Need Targeted Analysis? There’s An App For That – Soon
I argued in my last post that the analytical technology environment will change dramatically over the next several years. One key change will be a shift from multi-purpose technology environments to focused applications and data.
Analytical capabilities have—since at least the 1970s—been multi-purpose. In this model, users are given an extensive toolbox of analytical methods and procedures. The early analytical vendors called their offerings “statistical packages,” and they contained tools from ARIMA (autoregressive integrated moving average) to variance analysis (there are few if any statistics beginning with the letters W-Z). It was the job of the analyst or the decision-maker to decide what tools were appropriate for what analytical context. This, of course, required a high degree of analytical sophistication—one that many analysts and almost all decision-makers lacked.
Many data environments for analytics were also multi-purpose. The idea behind an enterprise data warehouse is to support a variety of analyses and decisions. Data marts, of course, are typically intended to support a single type of analysis, or at least a narrow range. The size and complexity of warehouses made it difficult to find the data you wanted.
Multi-purpose environments are desirable for their creators (IT organizations) and highly expert users; they allow lots of flexibility for expert analysts, and greater productivity for an IT organization. It is akin to the difference between a warehouse club and a personal shopper; without the need for individualized attention, warehouse clubs can use their labor very productively. And they offer a lot of merchandise to the expert shopper. The downside is that average shoppers—for goods or analytics—may not be able to find what they need.
Another problem with the multi-purpose approach is that it makes it difficult to determine return on investment. The tools and data warehouses are intended to improve multiple decisions and business processes, so it is difficult to determine the value to the organization.
Going forward, I think we’ll see analytical widgets or tools that are linked to a particular type of decision in a particular industry. If you want to forecast retail sales, you’ll have an analytical app for that. If you want to do physician targeting in a pharmaceutical firm, there’s an app for that. This might seem to be difficult to find the app you need, but there are about 250,000 iPhone apps, and people seem to be able to locate the ones they need or want. The apps will not only do your analysis, but will also guide you through the process of ensuring that your data are in good shape for it, interpreting it, and making a decision based on it.
Data marts will be single-purpose too, and may be assembled on demand from composite data sources. I think there will still need to be some “staging area” for data analysis, but I’m guessing it will be temporary in many cases. It’s just too hard to predict what data you’re going to need ahead of time, and various new technologies are making it easier to extract the data you need when you need it.
I am not sure that statistical packages and enterprise data warehouses will disappear completely, but they will be the province of a relatively small population of analytical professionals (as they are now to a large degree). They may not even shrink in absolute numbers of installations, but the growth will be in focused analytical applications and data environments that can be used by millions and millions of decision-makers.
