A Simple Way To Model ROI of Any New Feature
When evaluating a potential new product, feature or enhancement there are dozens of factors a product manager must take into consideration. At a high level these key factors can be broken down into three main categories — desirability (why does the target user want/need this?), feasibility (can our organization make this happen?) and viability (will this help the business and is it sustainable?).
While assessing the viability and eventually pitching an idea I’ve found it extremely helpful to go through the process of modeling out how this will impact the business. It helps to evaluate tradeoffs, opportunity costs and how many resources should be allocated to the project. Of course, any model or projection is fraught with unknowns, assumptions and downright BS. But the exercise itself is incredibly useful, especially when still considering and forming the idea.
These models and projections will vary greatly from product to product and business to business but this simple formula will help you evaluate these product decisions and assist in pitching your idea to your leadership and stakeholders at the early stages.
The ROI calculation
(reach of users * new or incremental usage * value to business) — development cost = expected ROI
Reach of users is an estimate of the number of users who will see, use or be impacted by the feature.
New/incremental usage is an estimate of the change in behavior or primary action being driven as a result of the project.
Value to the business is the ultimate outcome you want to measure. This is often measured in averages or percentages and could be anything from LTV, conversion rate, renewal rate, etc.
It helps to create different outcome scenarios for the models. I will typically use a bad, ok, and great versions of the model based on usage or margins to help provide guidance and benchmarks.
The ROI model in action
Here’s a real-world example of how this calculation can be applied when evaluating a potential development project.
Let’s say you are the PM of an e-commerce site evaluating ideas to increase sales. One idea the team has is to use re-targeting emails to attempt to recapture sales from users who add an item to their cart but leave without completing the purchase.
Your data shows that on average 50,000 users per week add at least one item to their cart but never check out. The average value of those items in the cart is $10. Knowing this information we can create three different models based on projected conversion rates from the recapture emails. These models would look something like this:
Bad Conversion
(50,000 users add item to cart and abandon * 0.5% conversion rate * $10 avg. per cart) = $2,500/week
Ok conversion
(50,000 users add item to cart and abandon * 2% conversion rate * $10 avg. per cart) = $10,000/week
Good conversion
(50,000 users add item to cart and abandon * 4% conversion rate * $10 avg. per cart) = $20,000/week
Don’t forget to factor in development cost
Nothing comes free in software development. Any project or feature that takes time from the development team has a cost associated. Most organizations will be able to estimate the development cost of the average team by sprint or month. In the example above let’s say the team estimates that the work will take 1 sprint (2 weeks) to complete. If our organization projects that 1 development sprint costs roughly $20,000 then we can use that to factor in the expected ‘payback’ time for any feature or project.
