Automation with a leaky business case and suspect ROI
Google “business case RPA” and you get over 200K results. At least the first two search result pages (I didn’t bother to check beyond that) spoke only about how amazing the business cases are. Some said, it is undeniable.
Most consultants, and consulting companies target a potential situation with labour cost reduction as the spear head. Usually, that is where most of the cost save is, too. Even if you consider work being done in an offshore location with cheaper resources. More often than not, the cheapest and junior most employee (say in IT/ITES), in the cheapest offshore location is costlier than a piece of software. This would, likely, hold for any other people intensive process run industry. Process Automation leading to labour reduction is almost a given. After all, how can inserting robots possibly not give cost savings, right?
Allow me to take a step back. A business case is expected to capture the reasoning for initiating a project or task. The question, however, is if labour cost reduction should be the only component of the business case and should a Process Automation led transformation be executed solely on the basis of labour cost reduction?
Till recently, having dealt with multiple of these process automation situations, the labour related use cases are what occupied mind share and what I pushed for. Until I was proved wrong at a recent client engagement.
Here is a billion dollar company (in health care payer space), with about 10,000 employees and over a 100K associates, starting to take baby steps towards DX. For a payer, the largest cost is incurred in claims (in paid and overhead) and that is what they want to keep down. As we got to know the company, its ethos and culture more, we realized that they prided themselves on their large people strength and the fact that the provided employment to a large number of people (full-time or part-time).
Publicly available numbers, and back of the envelope type of calculations showed us that their total expenses were ~$97mn, and thus an operating profit of ~$13mn. Even if we rather optimistically assumed their wage bill is 60% of their operating expenses, then the annual per person cost is <$6K. Given their size of the company, one expects their claims department to house (say) 800 employees.
Now, comes the interesting part. Assume, large big bang process automation (on cognitive or AI) which eliminates 35% of those positions in claims. I go 35%, because a large part of the work is intelligence based and not just brainless processing. So, the cost avoidance is ~$1.6mn, and thus only about 1.7% of their cost.
Is 1.7% saving good? Not too bad, in absolute terms. But, if you see the company’s culture, their strength and their very deliberate contribution to the society at large (as an employment generator), it is trivial and likely will do them irreparable reputational damage should they let go of 500 people. Cost outweighs benefit, and creates a large hole in the automation business case right there.
Should they still go in for automation, if there will be no labour reduction? Yes. Will ROI be reached as quickly? No.
There are yet avenues for redemption. The target for the business case is now different. Automation will have impacts on two areas – speed and accuracy of processing. This implies that claim processing will happen cleaner, quicker and more number of claims will get processed straight through. Just the efficiency brought in here will show indirect lagging indicators in new sales, marketing and renewals. All these will impact reputation and cause a virtuous cycle. It is unlikely that these positive impacts will be simplistic enough to catch on a spreadsheet right away. The correlation will start getting visible in a while. However, the impact is powerful enough to justify the investments to be made.
This is where the data generation, gathering and the analysis comes in to play. But, that is a story for another day.
So, clearly (and you knew this, didn’t you?) the validity of the business case is dependent not always on cost reduction vs amount of money spent on the RPA implementation. It could be on potential revenue growth, accuracy of results, reputation, social cause and marketing impact as well.
Data used in the post above are excerpts from information available on the internet as part of public disclosure. And, I have used the phrase Process Automation and RPA synonymously in the piece above.
How do you view Automation? Do you consider only RPA? If you would like to have a chat about possibilities of automation, please visit the 3nayan web site or write to us.
About The Author
