Building products is all about allocating scare resources in such a way as to generate the greatest return. So when we’re thinking about what to build, we’re prioritizing in terms of cost and benefit.
Sometimes this is easy — one path is obviously better than the alternatives. But what if you have ties — ideas with similar ROI?
We can borrow a piece of theory from the world of investment management to help us think about this. Here’s the idea.
Investors seek to maximize return, but to do so they must make risky investments. Every possible portfolio of investments comes with a particular combination of risk and expected return. Some portfolio configurations are strictly worse than others — greater risk for the same return, or less return for the same risk. We never want to choose one of those.
If we plot each portfolio on a two-dimensional graph, we can identify all of those portfolios that are strictly inferior to another possible portfolio. We ignore those. They suck. The ones that remain — the ones that are not dominated by any other option — are the valid options. We can draw a line to connect them.
We ignore the bottom bit of the line that starts to curve back towards high risk. What remains is the line called the “efficient frontier”.
The key idea about the efficient frontier is that any portfolio configuration on it is considered equivalent in terms of risk/reward. To choose between options, the investor has to turn to other factors, such as their personal risk tolerance. For example, a younger person might choose to accept larger risk (along with greater expected return), because s/he feels like there’s plenty of time to recover if things don’t work out.
The same thinking can be applied to products or features that a team is thinking about building. Each product has an expected cost and an expected benefit. Products can thus be plotted on a two-dimensional graph.
As with investment portfolios, some products strictly dominate others. Eg, for the same cost, we can get a higher benefit:
…or for the benefit we can pay a lower cost.
Sometimes it’s super obvious which products are good, and which are bad.
But sometimes, it’s super not obvious which product is better.
In the chart above, Product A has low cost but also low benefit, while Product B has high cost but also high benefit. Which one is better?
Here it’s obvious that Product C isn’t a good choice. For marginally more cost, you could get a much higher benefit with Product B. Or for much less cost, and only slightly lower benefit, you could go with Product A.
We can think of the line between the dominant products as the efficient frontier.
Consider a fourth product:
Here, the only product that’s strictly dominated is Product C. The others form an efficient frontier.
How is this helpful?
It’s useful to recognize that there isn’t always going to be just one solution that is the objective best. There will often be a frontier of efficient choices, and deciding between them will require you to turn to other factors.
For exploration in future iterations of this post:
- Create a spreadsheet model for product managers to use
- How to think about benefit:
- benefit to who (the company? the customer?)?
- along what time horizon (short term? very long term?)?
- How to think about uncertainty and error bars?
- Is it true that Product C above is strictly inferior? Why?
- Is convexity along the frontier the thing to be avoided?