regardless which metrics you use they are close to impossible to be predicted on advanced R&D products...in my 20+ years carreer I have seen hits and missed on ROI and break even points by 10x in each direction to the point that I consider them largely useless in the decision making process
Break Even time is the preferred method to measure how long it takes to realize the payback of the initial investment, measured in months. NPV is the total value (returns-cost) of an investment. It is measured in dollars. ROI is a single measure of return, measured as a percentage. It can be compared directly to an interest rate.
In simple terms, I believe it is ROI. NPV is used to measure how long it takes to payback an initial investment, and stops there. ROI is a continuous calculation, which measures the return in terms of the costs of the goods during each period of the exercise.
By the logic of this article, managers, accountants, and possibly the janitors should all study engineering. Then they would have the ability to understand why specific technical decisions are made and what are the implications for the product.
If you are really into ROI, what would be a better investment of the engineer's time, learning finance, or getting more advanced technical training in their field to make them perform better as engineers? Frankly finance people are less rare than good engineers, so unless you want to turn them all into accountants, financial concern should be left mainly to the bean counters.
There is a reason for specialization and why some people are into finance, and some are into management. One possible outcome of finance education in engineering is an increase in unsafe products, especially if they decide it is cheaper to allow x lawsuits/year than to increase the safety of the product. Of course the problem with that is you can never know what the cost is going to be. Maybe a good lawyer could bankrupt the company with one death or injury. By the logic of this article the engineer should also study law as well so that they can make informed judgements on the likely cost of unsafe products.
Also they should get training in political science so they know how to influence lawmakers to help protect the company and develop standards. Perhaps they should also study psychology and NLP so they can influence other people to accept their designs.
No I think finance people should stick to finance, engineers to engineering, and lawyers to the law. I think the world is diminished whenever a good engineer becomes a bad manager, or worse, a bad accountant.
Actually I agree with you. Ranking sets of projects is exactly the same as ranking different complete portfolios. Different terms for the same concept. Ignoring the granularity issue, filling a portfolio with the hghest ROI projects will deliver an optimal portfolio- it leads to a portfolio that has both, highest ROI and highest NPV compared to any other portfolio. Of cousre, with granularity of projects you may choose a lower ROI program that has higher NPV so you don't have resources doing nothing. At this level of debate, all sorts of other factors are considered in creating a proper portfolio. This brain teaser was just to isolate a pure financial look at the problem and illustrate a difference between ROI and NPV.
THis makes almost perfect sense. My earlier comment about using NPV to rank sets of projects can be illustrated by imaging that you only have 2 projects to fund--one with high ROI but low NPV, the other one with low ROI and high NPV. Again, you can't pick both. In this unusual case you'd want to pick the one with the lower ROI--just like you might pick the one with the lower concentration of sugar water, if that was the only way to fill (or nearly fill) the barrel. In the bond example, say you could only choose a $5001 bond at 6% or a $9999 bond at 5%. If those where the only choices for your $10,000, the later is the better choice. Yes, this is a contrived example, and usually picking the higher ROI projects are better. I would think that it only matters for the final project you pick, but I haven't worked through any scenarios to see if this is true.
Small or large, it is better to choose the high ROI projects first. I gave the small vs. large examples to show why NPV does not prioritize the projects correctly. You can also have large projects with large ROIs too. These are winners. iPhone, for example.
There may be some quantization effects at the margin- the next project on the list is too large for the remaining resources. Here you could shuffle projects a little. At some time, you are at the limit of the accuracy of the estimates anyway.
Larry, the analysis works for organizations that do not have resource constraints -- i.e., are able to do all the smaller but higher ROI projects. But what happens when resources are constrained?
To use your sugar water analogy, what happens when the organization cannot produce enough small mugs of highly concentrated sugar water to fill the barrel? Are they not better off producing a fewer number of large mugs that, although the sugar concentration is smaller, the quantity fills the barrel?
Mathematically, it is easy to show that the organization should add more resources to complete all those smaller, higher ROI projects. But in the real world, that is sometimes easier said than done.
Should you prioritize projects by NPV or ROI? That was the question behind my fifth and final scenrio. The answer is ROI.
If you rank order all projects in order of ROI, and keep selecting the highest ROI projects until you have used all your funding, you will create a portfolio that has the highest ROI...and NPV.
If you saw my analogy above with the concentratioin of sugar in different containers, you may have already come to this conclusion. But let me make an actual financial analogy. Let's say you have $10K. You see two US savings bonds you can purchase. You can by $1000 bonds at 3%, or $100 bonds at 6%. The ROI of each is 3% and 6% respectively. The NPV of a $1000 bond at 3% is 5 times that of a $100 bond at 6%. If you chose NPV, you would have loaded up your portfolio with $10K of 3% bonds. If you chose ROI, you would have loaded it up with 6% bonds. Clearly, investing in the higher bond rate is better, regardless of the denomination of the bonds.
NPV can mislead you in another way- combine two lousy projects into one, and you get the sum of the two NPVs, and they go higher up in the rank. Their ROIs don't budge much when combined- the combined ROI is between the two. Combining the projects didn't make them better, so this is another indication that ROI is a better rule.
OK, those of you that remember your IE instructor saying something about NPV ranking will reflect ROI ranking are now scratching your heads. Here's the key to unlocking that dilemma: This is true for same sized investments, and is true on a portfolio level. If you looked at 10 different portfolios you could fund, the ROI ranking of the entire portfolio will generally match the NPV ranking of the entire portfolio. This is not true for individual projects, because they don't take into account the returns you can get with the remaining funds. A small project with a high ROI may have less NPV than a giant project with low ROI, but because more money is left over to fund other good projects in the portfolio, it is the better choice as you rank order the programs.
The was an Engineering Economic Scxience major at Stanford when I went there for grad school. Though I didn't take any EES classes, I knew people who did. They related this peculiar calculation:
When do you decide to put a guard rail up on a mountain corner? Well, you can look at statitics of accidents on mountain roads due to driving too fast for a corner, the probability of the guard rail preventing a fatal accident, and the cost of the guard rail. What you finally end up with is cost/life saved.
But then how do you calculate the value of a life to see if it is worth it? They were instructed to estimate the income potential of the average driver (or passenger), and discount back to Net Prevent Value. Not their total lifetime income- the income for the remaining years of their lives. Yes- NPV- ignoring sunk costs and sunk income!
I was taken back buy this calous calculation. I think the idea was, looking at only economic factors, this maximized society's total net economic activity. Less guard rails and you lose income, which represents the value of the goods and service that individual would have produced. Invest more, and society is investing more than the goods and services they are saving. I'm not advocating this cutoff, but I found it interesting.
Whatever the cutoff criteria, rank ordering investments by cost/life saved is a good way to prioritize safety investments.
My Mom the Radio Star Max MaxfieldPost a comment I've said it before and I'll say it again -- it's a funny old world when you come to think about it. Last Friday lunchtime, for example, I received an email from Tim Levell, the editor for ...
A Book For All Reasons Bernard Cole3 comments Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...