Prioritizing With Uncertainty

September 29, 2016 Focus area: Scaling Agile

Product owners often prioritize user stories using the Weighted-Shortest-Job-First (WSJF) method. The WSJF prioritizes jobs based on business value, time criticality, risk reduction, opportunity enabler and job size. If the added value depends on other factors the WSJF does not always suffice, as it assumes that relative independent values are known. In this blog I will describe two of these situations. Alternatively, I will demonstrate the use of the economic theory to make a prioritization decision based on the expected outcome (business value). I shall provide a guideline for product owners for identifying expected outcomes and prioritizing based on those.

The great escape

Imagine you are in the Bayview Correctional Facility in New York City for a bank robbery (you did commit). You are bored and want to escape. You plan to escape to the haunted house of your deceased aunt in Beacon, where no one will find you. You asked a friend to park a car right outside the prison for your escape. Sadly, folding paperclips in prison provides you with limited monetary resources. You can only afford enough gas in the car to make it to the haunted house in Beacon on your own.

You plan to escape through the window in your cell. Unfortunately, you share a cell with Mister X. You’re pretty sure that he will call the guards if he sees or suspects your escape. You could include him in your plan and escape together. As Mister X is 2.10 meter tall and quite heavy, the car would  use more fuel if he joins. Consequently, you won’t make it all the way to the safe house.

You are faced with a difficult decision:

A. Escape on your own, but Mister X will probably tell the guard.   Though, once you’re out you will make it to Beacon unseen. B. Escape with Mister X. You (and Mister X) have to walk to the safe house for an hour. It is likely you will get caught during your stroll.

Though not as dramatic, Product Owners are often faced with equally difficult choices. As using resources for realizing one user story costs you (at least a part) of realizing another user story. It’s up to the Product Owner to decide which user story (functionality) adds most value to the product. In this case the WSJF does not suffice, as job size, business value (live a happy and free life), time criticality, risk reduction and opportunity enablers are very similar for each option. Another way to prioritize is required.

General Economic Theory suggests looking at the highest value of the expected outcome to help make decisions if the outcome of a situation is dependent on other factors. How likely is it that you escape in case A and B? For this you need to know: 1) what are the chances that Mister X will tell on the guards, 2) how far do you have to walk if you escape with Mister X? and what are the odds you will get caught while walking.

Assume, the probability that Mister X will tell on you is 55% (he doesn’t really like you). So if you escape on your own, the chance that you will live happy and free is 45%. If Mister X joins you on your break to freedom, he will not tell the guards. Unfortunately, the car will only get you as far as 2/3 of the way. The two of you have to walk for an hour, the probability that someone will see you is about is 35%. Logically, you are more likely to live happy and free if you take Mister X with you (even though you do not really like him).

Making money

A second example is less gruesome, but perhaps a more realistic choice people face:  should I save or invest my money? Say you have some savings (€10.000,-) and you wonder if you should invest all of it in a startup or create a savings account. Needless to say, if the startup is successful the pay offs are higher (1.3%) than the interest on your savings account (interest 0.6%). However, where pay offs from your savings account are guaranteed, this does not hold for an investment in a startup. You have to compare the expected pay-off of investing to the pay-offs of saving all of it. This implies that you have to take uncertainty into account.

After one year, you would get 60 euros in interest if you choose to save and a 130 euro dividend if you invest your money. However, you have to take the uncertainty aspect of investing in a startup into account. Let’s assume that you will always get your investment back whether the startup survives or not and that the survival rate of startup is 50%. This implies that you should expect to receive only a 50% return on your investment which equals 65 euro. This is 5 euros more than you would get when saving the money. Obviously, if the two  assumptions are true it is better to invest your savings in a startup.

These two examples show how to best make a choice if you are dealing with an uncertain outcome and dependencies. If it is not possible to compare the added value of user stories (epics or features) objectively, I advise you to investigate the expected business value or expected effect on the business value.

You can identify the expected (business) value using the following 6 steps:

Step 1. Define the uncertain factor (e.g. will Mister X tell the guards, survival rate of startups). Step 2. Identify the scenarios of the uncertain factor (45% that Mister X won’t tell, survival rate of startups is 50%). Step 3. Analyze the pay offs (generated business value) of each of the scenarios. Step 4. Identify the most beneficial scenario (highest pay off) of two user stories (Don’t tell Mister X, do tell Mister X). Step 5. Take the uncertainty factor into account and compare pay offs (highest business value). Step 6. Choose the user story with the highest expected business value.

You might never end up in prison and of course you will always invest wisely. Though, if you have difficulties choosing between user stories as Product Owner, ask yourself: what would Al Capone or Warren Buffet do?