It is common to read that one of the major issues with search or text mining technology investments is that it is basically impossible to calculate the ROI.
While I agree that calculating ROI is less deterministic than with other technology investments, I disagree that it can’t be done–it’s just that the calculation that has higher variance than others. Estimating the value of an organization’s information assets—specifically the value of each piece of information available to it at any moment in time—varies widely.
At a certain moment in time for any organization, its content, its information, has a precise value. This value does not always correspond to the cost of producing it, nor the cost of gaining access to it, but goes much farther. For example, a product brochure’s value is derived from its utility as a vehicle for providing detailed information to prospective customers, which helps salespeople do their jobs more efficiently (because they don’t have to spend time creating material), and minus the actual production cost (writing, design, printing).
The value of information is not constant. For a digital publication, the value of an article or a post peaks at around the time of its initial publication because it has greater visibility with readers as a new piece of content. Over time, its value decreases as it is replaced by newer, fresher content.
This decline is not the same for all kind of news. Gossip or political news lose value much faster than editorial, while restaurant or hotel reviews tend to maintain their value over a longer period of time. The same is true for enterprise content. A corporate presentation retains its value until a new version is issued, a customer-specific presentation or a sales offer lose value more quickly, although not completely as they can be reused in part for similar prospects.
To calculate the ROI of a text mining or search project, you have to calculate the value of finding information faster or the impact of not having access to critical information, minus the total cost of the project. Not finding a procedure in the company intranet could result in a costly compliance issue. Not finding an important presentation in time could cost you a customer. Not finding an existing patent could cost you millions in wasted research, and so on.
Obviously, the hard part is assigning a precise value to this information. However, this is no different from other estimates in a decision-making process. I think people often shrug off this valuation because, no matter which assumption you use to calculate it, the value of such investments is, for lack of a better word, huge. And because it is huge, “it can’t be true”.
In reality, not only “is it true,” but it is getting even more true the more information an organization processes. So, the next time you are evaluating the ROI of such investments, try to make this estimate under the most conservative assumptions, then take the number and divide it by 3. You’ll see that you will still show a great return, which will make you feel better about the investment. And rightly so.
Author, Luca Scagliarini.
