It’s inevitable, for those of us working in any data-focused profession there will inevitably come a time when we’re asked to evaluate or present our justification for purchasing, a new analytics system. There are myriad reasons for what triggers this. It could be that the tool we’ve been using is simply showing its age and we want something with additional functionality, perhaps management is pushing for us to use something cheaper, or maybe our favorite tool is no longer being supported. Regardless of why, we now have a decision, or series of decisions, in front of us, either individually or as part of an organization.
How do you decide to settle on a new analytics system?
Well, there’s a right way and a wrong way. The wrong way is pretty simple. It starts with someone uttering one of my least favorite terms: “best of breed.” What’s wrong with that you might ask? Wouldn’t every organization benefit from having a solution that’s the “best?”
Sure, but when someone says a solution is “best of breed” what they generally mean is that it’s the newest or the most popular. While that may sound appealing, every organization’s technology stack is unique and that’s going to have a powerful effect on what the right analytics choice is for you. A more rational choice would be to focus on the feature list but that too is problematic, as functionality has become so consistent across solutions. Instead, I suggest you focus on a combination of three factors: shortcomings, technology and process.
When I say shortcomings I’m referring to the reason that you’re abandoning your existing solution. Ask yourself what it really is that you’re looking to replace and how you can help your colleagues be more efficient in their work. Excellence in that functionality is obviously a key consideration to bear in mind when selecting a new solution.
What’s your technology stack like?
Another facet to consider has nothing to do with the technology itself: your organization’s culture. Why does that matter? Because even two companies facing the exact same pain point may find that it makes sense to deploy different solutions. Consider a small startup versus a larger, more established, organization. The startup may prefer a solution with less customization and more out-of-the-box functionality while the larger organization may prefer the opposite.
Why? Because a larger organization is likely to have more budget for headcount and can afford to hire someone, or even a team, to focus exclusively on analytics. By contrast, a startup is an organization with limited headcount where the person tasked with managing analytics is likely to be managing several other functions as well. As a result, they may have a preference for a system that provides more out-of-the-box functionality and better ease of use, the “minimum viable solution” if you will.
Just like it’s important to understand your organization’s culture before making a purchasing decision, it’s also critical to understand your company’s existing technology stack. No solution exists in a void, and that’s doubly true for an analytics engine, which by definition needs to process data from external sources. Even more important than the features of the solution you’re considering is the degree to which it will play nicely with your existing systems. It doesn’t matter how robust a solution’s feature list is if you can’t make use of its functionality with the technology stack that you have in place.
Why do new systems fail?
In addition to culture and technology, the final factor to bear in mind when purchasing an analytics system is the existing analytics process within your business. One of my favorite books on the subject is Phil Simon’s Why New Systems Fail, which cites research finding that a shocking 68 percent of IT projects fail. The cause of these failures frequently has little to do with the solutions themselves and more to do with existing processes at the companies deploying them. That’s because there’s a tendency within companies to try and shoehorn in faulty processes to new technologies.
Speaking of faulty processes, I’ve written about the need to have an effective data governance strategy in place and it’s worth emphasizing that if your analytics process is poor and your data is unclean then it doesn’t matter how good your analytics engine is. The results will remain poor. It’s a mistake to try and look to a new analytics solution as a panacea for ills that you should be fixing internally. If there’s no way to ensure that your data is updated and correct and that existing systems are in sync, then you’re far better off working to map out and consistently deploy a good data governance strategy than buying a new solution that may well only make the process worse.
Seek answers, don’t chase them
In short, choosing a new analytics solution often has much more to do with internal factors than with the feature list of a given offering. None of which is to say that comparing the functionality of different offerings is a bad thing. By all means, do so. However, the feature list of an offering has far less to do with the success of its deployment than approaching a purchasing decision with a clear idea of what your existing technology stack, culture and analytics process looks like and chasing after the latest “best of breed” solution is likely to end in failure. Philosopher Auguste Comte once said that you must “know yourself to improve yourself.” That’s as true with regards to an analytics program as it is to anything.
Recommended Read: For Retailers, the Time for Digital Migration Is Now