Real ROI is complicated: musing on where to spend your next analytics dollar (part 1/3)


first Google Image search result for "cold water", from
first Google Image search result for "cold water" from

How does your company measure analytics ROI?

It may be something like this:

ROI = ( Value / Cost ) - 1

I believe the world of analytics has oversimplified view of ROI, ignoring important influences on achieving the purported value solutions create.

In these three posts, I discuss two overlooked factors in getting the maximum value of your financial investment in analytics. I hope that considering them and the full spectrum of investment options will lead people down unexpected paths, to the benefit of their organizations.

Value of new effort should be defined as the delta compared to the value of your current solution

Like redoing your kitchen, solution ROI depends heavily on your current state. My last house, still with its original kitchen from 1949, provided a good return when I remodeled it in 2015. The value before and after was as different as night and day.

But the same math did not apply to my current house, whose kitchen was remodeled just a few years before I moved in. While I did not care for the choices they made, the cost to redo it would be the same as my previous house, but the value would only marginally increase. Later, when water damaged several cabinets, and they had to be replaced anyway, the ROI was more acceptable, so we went ahead with the project.

What does this have to do with analytics? If you currently have no solution for a business problem or to support a business process, the value created is all to the good. But if you already have a solution -- even if suboptimal -- we should be measuring only the increase in value relative to the cost of achieving that increase. 

The reason I have been thinking about this lately is the trend among companies to migrate between similar analytics platforms, including much of the content. It is particularly surprising in the budget-strapped economic environment of the pandemic. Migrating content between similar platforms is a purely technical exercise, like allocating finite resources to tear down a house and rebuild it across the street. It is unlikely to be valued by the business users on whom you rely to convert your solutions to business value. This is not a comment for or against any platform my company supports; rather, for most use cases, it is the use case that matters, while all major analytics platforms can support them reasonably well. 

Prioritizing work that supports unmet business needs is likely to be higher ROI than like-to-like content migration. Selectively purchasing new technology to meet use cases your current technology cannot support and selectively migrating content when the business case justifies it is a saner financial approach, if you plan to eventually sunset a platform.

"Cheap, fast, and good: pick two" applies to this situation. Pragmatic migration approaches, even those advocated by the vendors themselves, span years. Maximizing migration ROI is likely to involve a phased, deliberate process (cheap and good), not a big bang migration of everything at once (fast and ???). So if you plan to migrate, start now and start slowly.

Part 2 will be: Value is 100% dependent on what people do after you deploy a solution

Contact Form


Email *

Message *