Cost-Benefit Analysis: Startup Decision Making For Geeks

Decisions make or break startups. So how do you decide whether or not to attempt that project with a lucrative markup? How do you decide between two different alternatives? Who should i marry July or Emily? Turns out that you can analyse the costs and benefits of your decisions, within a “Rough” framework to arrive at answers that make a little sense. The answers however are dependent on some of your assumptions about the variables within the frame work.

Standard Cost Benefit Analysis

In standard cost benefit analysis all one needs to do is make assumptions of the costs and benefits of a particular course of action.

For example if you want to do a cost benefit analyses for project X

First Costs……

Operating costs (Cost of resources used for project): 2000$/month

Fixed cost (Cost Of Project acquisition): 1000$

Opportunity Cost (Cost for opportunities you missed): 1000$/month

Then Benefits…..

Project Revenue: 7000$

Now Net Benefit/Cost (-) = (7000) – (2000*months + 1000*month + 1000)

If months = 1, then net benefit is 3000$ so you go ahead and take the project.
If months = 3, then net cost is 3000$ so you don’t take the project.

This is the bare minimum cost-benefit analysis framework.

Cost benefit analysis for geeks

Ok now onto the geek flavor of cost benefit analysis. This is something i developed and use, so pls do point out any mistakes you can find. Anyway the crappy thing about the above frame work is that there is no flexibility. Your guess for the fixed costs might be wrong and hence your whole analysis might be screwed. So in my framework i make room for uncertainty in the guessed values.

So instead of saying my fixed costs will be 1000$

i’ll say my fixed costs could be 1000$ with 80% probability, 2000$ with 15% probability and 4000$ with 5% probability.

So now the above equation becomes

Cost (-)/benefit = 0.8*((7000) – (2000*months + 1000*month + 1000)) + ((7000) – 0.15*(2000*months + 1000*month + 2000)) + 0.05*((7000) – (2000*months + 1000*month + 1000))

Hence the framework becomes.

Cost (-)/Benefit = (benefit1 – cost1)*Probability Of Outcome1 + …….. + (benefit n – cost n)*Probability Of Outcome n

The above two frameworks are only for cost-benefit for a project with a one time payoff.

What if the decision to be taken requires investment and the Return on investment is expected after many years? All you need to then is to use the time value of money paradigm, which is described in my post startup valuation for geeks.

So if you see any mistakes or have any doubts please do leave a comment. Bye

-Suman

PS: I know the post sucks and is humor less but lately my cost benefit analysis shows that its costlier for me to write a long post with attempts at slapstick.

Planning For Success

Why do you think the same five guys make it to the final table of the World Series of Poker EVERY YEAR? What, are they the luckiest guys in Las Vegas? Mike McDermott, Rounders

Can success be planned for? This I guess is a fundamental question every entreprenuer should ask himself/herself(Yeah yeah politcally correct blah blah) and strive to find the answer to. The very notion of success being planned for, or in other words ,the predictability of success for a course of action seems absurd at its best. Are we naive in asking ourselves the question when we know that 9 out of 10 ventures fail? Well I think and hope that we are not. So how should we even attempt to find an answer to this question?

The Genesis
I like many other people used to, and still to some extent take decisions based on gut instinct. There was no formulaic way in which decisions or ideas were approached. Then during the short time I was co-founder of Advetta i met Rajan my antithesis in regard to ‘decision making’. Interaction with Rajan was one of the best things I retained from my stay at Advetta and I learnt a lot from him. He was a guy that played chess(metaphorically of course) all day long. For every step to be taken, we would bring to the table the analysis of all the possible consequences we could think of for that step and Rajan would try to quantify in a mathematical (I know i lost all non geeks here) way how one step was better than another. I am not saying we always made the right decisions or such a process will always lead to the right decision, but atleast we were working within a framework that attempts to reach the best possible decision. Which made me ask myself can’t we extend this to plan success?

Success? Planning? What does it mean?
First lets try to understand what “planning for success” means. The only thing that is certain right at the outset is that success is never certain (Hehehehe pun intended and gotcha!). Success here is not about an absolute certainity but a probability which in English would mean “we don’t need to win or loose we simply need to win more times than we loose“. When we attempt to plan for success we are attempting to create a framework or science which would increase our probability or chances of being successful. With this framework if we fail only 8 out of 10 time we are already more successful than the average Subramaniam/Kumar (the india version of the average joe). So to put it in a nutshell….

