So you’ve built the business case for solving your company’s big data analytics problem.
You’ve set up the first necessary steps.
You’ve even corralled resources and you’re getting ready to build a great team.
And suddenly, the question of software pops up.
In today’s world, when technology appears to be replacing people everywhere, you might be wondering—if I get software, will it replace me too?
The answer, is no.
In fact, play your cards right and software can play a role in making you an analytics superhero and an invaluable part of your company.
Let’s look at the facts. While data analytics software can be great, it’s nowhere close to replacing the deep knowledge and decision-making skills you have to offer. Instead, its role is to make things easier and more efficient for you while helping you shine.
In my two previous articles on this blog, I talked about whether you should build or buy your data analytics platform and also, best practices for deploying your solution. Don’t forget to take a look!
In this article, however, I’ll be covering the last part of making your big data analytics journey successful—and you into the analytics hero of everyone’s dreams.
Here’s something to consider: as the analytics hero, the strategic and tactical decisions belong to you and your team. I want to run over some best practices that will help you achieve this and perform your role in true heroic fashion.
I break this down into five steps:
Now let’s delve into what all of that means.
You might be tempted to rip and replace. Why not? A clean start, a fresh beginning, etc. But doing that pretty much always comes with unforeseen delays and complexity.
That can create bitterness and tension in the community, especially among those who implemented the previous solution and are still attached to it, or were advocating for a different solution.
Instead, always keep your project’s main purpose first and foremost in your mind. Are you delivering the proper function to the business? And are you continuing to support the communities that are already using these systems?
Keep those communities happy, and they’ll be happy with you and your project.
Another great reason to keep your previous systems running: as you roll out your new products, they may have issues that you weren’t initially aware of. Incorporate some backup options in place until you’re certain you won’t need them anymore, and then roll out a phasing-out plan that keeps everyone, including those late-technology adopters, happy.
Okay, let’s get this out of the way. I work for a big data analytics company. Of course I want you to buy my software. But even more than that, I want you to succeed. After all, if you don’t succeed, I won’t either.
So here’s what I have to say about buying big data analytics software:
One size does not fit all.
Purchasing independent software solutions for your every need is tricky, and the criteria vary. And it’s highly unlikely that you’ll find one magic solution that can do absolutely everything you need. So here’s what I recommend.
Take your list of business use cases and their value, and use that to prioritize which software capabilities are absolutely essential to changing the business, and which are merely very nice to have.
Examine this list, put it aside and take it out to look at again, and go out and buy what you can’t easily build.
For example, with data integration and preparation, reaching out to data sources is key to streamlining delivering analytics. You’ll have to buy platforms that support the sources you need.
By doing so, you have an independent software vendor that delivers maintenance and support on diverse data environments without distracting your team and causing a small flurry every time a new source comes around.
So don’t just look at what you can easily build and what you can’t. Also look at what will require constant upkeep and maintenance, and whether the cost savings will really be worth it if your team is constantly chasing after new updates.
Remember your basics. What does that actually mean?
Here’s my perspective. In the world of big data analytics, the project owners have sometimes been a little … cavalier. They haven’t always thought about the needs of the larger organization.
But this is your opportunity to change that. Champion yourself as the big data analytics hero for everyone. That way, you’ll find it easier to get them on board.
Here is what you can’t forget though:
Prioritize these and roll them out throughout your project’s lifecycle. They’re important enough that they shouldn’t be an afterthought. And don’t push off your metadata management or data modeling; it’s time consuming and wasteful.
Refer back up to point number two, “Buy What You Need” if you would like a refresher, but I’ll add a second point to that here. Another advantage to buying what you need is that it provides you with what you need to do more customized builds for your business.
For example, I have a client that designs and manufactures ocean-faring boat buoys. Unfortunately for them, no software package can meet their unique needs. So they have the opportunity (and the excuse) to build what they want.
And when you’re building only what you want, well, that’s the fun part! Go out there, build it and have a blast.
Creating your big data analytics solution is going to be a constant work in progress, especially since you want one that allows you to deliver agile solutions.
Keep in mind, this does not mean going out and building new things every day. Just think about the time and effort it takes to maintain all these custom solutions. Eventually, you won’t have any time to think if you’re busy maintaining custom-built applications.
What it does mean is selecting a good mix of out-of-the-box products that meet your criteria for scalability, governance, analysis, visualization and management.
At the same time, take your homegrown projects and understand which ones are unique to your business and because of that, are truly needed. Which ones can be slowly moved and managed and adapted to streamline your solution cycle?
Constantly evaluate your big data analytics solution and see how you can optimize it — it’s how superheroes stay ahead of the game.
This recipe for success is made up of my own observations on what has succeeded in the market.
But every deployment is different. Take these learnings as your baseline, but as your own big data adventure unfolds, tweak it with your newfound knowledge as you become a fully fledged big data hero.
As you continue along this journey, you’ll learn to save your business and usher in a data-driven approach to operational excellence!