With the connected world in thrall of big data and analytics, why are we talking about a decades-old technique called Business Intelligence? Aren’t they two ways — from different generations — of saying essentially the same thing?

Actually, no. While both involve application of statistical analysis to collected data, they work in different worlds and have different objectives:

  • Big Data Analytics or BDA as the technique is predominantly used today, looks at stored data from customer transactions and site visitors, applying analytical techniques to divine how best to preserve and grow one’s market, essentially looking outward
  • Business Intelligence or BI, around since the 1980s, mainly looks inward at the data created by existing internal processes — the only data available when it was developed — with the objective of improving internal efficiency. Firms like Cognos, now part of IBM, and Crystal Reports, now a SAP product, have long been active in the BI space

These differences are more substantial than they might appear. For example, because big data is collected from external transactions that cannot be rigorously audited, it is by definition the trusted source for all subsequent analysis. Business Intelligence on the other hand, uses business data created internally that if not collected properly or has gaps, will produce answers that cannot and should not be trusted.

Why Has BI Resurfaced?

Reliable studies indicate that as much as 80 percent of all BI projects fail. So why the continued interest?

The answer can be traced to a new wrinkle in the BI world: Self Service Business Intelligence (SSBI). The idea, we are told, is that operational users need not depend on the IT department’s enterprise level BI architecture, expertise and tools, but can instead acquire their own BI tools and use them to develop the answers they need. Cynics might restate this as: “the software community, struggling — many without much success in recent years — to sell high priced Business Intelligence systems to IT groups, decided to 'jump the shark' and go directly to the user … and the user’s budget, to sell their wares."

Even Microsoft is touting the SSBI capabilities built into MS Office and SQLServer, with Oracle, SAS and others on the same “self service” warpath.

Is Self Service Business Intelligence For You?

But if you’re an operational manager, can you wade into this new landscape and survive? The literature is all over the map so unless you’re a real bookworm, it may not be of much use.

My own penchant is to use a process we know something about as an analog to the SSBI adoption process: let’s use building.

You wouldn’t decide to build a structure yourself unless you knew what you were doing … and so it is with SSBI. Unless you, or someone on your staff, knows a good bit about business process analysis, you are probably out of your depth, regardless of what the software salesman tells you.

You wouldn’t decide to build a structure yourself without access to appropriate raw materials … and so it is with SSBI. If the data maintained in your corporate systems isn’t accurate, complete and organized for efficient access, you probably won’t be successful at getting answers to your questions no matter what kind of software you buy, and because you don’t own that data or the processes used to create it, you probably can’t fix it yourself.

You wouldn’t decide to build a structure yourself unless you could afford the costs in time and materials … and so it is with SSBI. The term Self Service can be misleading: it isn’t cheap no matter how you approach it, and if you don’t have or can’t get the budget to do it right, you shouldn’t start. That’s why the building trades use highly detailed estimates before they start, and so should you lest you end up half-done and out of funding.

You wouldn’t decide to build a structure yourself if the people around you — like zoning boards, planning commissions, building inspectors, etc. — are opposed or indifferent to your efforts … and so it is with SSBI. Deciding to do BI yourself is likely to run afoul of other parts of your organization, most likely the IT group that has run BI (or delayed getting it) up to now, and “self service” doesn’t mean “self reliant.” If it isn’t at least partially a team effort, it’s going to be a real challenge if possible at all.

You wouldn’t decide to build a structure yourself unless you knew what you wanted it to do for you … ditto for SSBI. Unless you have a pretty good idea what areas of your operations aren’t performing at full speed, you will have a difficult time structuring the BI analysis processes to answer the questions that will help you improve things. At best, you will spend time and money following dead-end roads before you happen onto the answers you need.

Know What You’re After or You Won’t Find It

The goal of this analogical exercise is simple: if you want to consider SSBI for your organization, approach it like you run your business. Ask the questions that will illuminate your goals, costs, requirements and challenges, and make sure you understand the answers before you jump. You may find that merely asking the questions will expose solutions to some of your problems. In any event, document your reviews in a detailed requirements statement.

When — or if — you finally sit down with the firms that would like to sell you SSBI software or services, have your detailed requirements statement handy so you can say “this is what I want to accomplish, with these resources and constraints. Can you do this for me, and if so at what costs that you are willing to put on paper?”

The word “intelligence” may be part of its title, but SSBI won’t do you much good if you don’t approach it in an intelligent manner.

Creative Commons Creative Commons Attribution 2.0 Generic License Title image by  DanDeChiaro