Feature Value = Unit Impact x Multipliers +/- Alternatives
There are many ways to prioritize features. See this article for 7 ways to prioritize features including Kano, RICE, priority scorecard, etc. Here, I suggest a mechanism for PMs to rationale Feature Value, input into the overall prioritization process.
Context: As a PM, I sometimes need to rationalize feature importance without actually putting the actual prioritization for my product team.
Why? First, I want to scale myself beyond one single product. If the entire product team depend on me as a PM to make every single feature prioritization, I would be the bottleneck of the team. Second, I may have more junior PMs in the team, so I need to find a way to give inputs to their product prioritization via a strategy discussion, without actually do the prioritization for them.
Solution: Feature Value is a formula that captures product strategy. Since it should reflect a product strategy, there is no right or wrong formula. A product team needs to come together to rationalize the formula together and continue to refine the formula throughout the product life cycle.
Let’s start with a base formula:
Feature Value = Unit Impact x Multipliers +/- Alternatives
- Feature Value: a numeric value of the importance of the feature to its users. The higher the value, the more important the feature. PM/UX should base on their user research (quantitative/qualitative) to judge feature value. The engineering team can give an estimation of the level of effort to implement the feature. Final feature prioritization takes into account Feature Value and Level of Effort.
- Unit Impact: only positive value as a single unit measurement of impacts. If you are working on improving the productivity of a developer tool, feature value is time-saving before and after having the feature on an interaction with the tool. You can assign discrete value to it. For e.g., 3 is high value (save 3–5 min per interaction), 2 is the medium value (save 1 min per interaction), 1 is the low value (save less than 10s per interaction).
- Multipliers: only positive value — There are multiple types of multipliers, I suggest 2 types for my products. Scope multipliers are the numbers of users this feature will impact. New user multipliers are how this feature will impact new vs all users. Usually, only Scope multipliers are enough. However, for specific products where customers might start to develop features complacent, you will find new user multipliers a refreshing way to improve product development.
- Alternatives: can be either positive or negative. If there is an alternative to a feature, it can be both helpful and harmful.
When you actually do want your customers to take these alternatives (at least in the short-term) so that you can focus on something more important, give this alternative a negative value. It will reduce the overall value of the feature.
When alternatives are actually competitors or things that you do not want customers to use, assign a positive value to alternatives. It will increase the overall value of the feature.
Application example: Let’s start with a simple formula for a B2B product in its early growth phase — the product is still acquiring new users but have a stable customer base.
To apply this Feature Value framework, here are the steps:
- List down all user problems and their corresponding solutions/features
- Align a formula with your product team
- Assign a score for each feature and compute the Feature Value of each feature
- Don’t take Feature Value too literally (Feature Value of 5 is not that much different from 4), but use your judgment to group features into big buckets and prioritize accordingly.
Step 2 requires strategic thinking to ensure you align your product strategy to feature prioritization. Once Step 2 is done, step 1 and 3 can be pretty mechanical.
Most of the discussion should be in the last step. Feature Value is only a tool. Eventually, PM needs to use judgment to see if the output of this model fits Product Sense.
For example, for my product, I am worried that this product is becoming too complex: a Photoshop of its kind, so I want to use new user multipliers to search for features that existing users might not value but new users would find extremely important. Using new users multipliers, many usability features, and simplification of the product would raise to the top bucket of prioritization. So, the model works out per my expectation.
