Make smart decisions, the smart way

Make smart decisions, the smart way

Authoring
Sergi Salvador Lozano
Web Developer at Tyris Software

Comparte esta página:

Organically growing your project’s Business Intelligence

What do you mean with that clickbait title?

So, since a few years back, you kept hearing more and more about this “business intelligence” thing (mostly by its initials, BI), and it sounded like a buzzword at first (not unlike “smart” or “the cloud”). But then you also learnt that many big companies were dedicating entire departments to it, so it has piqued your interest, and you have started wondering if BI has more to it than just marketing speech:

“can this be truly useful and worthwhile to my enterprise?”.

Short answer: Probably yes.

Slightly longer answer: If your enterprise has room for improvement and decisions are going to be made for that purpose, then yes, as long as you approach it sensibly (isn’t that true of just about everything, though?).

In this article I will try to cover, at an introductory level and in layman terms, the “whys” and the “hows” of business intelligence for those companies or projects who are considering implementing it.

I will be focusing on software projects, as that’s what I know about, but it should be noted that BI can be implemented in a wide variety of contexts.

Why business intelligence?

Turns out that, for all the complexity that can potentially be found under it, the main idea behind BI is exceedingly simple (and also over sixty years old, believe it or not):

In a way, the traditional approach to decision-making can be seen as an “artisanal” process: a person or team uses their personal expertise to decide what should be done, when and, to some extent, how. While this doesn’t necessarily lead to bad or suboptimal choices, in those instances where it does the problem might go undetected, making a project worse for it.

Business intelligence entails a paradigm shift, in that a major part of this expertise is redirected not towards straightforwardly answering a question, but towards deciding which questions should be asked; then, through the collection and analysis of usage data, an informed answer can eventually be formed for each of them.

Business intelligence, the unintelligent way

Great then! Let’s jump right in! … or better yet, let’s sit back for a moment and contemplate how some intuitive ways of getting started can actually be counterproductive.

As with any new technology, it can be tempting to try and get a full understanding of each and every nook and cranny before starting any actual implementation; this cautious approach is not necessarily bad in general, but if taken to the extreme (and your team is not already somewhat familiar with BI) you will be biting off way more than you can efficiently chew.

Business intelligence is a very wide umbrella that covers many solutions and strategies, more than most projects will ever need. If you try to get a deep understanding of every one of them, you will be investing vast amounts of time and effort that could be better directed elsewhere.

BI

For similar reasons, I wouldn’t recommend purchasing premium BI products or suites (as great as some of them can be) before you and your team are more experienced, and can accurately assess which features are really needed for your use cases. Even though that’s slowly changing, business intelligence was initially born as a set of technologies aimed at big businesses who were used to working with vast amounts of data, and price points tend to reflect that reality.

Finally: BI is a multi-layered process where all levels need to work together in order to achieve their purpose. Approaching it vertically (one layer at a time), which most commonly translates into aimlessly implementing lots of data mining for future analysis, can bring problems of its own. Data storage can rapidly get out of hand, part of the mechanisms of data collection that turn out useless, and it can even have a perverse effect on the next layers (the data you are already collecting might influence its analysis, instead of the other way around).

Instead, let’s take a look at a different way of bringing business intelligence to your project.

The way forward is organic, and compartmentalised

Even though BI is a wide discipline and there is not a single right approach, in order to get started fast, it’s a good idea to commence small and grow from there. More than anything, it is crucial to have a deep understanding of the needs and possibilities of the project or business itself, break them down, prioritise them, and only then start addressing them one by one.

For every objective, a BI workflow can be implemented to look somewhat like this:

  • Implement data mining:

The team in charge of implementation (in the case of software, the developing team) then add to the project the necessary meters to collect structured and consistent data for those dimensions.

  • Collect usage data:

Now it’s time to let the system work. At certain points during this phase, controlled changes can be introduced in order to measure how they affect the goal.

  • Analyse collected data:

The analytics team then transforms the data into new formats (like statistics or graphical representations), that allow them to draw conclusions about the matter being examined.

  • Make decisions:

Finally, the people responsible for business decisions can do their part with solid data in hand to inform their choices.

BI Chart

I should reiterate, that the previous workflow (which I chose because of its simplicity) is just one of many possibilities when it comes to implementing business intelligence: it all depends on the goals that must be achieved.

Some projects rely not only on internal usage data, but also on external information, while others opt for workflows that automate the process in ways that allow for real-time decisions.

However, the general points that I want to convey remain: grow your BI organically, horizontally, and with clear priorities; and understand that there are multiple disciplines involved, multiple layers that, while they benefit from good communication, require different backgrounds of expertise.

TL;DR (Conclusions)

Business intelligence is the process of using large amounts of collected data and analysing them, in order to improve the decision-making process for a project or business.

It can be worthwhile to consider for most enterprises, but it should be approached sensibly: be aware of your needs and don’t aim to implement every BI-related technology unnecessarily.

A good way to get started is to build your BI following an organic workflow (starting small and growing from there according to your needs). BI is a multi-layered process, and horizontal approaches can help preserve cohesion between those layers.

Authoring:
Sergi Salvador Lozano
Web Developer at Tyris Software

Comparte esta página:

Cerrar menú