Hi all, I have been receiving many questions on what does it mean to be "self-service". What is the self-service data discovery message about? How is Sense Self-Service Data Discovery if they need to know script, set analysis etc. Does self-service data discovery mean the users don't need data modeling or scripting skills, etc. Here is my take on this. Always happy to have your input to spark greater discussion.
I call this:
PARTS of the SELF-SERVICE DATA DISCOVERY MESSAGE
(Keep in mind - it depends on what the audience, roles and responsibilities are of both the organization and its people. A department, an enterprise, a partner, a OEM - all will approach these differently. But in then end - the same fantastic results will emerge with Qlik Sense driving the analysis.) Qlik Sense fits in the way YOU do business, not the other way around.
From the perspective of a non-technical business users with basic knowledge of accessing data sources through simple connections to local databases and column selection - or even through the use of spreadsheet and flat files .. data modeling skills are really not necessary. These people want quick answers to business questions and possibly want to share their findings. These individuals are great candidates for Qlik Sense Desktop. These are the same people that may be using spreadsheet tools or other local desktop data visualization tools. There are many questions that can be answered using the product as it is without the need for data modeling skills, set analysis, scripting etc - in this manner. Now, perhaps there is more data needed to make a decisions and it is not available locally, perhaps it is available in a well defined DW, or another external source - setup with appropriate access and security. Qlik Sense Desktop can connect to these sources and bring that data into the discovery process for analysis. No data modeling needed. This is 1 part of the self-service data discovery message - FOR this type of audience.
Now - perhaps some more complex logic or calculation is needed for some dimensions and measures. 1st - are these already defined somewhere and stored and aggregated in a DW? If not - would they need to be defined at the source for everyone (DW) or just in the BI application layer? This is where having some skills with the expression syntax may be necessary - which in this case, is no different than that same person using Excel formulas or any other tool that offers the ability to define calculations. However if more complex logic is needed, the appropriate individual with those skills, can define these in Sense, so having these skills is helpful. But, once created, these calculations can be defined in the library of the Qlik Sense app and reused to create new analysis for those who do not have these skills. It is also possible to provide a templated Qlik Sense app - with the data model and library populated with measures, dimensions and even common visualizations to be reused. SO this is another part of the self-service data discovery message.
Now - let's take this to a broader level. As a business user my local discoveries and collaboration have determined that this is something worthy to share and possibly can be used to derive new KPIs for my business and perhaps even make it available as an application for the enterprise. However I need to promote this to the Enterprise. In order to do so, the appropriate people, data, aggregations, transformations and expressions, etc. needed to be folded into the equation. After planning and discussions take place at the Enterprise level (assumed) - the people WITH data modeling skills, scripting skills, etc. could then be brought in to centralize and optimize this for the Qlik Sense deployment. This of course would now be deployed to Qlik Sense Enterprise Server, which now has a layer of monitoring, governance, security and management controlled by IT or the appropriate BI staff. The office has also decided to make my app available as a centralized template (similar to that of a semantic layer) - now with nothing more than a web browser on a machine or mobile device you can create new discoveries, stories, analyses for your own personal use - or decide to publish them to a common area for all to use or share (providing appropriate security) - all controlled and centralized under the Qlik Sense environment. This is another part of the self-service data discovery message.
So in summary - the self-service data discovery message has many different parts that can involve many different people and skill sets (if needed), it really depends on the business requirements and needs of the organization. Qlik Sense provides you with many choices on how it should be used and deployed within an environment. Sense fits into the way YOU do business. Locally, Departmental or Enterprise, skill sets of all kinds are needed with any BI solution, its up to the individuals and business to identify what is the best fit for the solution.
Re: Parts of the Self-Service Data Discovery Message
I am keen on Self Service and believe [hope?] that Qlik Sense will be a useful tool for facilitating this.
I have so far been a lurker re Qlik Sense, merely reading the views, questions and answers of others. Personally I shall hold back from doing anything decisive until the Qlik Sense Enterprise Server is released and the licensing model formally announced. I have seen people badly burnt by other vendors' "free" offers becoming painfully expensive for any practical usage within an enterprise.
The current freebie Qlik Sense Desktop limitations that include only allowing installation on desktops os's and not server os's, along with the hardcoded expiration of the current release at the end of the year pose unknown risks that I do not believe to be acceptable for me to take. Hopefully when the licensing model is announced then risk analysis and internal costing will be possible and I am optimistic that risks / costs will enable viable Self Service BI, with Qlik Sense being one of the tools involved in this.
For me the crux of Self Service BI is Central IT deciding to lose control by encouraging the great unwashed to use a tool such as Qlik Sense. However losing control needs to be done at a suitable time and in a suitable manner. It will result in the desire for a single version of the truth being compromised, but as long as one accepts this then great benefits could be gained.
Einstein surmised that observations are relative to the conditions they are observed from. So in any organisation the observations of say Sales, Finance and Operations are most likely to differ. The grail here is to embrace and accept different viewpoints, whilst maintaining a single version of the truth for use in formal outputs such as Annual Reports / Audited Accounts.
Re: Parts of the Self-Service Data Discovery Message
Bill - thanks for your contribution - it is input like this that is very valuable to us. I encourage you to post again to give us an update on your experience once you have learned more about Qlik Sense and see how it will / will not work for you.
Just to be clear, in regards to IT voluntarily 'losing control' - as I have interpreted by your post - that is a decision that IT can only make and is not a direct result of the product itself. In regards to self-service, I guess, 'Losing control' is a strong word, I would prefer to address this as providing less restriction, more flexibility, while still maintaining a level of governance and consistency. For example. Sense Enterprise includes centralized and comprehensive libraries, security, monitoring, auditing and governance that can conform to the many different ways organizations do business, This in turn lets the users explore the data they need to make decisions, while giving IT confidence that the information they are drawing upon is consistent.
Thanks again, I look forward to learning more about your experiences with Sense.