How to measure the ROI of your Analytics team: Part 1 = Profit
I’ve shared previously why this is so important. In a similar way to the way Marketing has been criticised for decades, a Data Science or Analytics team cannot afford to be seen as just a sunk cost or vanity project. In his usual practical style, grounded in commercial experience, in this post, Harry makes the case for profit as your key metric.
Harry Powell is currently Director of Data & Analytics at Jaguar Land Rover. Prior to that, he has lead Advanced Analytics at Barclays Bank & had years of technical experience in data & commercial roles. So, over to Harry to introduce the case for profit as your measure of Analytics ROI…
You are expected to show Payback
Not before time the data science world has woken up to the idea that however clever you are and however cool your models might be, the rest of your organisation is going to expect some payback sooner rather than later. The honeymoon will sour if the business doesn’t feel measurably better off.
All those data dinosaurs will have been proved right; it really was just another passing fad. Most companies may like to talk about research, but they don’t really want to do too much of it. It just annoys those who have to generate all the money to carry those deadweight vanity projects.
Maybe that’s a bit extreme, but they have a point. You have to earn your seat at the transformation table.
But what should you measure to show ROI?
Driving change with data is hard. You need to focus on what matters most and deploy all your resources on that. And then you need to be able to show that it has changed for the better.
And so we come to profit.
Since the 1980s, business cases have been framed in the language of “value”, a shorthand for shareholder value. Since shares were priced on all the information available about the future prospects of a company, and since the role of the company was (thought at the time) to be to generate returns for shareholders, it seemed reasonable to assess business activities on the basis of how they contributed to shareholder value.
But value is a somewhat nebulous concept that can be bent pretty much any way you want. It is after all very forward-looking, encompassing all future cash flows (that’s a lot of years into the future). Those cash flows need to be discounted at some rate (and it turns out that no one really knows what this should be). Then there is asset value, brand value, technological capital etc. There are enough levers to pull, and enough uncertainty involved, that you can pretty much tell any story you want to get an investment approved. Value tends to be preference disguised as science.
Why profit is better than shareholder value but not perfect
But it’s worse than that. Because next year’s cash flows are only marginally less valuable than this year’s (especially now interest rates are low). If your project fails to live up to expectations you can just push back the returns and hopefully, by the time any failure is obvious, the world will have moved on and no one will be watching. So value hardly provides a robust guide for where to prioritise your effort.
Profit is a blunt tool. It is just sales minus costs in a given period. It doesn’t include all sorts of good stuff that might result from doing a project, like learning, or customer net promoter score or quality improvement or whatever (You would definitely want these kinds of things to be considered on your scorecard).
It too is not beyond manipulation, although that manipulation would be at the corporate level, and it would hopefully be identified by the audit team. But at least it’s clear. It happens in an agreed near-term time frame. It is regulated and reported in the accounts. It can be checked – you have a whole finance function designed to do this for you.. Compared to value, there is little room for disagreement.
The value of setting a profit target as your ROI metric
By setting a profit target, you have a hard target that everyone can see. It demands that the data analytics team focus on realistic goals, not fancy pipe dreams. It insists that insight is connected to action that matters. And it requires that that action can be measured. Focussing on profit will result in clear goals that are delivered.
Of course there are other material considerations. You will want to show that profit can be converted to cash; that the profit can be scaled and replicated year on year (sustainable growth); and that you have grown a capability which can find and deploy further profit opportunities (value multiple).
But start with a clear, short term measurable goal. And my recommendation is that profit should be in there somewhere.
Are you holding yourself & your team accountable to increase profit?
Previously, I’ve also argued that it is helpful for data leaders to hold themselves accountable for profit targets they don’t fully control. It is all just part of building the reputation that you need for joined-up working & commerciality.
But what about you? How are you measuring the return on investment (ROI) of your analytics or Data Science team? Is profit working for you as a metric & if not what alternative have you found? Please share in the comments boxes below…