Using economics to inform product delivery
A common topic that comes up in our team is how to deliver our product road map. The standard question we are facing is the choice of building for speed vs building for future.
Build for speed: Quick implementation in a relatively narrow way.
Build for future: More general implementation, that often comes at higher immediate cost but is easier to extend in the future.
To illustrate this, imagine you have two ways to deliver a feature. A quick implementation, i..e building for speed, will come at a cost of say $5,000. A more general implementation, i.e. building for future, can deliver the same feature but at a cost of $10,000. You have future plans for extending the feature. Delivering this on top of building for future would cost you another $5,000, but building for speed would require major changes which would cost additional $12,000. You can see the summary of the two options in the table below.
We can start by comparing the total cost of the two options.
TotalCost = Cost(now) + Cost(future)
The total cost of building for speed is higher, but much more of the cost is in the future. This is a standard question of cost now vs. cost in the future, and fortunately economics offers a few useful concepts to help is understand how to make such a choice.
Introducing discount factor
Economics knows the concept of net present value. When people are given the choice of $100 today vs. $100 in the future, almost all will choose the money today. This is because money in the future is worth less today. Just how much less depends on the discount factor, which is applied to future income. A 10% discount factor means that $100 today is worth 10% more a year from now.
You can use the same concept to think about the net present cost of implementing a feature, where we discount the future costs as they are worth less in today’s terms.
NetPresentCost = Cost(now) + Cost(future)/(1+DiscountFactor)
While building for speed has a higher total cost, the graph below illustrates that as the discount factor rises, building for speed becomes less expensive relative to building for future.
This is because building for speed has higher future costs, but these aren’t worth as much in today’s terms.
Adding opportunity cost
There is another key input to take into account — the value of the features we could be building now instead of the additional immediate cost of building for future. In economics this is known as the opportunity cost — the cost of giving up something. Adding this to our model, we get:
NetPresentCost = Cost(now) + Cost(opportunity) + Cost(future)/(1+DiscountFactor)
Fixing the discount factor, the graph below shows that as opportunity costs rises, building for future becomes relatively more expensive.
This is because higher opportunity cost means we are giving up more value that we could be building instead of the higher upfront cost of building for future.
The uncertainty of the future
Product road maps are not set in stone, and there is always a chance the features we think we will be building never see the light of the day. We can model this by assigning a probability to the future features. A 60% probability means that there is a 40% chance we will not end up building the features building for future tried to address from the beginning. We can use this information to calculate the expected cost of both options.
Expected(NetPresentCost) = Cost(now) + Cost(opportunity) + FutureFeatureProbability * Cost(future)/(1+DiscountFactor)
The graph below shows that the more uncertainty there is about the future (i.e. the lower the probability of the future features), the more expensive building for future is relative to building for speed.
This is because the lower the probability of the future features, the more wasteful it is to be preparing for them now.
Unfortunately, it is almost impossible to correctly estimate the inputs for the above equation. But the aim of this article is to explain what a team should be taking into consideration when making decisions about product delivery. This is particularly important for early stage products, which usually have the following characteristics:
High discount factors: Getting capital through fundraising is expensive, both in terms of time and equity. It is not unusual for high growth start-up valuations to increase 2–3x from one round to another.
High opportunity costs: Building the third feature usually drives more relative value than building the 101st feature. Sacrificing new features in the short term thus has high cost.
High uncertainty about future: Product road maps is a constant evolution.
Therefore, teams building early stage products should usually opt to build for speed. As the product matures, this will slowly shift towards optimizing for total costs.
