users.
If R is too “open” for Qliktech, there are plenty of commercial alternatives to it. I know that many Qliktech Partners have need for extra analytical functionality integrated with Qlikview.
2. Virtual Memory for Qlikview: an ability to expand the database beyond RAM
Qliktech Competitors (e.g. Spotfire and Tableau) have similar to Qlikview in-memory columnar databases/datastores, but their Data Engines are capable to allocate data outside of RAM, e.g. on disk and therefore they are able to handle much larger data sets and scale (theoretically, at least) unlimitely. In the era of “big data”, cheap fast disks, very fast (and relatively cheap) SSD and soon to be available PCM (Phase-Change Memory), Qlikview should have a full support for virtual memory and an ability to take advantage of all this storage options.
3. Qlikview has to be able to read Microsoft DataCubes!
With upcoming new version of SQL Server 2011 Microsoft has at least 3 of the most sophisticated and useful types of OLAP Cubes on market: traditional SSAS (SQL Server Analysis Services) Cubes in SSAS Multidimensional Mode, “new” (Vertipaq) Cubes in SSAS Tabular Mode (no need for Sharepoint anymore!) and “local” PowerPivot-based Cubes. Obviously Excel/PowerPivot can access all 3 types of Cubes.
One of reason why Tableau is growing faster then anybody (including Qliktech) and probably will outsell Spotfire in 2011 is because Tableau can access all types of Microsoft Datacubes. If Qliktech wants to keep its leading positon in Data Visualization market, it has to be able to natively read and integrate with all 3 types of Microsoft Datacubes. Currently Qliktech and Qliktech competitors (except Tableau and Omniscope) cannot do that.
4. Server-less distribution of Qlikview applications with FREE Desktop Qlikview Reader
For many years Qliktech enjoyed the business model based on a requirement that customer with multiple users (unless customer wants to pay for Qlikview Desktop Pro for all users) needs to buy a Qlikview Server. This has to change, if Qliktech considers Tableau Software as a company who can lure customers from Qlikview. My estimate that in 2011 Tableau sales will reach 20-25% of Qlikview sales and if trends continues, it will be 40% in 2012, 50% in 2013 etc.
One of reasons for Tableau’s success (it has more than 100% YoY) is that they have FREE Desktop Tableau Reader, which allow to distribute Tableau applications/visualizations as files without any involvement of Tableau Server. The same model successfully used by Visokio for their good Omniscope product. Spotfire has similar distribution approach because everybody can have a free Spotfire Silver account (which comes with limited Spotfire Desktop) or use Spotfire Enterprise Player.
It is actually easy to prevent further Tableau expansion, if Qliktech finally will allow as an option for users to have and use FREE Desktop Qlikview Reader. Such a license will be very different for Qlikview Personal Edition (which does not allow any distribution of QVW files): Desktop Qlikview Reader will not allow create any Qlikview applications and QVW and even QVD files but it will allow to open and interact with any QVW files, created (by somebody else) by Qlikview Desktop/Developer tool and posted on any file server or website.
5. Free Qlikview Public Cloud Service (with limited data sources)
Qliktech rejected it on the ground it is rather marketing then technical idea and because the mentioning of competitors on QlikCommunity forum is not appropriate (both reasons made me smile and laugh).
I know only 3 ways to improve and learn from: smart (from own mistakes), smarter (from somebody’s else mistakes) and Smartest (from somebody’s success). Qliktech refused to be Smartest or even Smart by rejecting this obvious idea. Tableau, Spotfire and Microstrategy agreed with me and they look smarter to me in this regard.
Qlikview has a successful competitor – Tableau, who is 5 times smaller but growing faster (I thought it is not possible) than Qliktech (in fact at least twice faster with more than 100% YoY)! One of examples of Tableau tremendous successes is a free Tableau Public Service, which quickly helped Tableau to establish and promote its own brand and products. Tableau Public is extremely popular, getting million+ new users every months and since inception it got 20+ millions of users! This success forced Spotfire to offer Spotfire Silver for free! WSJ and many newspapers and millions of bloggers using Tableau Public and indirectly promoting their brand. Spotfire Silver 2.0 does that too, may be even better then Tableau Public. Microstrategy just release Public Cloud (personal edition) for free.
The Smartest way for Qliktech to respond is to create a similar or better public cloud service, something called like Qlikview Cloud or whatever name marketing wizards from Qliktech will come up with. Of course, that service should restrict the size of involved datasets and the variety of datasources. My personal suggestion is to take a look on http://www.qvsource.com/ in addition to standard sources like Excel and flat files and to QVW files of certain origin (e.g. published on QlikCommunity or Certified by Qliktech for Qlikview Public Cloud – similar to some certification Qliktech is already doing for some QVW files generated by Qlikview Personal Edition).
