While Big Data and its associated phenomena and pomp are very popular these days, it seems many organizations are missing a lot of opportunity to improve their results and their businesses with the data that is already right under their nose. I am talking small data, the data that you normally have access to through your regular reporting functionality, or the data you look at now but look at only as a one off report and not in conjunction with other results that may prove instructive about some of your assumptions and results.
Why small data? Big Data projects are often associated with cost overruns, scope creep, and insufficient results to justify the return on investment. Small data projects have all of the advantages of being smaller: they are easier to set up, easier to prove a win with, and easier to manage since you understand the scope of the data with which you are working.
Plus I think there is a really good argument to be made that thinking small is really more effective, at least in this stage of the game. Exclusively focusing on how big the data is misses the point. Small data is just as important, if not more so, and it can be much easier to learn how to apply questions and derive insights from small datasets that you can easily wrap your head around than it can be to try to find a needle in a giant haystack of unstructured data. Put simply, the real point here is about developing the right analytical skills and learning to ask the right questions of your data and then discovering how to apply those insights in productive ways that really accelerate your business.
The key skills to develop and hone—easier on small data and likely more effective to start with, too—are
- What data do I have now and what does it tell me about where I am now? How is this my starting line and where do I want to go?
- What does this data tell me about how I got here and what does it mean? Why am I here, or conversely, why am I not where I want to go?
- How do I get from where I am to where I want to be? With what this data tells me, what are my next actions?
There are some additional benefits for thinking small too. For one, obtaining and interacting with a small quantity of data is usually one of those projects you can begin immediately with your own authority and not have to go requisition supplies, equipment, and staff time. This lets your team operate in a more agile fashion, responding to marketplace needs or good ideas that you get from customers and partners or internally in a quick way without a lot of formality. Small data is much more available as well: it can generally fit in a large spreadsheet so it is easier to work with and understand with the tools you have Available. It is also easier to reconcile your efforts with a defined return on investment: you spent a week looking at abandoned cart logs reconciled with web traffic and you were able to tell, for instance, that a slow down on your site caused you to lose customers, so you are able to say with certainty that fixing that problem would result in an average of 27 sales restored each day, for example. Quick and easy wins like this lead to the confidence and consistency that are needed to begin or expand a data program into bigger pursuits. Finally, it is also easier to take a small known data set and find both tactical as well as strategic insights from it. Consider web analytics, which can tell you if your current marketing plan is resonating with your website visitors (by looking at the length of time on a web page, your bounce rate, and the path visitors take when on the site) and well as how well your site is laid out against a conversion goal (by setting up a standard goal path and measuring conversion rates against that path). The larger the data set, the more difficult it is to find and integrate both types of actionable data into your planning.
The bottom line here: try to avoid getting blinded by the quality of the data. Do not get lost in the haystack. Start small, with data the scope of which you understand, and search for meaningful insights and results. Then lather, rinse, repeat. Hone those analytical skills and you will see a payoff quicker than trying to ask the right questions of a universe of unstructured data.