[twitter-follow screen_name=’@dinojain’ show_screen_name=’no’]
In my last blog post, I talked about how to dig yourself out of data debt. One of questions I received concerned how to avoid data debt to begin with. I’m happy to oblige with an answer. Data debt is the practice of implementing new technologies non-strategically—thus creating a gap between spending and ROI, or debt. I can’t tell you that you can avoid data debt entirely, but there are questions you can ask of your organization that will help guide your decisions and reduce your exposure.
These questions arn’t new, but the context in which you consider them is. Technology has changed rapidly in the past 5 years, and the pace of change is only accelerating. Where once you may have given a particular answer, that answer might change given the availability—and suitability—of technologies for your organization.
- Does it fit our strategy? There are almost innumerable new analytics technologies on the market. Don’t buy something just because it’s new and powerful. You might not need all the capabilities a tool set offers. Be realistic with your needs and your resources. Even if you do believe you need a specific technology, be sure that it will integrate within your existing IT strategy and infrastructure without inordinate disruption. Implement incrementally and prove the concept to deliver quick wins and tangible value.
- Are we looking at the future? The market is littered with technologies that were cutting edge just a half-decade ago. The problem with most of them was that their analytics and presentation layers couldn’t evolve with changing analysis demands—especially the need to process unstructured data. Avoid data debt by choosing a tool set with the data storage, analytics capabilities, and presentation layer that will handle the data types and analysis needs you have now, and those you anticipate having in the future.
- What’s the payoff? When you implement a new analytics capability, you must be able to clearly define the goals of the project—the payoff. Do you need deeper insights into your business? The ability to more accurately predict what will happen based on scenario modeling? The ability to prescribe actions based on those models, thus enabling more efficient operations and better decision-making? The technology you choose, and every decision you make about the project, must be driven by those goals. If they’re not, you’re asking for lowered ROI and data debt.
- Can we do it? No company has unlimited resources. If the technology you want to implement will require more resources than you currently have, admit it, plan for it, and find a way to pay for it. If you try to implement an analytics project with a short staff or by cutting monetary corners, you will most likely fail, and your data debt will grow exponentially as your ROI flatlines. If you don’t have all the resources you need when you start, my advice is the same as it is in Q1: implement incrementally, show quick wins, and get more funding as you go.
- How can we Ensure ROI? You can’t. But you can try to make sure the plan you start with is close to the system you end up with. First, develop a realistic picture of what the technology will cost, both in time and capital, to implement, then plan for points where costs could balloon—unanticipated hardware costs, difficulty integrating new capabilities into the current technical architecture, unanticipated customization—the list is almost endless. Second, avoid too much scope-creep. Clearly delineate which capabilities you’re implementing. If a requested capability is outside the scope, park it and implement it in a later phase. If you know what your actual spend is, and you stay within your scope, you should be able to ballpark, and meet, your ROI.
My final piece of advice is this: The environment will change during the lifetime of an analytics project. These days, it’s changing faster than ever. Be prepared to meet—and embrace those changes—by using the tools you need, and enough resources, to build your system incrementally and via a well-organized, informed method that shows value early on with quick wins. Your system won’t be perfect, but if you’re disciplined and honest with yourself, it will most likely be successful and keep you out of data debt.