There are many ways how to pet a cat. Obviously, the belly rub is at the top.
Similarly, there is plenty of ways how to try to calculate ROI on your data analytics solution/function.
Of course, mostly you will see ROI in phrases as “Get ROI on your data analytics.” etc. on the first pages of any consulting or technology pamphlet. You won’t find how do they want to calculate it, after all, it is your problem, not theirs, isn’t it? You will be trying to figure it out one year after all the money is spent.
Surely by now, you are thinking “What is the exciting acronym in the title?”. Drumroll… Answers For Genuinely Asked Questions. Yes, I love my nonsense acronyms.
Also, before you will jump in with “How the hell would you quantify this?!”. Yes, I agree. In a normal world, it is pretty much impossible to quantify. Same as any other ROI on data analytics. Mostly you will see ROI based on automation of a manual process, which is not the knowledge you get from data analytics.
Since that is out of the way, let’s break it down a bit.
What should be common for us data folks, is tracking usage of our tools. Who runs which reports, dashboards, accessing self-service BI tools. When do they do that, how often etc. Understanding this is great because you can use it for your own work prioritisation and any BAU work.
When you try to use it as an argument for ROI, that is where it falls short. Let’s imagine that you have a person who has run 60 reports over the last week. Most likely we would say “Brilliant, our tools are used amazingly well.”
However, let’s look at it from the analyst’s point of view. They are being asked: “How many blue jackets have we sold in the UK year on year in the last 10 years?”. The person has just spent a week trying to go through a chunk of your reports trying to find if the answer is somewhere or if they could combine some reports together to answer it. Failing miserably in the end.
Great usage volume, not much of ROI.
“Why is there the Genuinely?”. It is there to cover “If you build it, they will come”, which we are all guilty of. Some of us more than the others. If you are on the “more” side of the scale, you might start to excuse what you build with “They CAN answer these questions.” and you will use that for the ROI justification. It’s not always bad, but be mindful that CAN use it, doesn’t mean that it is useful. However, this is a topic worth a separate post.
You want to also think about the impact of the answers. Naturally, not all answers are equal. “How are we doing in meeting our H1 public targets?” is most likely not the same as “Did a specific person changed a profile photo in the last 30 days?”
You have probably realised that this is more of a thought experiment than a calculation for your project to submit to your CFO.
And how can you go about this thought experiment? You should have an idea about how your tools are being used. Talk to your users and let the conversations challenge your assumptions.
*Replace this with data science, data warehouse, BI, MI, reporting etc. based on your phrase of choice
Just above average 