10 MayCost savings: what to legitimately count when figuring out the benefits
What’s the benefit? Quantify the savings! Both well known chants in our industry and in other industries. Benefits have to be quantified. Fair enough. Quite often, as in many things in the financial services industry, I have seen some math that just does not stand up to much scrutiny, both in terms of reflecting genuine savings and in working out costs. This week we’ll look at benefits. Inside a bank there is most often a figure used for the average cost of a head. Front-office and back-office are two different beasts, rightly so. The typical figure for “fully loaded costs” in the back-office is around $200k per annum for staff in Europe or the US and some $50k. That is the cost of one more head. Generally, that is made up of direct costs and allocated overhead; all those nice charges from the body corporate for real estate, HR, finance and the like. The exact numbers are not important per se. What is important is to consider what will be saved if you have a project proposal. Generally these proposals involve spending money on systems in the hope of saving people. If you can save heads, what is the value of the saving of one head, or FTE, if the “fully loaded cost” is $200k? If this were a question on a CFA exam, you would be asked to select from the most correct of the following answers:
- The direct costs of that head
- Something else
Let’s look at these in turn. Option A. This would imply that the firm instantly saves the fully loaded cost. But , the clue is in the name here; “fully loaded”. We need to go back to college days and look at at that most dry, but very useful, of topics: cost accounting. All that allocated cost for rent and the like is what the bean counting fraternity call a “sunk cost”. Your rent won’t change, the number of HR do gooders you have to pay for won’t change, unless of course you are RIF’ing the HR folks, where a bonus might be in order. All you will save is salary and benefits. So A is not right. B would be a guess since you don’t know the allocated costs. C is close; the direct costs. That would be a good answer and better than most people do. D is actually the most pure answer, because on top of C, you would have some redundancy costs. Or at least you would normally.
Now comes a but and it is a big one. If you cannot really save as much from a discretionary project as you would like or need to, you need to place a value on the intangible benefits, then lots of projects will not get done and at the margin you will likely hire more people. Particularly in outsource locations like India, it is very tempting to hire more Indians. That of course has an immediate upside as you for sure will not have more Chiefs than Indians, but more seriously, it would be wrong to use 100% of the fully loaded cost to work out what this costs. More importantly, there is a question of scale. Normally you want to invest to make things less manual and save money. The collective term for “less manual” is normally “improve controls” because good back-office staff know that when volume rises, just throwing people at the problem will not ensure control. Again, college days, or even A-level economics, will help us here. The Law of Diminishing Returns. If you keep adding resources, you do not get linear returns. Another factor here is time and coordination; if you had lots of operational issues, for example unmatched trades, even if you could hire and train lots of staff, will your systems allow them to all work in parallel easily without getting in each other’s way? And can they het the work done in the working day. There is not much point calling a German bank about a problem much after 5pm; they’re on the S-Bahn on the way home or already down the pub. Same is true after all when you get on the case of your IT people to buidl things for you. If it takes five guys five man years to do something, adding more resource half-way thru the project to get things done faster will not double productivity. My own maxim on this is to remember that “control does not scale”; it is though hard to put a value on this, but it is a case to be made.
Lessons Learned: It is easy to overstate tangible benefits. Most folk do and most shops won’t notice. However, it is also key that you understand the control story; will what you are proposing help get the job done in the hours available. When challenged, chant “Control doesn’t scale”. If a more easy to understand version is needed, “remember, nine women can’t make a baby in a month!”. This is of course, somewhat non-PC, and I hope that no members of the fairer sex will take offense as it is the best expression I have heard to describe the case.