Business Intelligence, or analytics solutions, if you prefer, are notoriously difficult to calculate an ROI on.
The investment calculation of an analytics project has all the same cost/investment variables of a regular IT project. It is the return portion that is most problematic, especially if you are going to be challenged on the hard savings generated when you are not replacing an existing system with a presumably a lower cost one. If your analytics project is going to save lives then perhaps it doesn’t need an ROI, but if the benefits are not as compelling then things become more difficult.If your investment is not replacing an existing system, then the return can only be based on enhancing human performance, replacing human performance, enhancing human decision making or replacing human decision making.
If you are trying to replace human decision making, then you fit nicely into the traditional business case model, so even if it increases scope, propose a solution to fully automate decision making. The value you are offering is eliminating current, or future, employees and replacing them with a technology investment. Regardless of how unpalatable the equation is, everyone understands it and comparing your utterly awesome analytics solution against , say, re-furbishing the four oldest stores in the chain, replacing that production line equipment, purchasing a competitor, or opening in a new sales location is, as they say, “easy-peasy-lemon-squeezy.”
But if you are not replacing human decision making, only enhancing it and not replacing human performance, only enhancing it then calculating an ROI is now decidedly more in the “difficult-difficult-lemon-difficult” category.
Your solution may be offering wonderful opportunities to do things you could only previously dream about, but because it isn’t eliminating anything, it isn’t freeing up cash to directly finance itself. You have to explain how enhancing human decision making will boldly create value where no value has been created before.
Unless presenting actionable information is the weakest part of your decision value chain, then anything prior, or subsequent to, presenting actionable information can prevent new value from being created. You must include strategies to overcome those issues in your ROI calculation if it is to remain credible. For example, one attack against your ROI could be the contamination from poor data quality, but more on that another time.
When you are only improving human decision making, not replacing it, then your ROI must explain how it will neutralize those decision-value chain issues so they won’t be the death of your killer analytics solution.