Qliktech rejected it and claiming that this is already offered – sounds like they do not understand own product then.
I am actually asking for this functionality since 2008. Majority of Qlikview users like Qlikview so much, so they do not wish to use other tools even in case if they need Data Collecting and Data Entry Forms. I personally found the way to integrated http://www.spreadsheetweb.com/ as my Data Collector for Qlikview, but it must be done at Qliktech internally.
Most users want to submit data to existing table in existing datasource or database directly from specialized QlikView Sheet Object. QlikView actually has a few features like Input Box (allow to enter data into variables) or Input fields (allows through usage of SQL Insert/update/delete statements to put data into table) but none of them allows you directly connect to data source, choose table and construct good looking Data Entry form with all data validations and enter their new records to designated table though such Data Collector. I suggest That QlikTech will add this new Sheet Object.
I can suggest to consider it from other, different angle too. You know that you can fillout your taxes by filling-out the free 1040 PDF-Form using free Acrobat Reader. The same thing can be done by using free Qlikview Personal Edition to fillout the pre-defined or automatically generated by special Wizard free QVW-Forms which will collect data into QVD (or alternatively QVX files, QVX format was released with Qlikview 10 with intention to replace the QVD format, see description here: http://community.qlikview.com/servlet/JiveServlet/downloadBody/2297-102-1-2325/QlikView%20QVX%20File%20Format.pdf) files, which cannot be used as Qlikview application but only as data transporting/collecting device/file. Appropriate and licensed QVW application will read it later (e.g. using partial data re-read) and enable server-less (it can be done with Qlikview Publisher in picture too) distribution of fresh data if needed!
I used QVD files a lot, since reading from them is much faster and than can be used for Data Replication: backend application reading from data sources to QVD files and front-end application (published through QlikView Server), reading from QVDs (or QVXs). Neither QlikView Documentation, nor Qliktech employees in a mood to talk about internal structure of QVD files. It will be nice to have an utility (CLI or UI), which can read from all possible datasources (remote or local), using ODBC or any other data access technologies and store all data into QVD/QVX files. Than these files can be shipped to QlikView applications for data visualization, but backend Data Collection and Replication jobs can be done without QlikView installed. I am wondering if anyone can describe QVD structure and may be instruct me and QlikCommunity of how to create, write, read and modify QVD files without QlikView?
FYI, nice people at http://www.vizubi.com/ already know how to read and write QVD files without Qlikview involved, so why Qliktech itself cannot open QVD format and provide some nice QVD itilities? And finally: what preventing Qklitech to create Excel 2010 Add-in, which will enable Excel read and write into QVD/QVX and that include ability to directly connect PowerPivot Tables and Local cubes with QVD/QVX. That will make Excel a very friendly Data Collecting tool for Qlikview…
Can’t agree with you more on R integration. I don’t understand why they’re behind the 8-ball on this. There was some discusion of specific use cases by John Trigg, to which I replied in that post. Since then, nothing has gone further. This is disappointing as it doesn’t, in my view, further the power of QV.
Jaspersoft was smart enough to partner with RevolutionAnalytics to provide a solution for QV integration, but this comes at a price (RevoDeployR, that is). The same software package seems to work with QV, but I can’t even say if this is an end-solution (basically the graphics generated are png files – but someone clever might add some functionality to a QV dashboard which can make this more interactive and useful.
Someone posted an attempt at a solution to integrate R with QV using statconn DCOM software. While this is much like how RExcel was born, it is a rather clunky and code intensive solution in my opinion, but doesn’t run the capital cost of a server class R installation.
So 6 of one, 1/2 a dozen of the other. Maybe QV will see the light and begin co-development of smartly integrated pathways from 3rd party statistical software, particularly R (the other high $$$ ‘big guns’ already integrate embedded dashboard output functionality – while you can build this within R (using RHTML), too – folks may not be too happy with dev time – just makes sense to simplify reporting tasks).
Correction to the above… Jaspersoft did not partner with REvolutionAnalytics to provide a solution for QV, only for Jaspersoft. However REvolutionAnalytics RevoDeployR package allows QV to call a remote R server instance so that QV can render R output.
Sorry about the misprint – I was emotionally engaged in the need to endorse R integration with QV haha.
FYI, I’ve developed two reusable Java libraries to create and process QVX and QVD files independent from QlikView. The QVX lib was already integrated into Pentaho Kettle and Lavastorm Analytics. Feel free to contact me if you have a use case for it.
Great ideas! Qlikview actually has several ways for R integration – including Revolution Analytics R and MadLib, Alpine (Pivotal) – see bigdataconsumerkit.com – that uses Intel Hadoop and/or the Pivotal Integration for bigdatavirtualkit.com. There is also an integration to use SAS Analytical Base Tables…