“a good plan for success should ensure a higher statistical probability of success”

Is this too much to ask for? Can there be no scientific (systematic) way in which we can approach this problem and attempt a solution. To the casual observer this might indeed seem so, these are the people that feel its a game of chance and that most successes are just a result of pure dumb luck. Yet what they fail to realize is that many people consistently make a killing at stuff that to the layman seems like gambling. “Consistent success” is the key factor to consider here. At this point lets indulge in little logical rambling shall we….

  • Consistent Success means success at higher probability than the norm
  • Probability that this consistent success is because of luck – Is very very low (decreases as consistency increases)
  • Probability that this consistent success is a result of a system that works(atleast for the time) – Very Very High

So a person who is consistently successful in a high risk market is advertently or inadvertently using a broad system or a broad set of rules to do it. Now that we have established that consistent success in a high risk environment is more the result of a good framework or system than luck let us look at one such scenario.

The `Stock Market’

Speculation in the stock market is the crucible for our assumptions that success can be planned for. There is probably no better or easier way to demonstrate that in the long run luck won’t get you anywhere and even simple well formulated ways of building and maintaining a stock portfolio can result in enormous returns. In fact speculation in the stock market has given rise to computer software that can consistently make profits given the appropriate statistical data regarding the stock market and external factors. In essence…

Statistical Data => System => profit

So a system working on certain assumptions and real world data can consistently generate profits. Another important lesson from stock speculation is “Dont put all your eggs in one basket” aka Diversification. Spread your risk, diversify your investments on many levels. Such a strategy would more often than not give you better downside protection.

Coming to the high risk world of Technological entrepreneurship, can we come up with some sort of a broad system to build ventures which would more often than not succeed? Well that is our hope at this point, we can’t say wether we or anyone else can come up with one such system, but we intend to attempt to find the solution as scientifically as possible. Here unlike the stock market there are hardly any precedents and they keep changing fast. But how can one build or discover such a system not just for startups but any other field where a systematic approach to success is possible. Following is a broad list of guidelines we intend to follow to arrive at such a system. Note that this is not a list on how to achieve at success but how to build a plan that delivers success at a high probability.
1. Define goals the plan or system tries to achieve

You can’t reach your destination if you don’t know what your destination is. Think what you are planning to achieve. If you are not too sure about the specifics define broad goals(hehehehe that is what we are doing), you can fill in the specifics later. For example my goal could be “My plan or framework should achieve 80% leads to sales conversion.”

2. Trial and error

Try and avoid error part of this as much as possible
– Learn from other peoples mistakes and successes.
– What works, why it works…
And trial and error is as necessary to arrive at a comprehensive plan for success if you are not committing errors you have luckily stumbled upon one possible path to while all other paths are closed. Only through friction can the true nature of the system be perceived.

3. DATA, DATA, DATA, DATA, DATA, DATA…

Oh and have i mentioned data. The crux of any scientific study is based on data and lots of it. Collect data from every angle possible, on everything possible. Once the data is collected and observed an inference or a hypothesis can be reached that tries to arrive at the end result from the given data. This at the very highest level is an attempt to discover how the system produces observed results for done actions.

And now the above paragraph in english. Imagine you own a chain of restaurants and you notice one day that the profits in winter are low. You don’t understand why? Now you go back to all the detailed data you have been collecting regarding your chain. You see that your sales in colder regions of the country are falling off drastically when compared to your partners only in the cold parts of the country while your sales in the warmer parts of the country are more or less the same. Once we have things in this perspective we can reach one of several conclusions…
Your chefs suck in winter – low probability
Customers in cold places are crazy – low probability
Your supply chain is not sourcing fresh ingredients in winter – High probability
Once again data saves the day!

At the end of all this, what we hope to have is a comprehensive system that delivers success with a higher probability. So begins our journey into the abyss of the unknown, unheard and unseen (A bit of melodrama for our female readers just so they know we may be smart but we are also sensitive and caring(who cares girls always seem to go for the jerks with a high probability (that should be your “scientific” plan for bagging a girl, be a jerk(wow too many brackets))).

-Suman