Three Misconceptions about Self-Service BI

By July 15, 2015 Articles No Comments
douglas-reynholm
Learn about some of the common misconceptions about self-service BI and the future of conventional BI in the light of self-service BI. That is, with the introduction and proliferation of self-service BI, what happens to conventional BI, which some are also calling traditional or classic BI.

Conventional BI

To be clear, by conventional BI, I mean data warehouse for central data repository, ETL, i.e. to extract, transform, load data in batch mode from multiple data sources, and data presentation, visualization from relational database and multidimensional cubes.

The Sales Call

douglas-reynholm

What prompted me to write about this was a recent sales conference call for a custom financial BI solution. With operations in Europe and Americas, the prospect needed multi-company consolidation from multiple ERP systems and multi-currency reporting. Complex, but doable. The man with a commanding voice said, “We have been working on and enhancing our business intelligence solution for the last few years. It’s working generally well for our operations and sales guys. But it’s all traditional BI.” He almost sounded apologetic as if he’s been leading his corporate IT back into the dark ages. He continued, “We are now looking into deploying more cutting-edge technology and tools. So, instead of traditional BI, for our finance guys we want to utilize self-service BI.”

1. Self-Service doesn’t mean Inexpensive or Limitless

First, self-service or do-it-yourself doesn’t mean inexpensive or limitless. Self-service BI relies heavily on in-memory data processing and analysis. Memory is expensive. Storage space is not.  Also, in-memory data processing and analysis capabilities based on available memory can be limited, and may not yield optimal performance. Moreover, it doesn’t just happen. Somebody’s still gotta do it. You’d still need a talented (read: hard-to-find, hard-to-retain and well-paid) employee who can not only grasp and articulate analysis requirements, but also clearly understands basic technical concepts such as data sources, entity relationships, data types and so on.

2. Self-Service BI doesn’t replace Conventional BI

Second, self-service BI doesn’t replace conventional BI; it augments it. It’s “cool” because it enhances your intelligence gathering capability and data visualization experience. But you’re not really choosing between conventional BI and self-service BI, as an overall corporate strategy. You’re going to leverage them both. And you are going to a apply the right solution depending on your specific BI requirements (data gathering, presentation, acceptable latency, etc.) and don’t forget, the budget.

3. Don’t do it just because “everyone” is talking about it

Finally, don’t “do” self-service BI simply because everyone is talking about it, and therefore must be cool. Understand what it is, its promise as well as its perils, before incorporating it in your corporate BI strategy. (More here: pitfalls of self-service BI and how to avoid them.) Also, evaluate what type of solution would meet your needs for data consolidation, analysis and collaboration.

Always Be Closing…

… except this time

maurice-moss

I told him.

I cleared my throat, and tried to explain this to the man with a commanding voice on the other side of the line. “To serve your finance team better,” I said, “you’d rather build the foundation with conventional BI. It makes more sense for two reasons, one: your IT team already has conventional BI skills, and two: self-service BI will not be appropriate to handle the complexity of your data collection, analysis and collaboration needs. With conventional BI, you would import and consolidate data from general ledger, budget, forecast etc. in your existing data warehouse; then deliver financial statements and dashboards depicting financial ratios. With this foundation built, you can provide users with self-service capabilities for ad-hoc analysis and data visualization.”

As I said, I tried to explain. What failed to convince him was his misled convictions about self-service BI and conventional BI. I’m pretty sure, it wasn’t my high-pitched voice.

Conventional BI is here to stay

TV didn’t kill radio. Web streaming didn’t kill TV. Sure, there were some permanent changes along the way, such as the demise of big box stores. (Remember Blockbuster and Tower Records stores?) But today, we have more options with devices (phones, tablets, laptops) as well with delivery (on-demand podcasts) for enjoying our favorite radio and TV programs. And, broadcasters also have more options with making revenue (web advertising).

As self-service BI gains more ground, appliances needed for conventional BI will improve in efficiency and affordability in the years to come. This is evident in the upcoming release of SQL Server 2016. After primarily focusing on inclusion of self-service BI features in SQL Server 2012 and 2014 releases, Microsoft has given attention back to conventional BI components, while simultaneously improving the self-service BI features. (More here: Microsoft breathes life back into SQL Server 2016 BI On-Premise) Further, data storage and analysis delivery will offer more options and flexibility. Cloud-based data storage and mobile-based analysis delivery are reality today and getting more flexible, secure and affordable. (More here: Why the cloud is key to on-premise SQL Server 2016 and here: Microsoft acquires mobile business intelligence leader Datazen)

Conventional BI is here to stay, and it will only get more powerful and flexible, making Business Intelligence, in general, more pervasive than it is today.

PS: Note to self

Add a reminder on the calendar to call back the prospect in a couple of weeks.


Ashu Shendé is the Managing Partner at IQubz, a BI consulting firm based in Southern California. Contact Ashu via LinkedIn or by writing to ashu at iqubz dot com.