Aiming for a New Target: The Perfectly Featured Product presents an argument against featuritis (and provides some explanations as to how those situations arise).

Wengel model 1699 - your product nightmare

The author (John Devanney) suggests that UX Designers keep heed of the KISS principle. Which, though aspirational, is not particularly informative or helpful.

How can the design team determine the ideal feature mix?

Is there an established methodology for this early phase of product development? What UX techniques could be employed to assist choosing which features to bundle into a product, and what are the pros and cons of these techniques?

4 Answers 4


Perhaps the best advice I've heard in this area is "the design team is wrong" - they can't do it in isolation.

Instead, determine the Minimum Viable Product - the version with the smallest possible feature set that's of actual use, release that and act on customer feedback as you iterate.

[This isn't to say that you only use customer feedback, but that you can't get any until you release and that it's more valuable than you might think.]

  • Minimum Viable Product is more of a business concept than a UX concept though, ya? Commented Mar 10, 2012 at 8:14
  • The MVP is the goal, I like that .. any suggestions on how to discover what features go into that?
    – Erics
    Commented Mar 10, 2012 at 8:29
  • @DesignerGuy - I don't think the two are independent. A UX concept that doesn't consider the business is (at best) an expensive flight of fancy; a business plan that doesn't consider UX is (nearly) doomed.
    – Bevan
    Commented Mar 10, 2012 at 8:41
  • @Erics One approach that I've seen work: Start by brainstorming all the features you'd like. Trim the list down to just 10 items by deferring everything you can. From that list of must have features, pick 3.
    – Bevan
    Commented Mar 10, 2012 at 8:46
  • Chasing links from the MVP wikipedia entry led me to this comment elsewhere: using multivariate testing to determine MVP. An interesting approach - any experience with that?
    – Erics
    Commented Mar 10, 2012 at 9:00

My thoughts-

  1. Facilitation games like Monopoly or else provide each participants a certain set of "dollars" where they would invest in capable features that are good for the product. You can skin out important features through this.

  2. Ask participants/stakeholders to do Dot Voting on each of the features - to find where the major share lies and to prioritize the options.

  3. Do a survey or conduct audience poll to find whether internal team and external team match with the options or features available.

  4. Best way also could be to start with a rapid prototyping with quick dispensable ideas/feature list and test it in wild. This needs time and also provoke good insights to rejig the feature matrix and align with majority of users expectations.

Look for judgemental analysis such as "feature creep" does not dissolve or sink the product in totality.


A note upfront: I assume 'early phase of product development' in the question refers to the development of new features into an existing product. The referenced blog post is also primarily about products in later stage of their product lifecycle. The stage where you've got a product that hopefully got traction and has active users. Some of the users demand new features. Probably the most active users, the ones you really like. Their needs might be special but since you value them you add the features. At the same time the product gets more complex and harder to learn for new users. Ask yourself: Can the number of power users promoting your product compensate for the number of users that sign up and never get active due to the complexity?

A common way is to 'ask' users during user research. While I think it's important to listen to customers it can also kill a product at the same time, if you kind of misunderstand. It's crucial to find the right balance between maintaining the current (core) product and adding new features to it. This balance needs to be analyzed and tailored for each product. I can't think of a general rule of thumb but I have some thoughts:

  • start from the product you've got
  • watch real users using your product and look into the usage numbers
  • what features do they value most?
  • iterate on these features first to make sure the UX is outstanding
  • do not (never!) neglect your core product!
  • listen to users (not only the loudest ones) -> what else do they want?
  • look at similar products and competitors
  • ask: would users pay (extra) if this feature was there?
  • prototype a new feature and show it to real users
  • do they just like it or do they love it?
  • decide whether you need to iterate on the prototype or implement it
  • consider implementing it for a small group first
  • based on the feedback, kill it or keep iterating on it
  • give the new feature a seat in the back (i.e. not in the main nav)
  • do promote it to a front row seat until the new feature has proven to be great

UX techniques: Besides the standard techniques for testing a feature wireframe or prototype and collecting feedback from users I'm not really sure if there is a particular UX technique for prioritizing features. Assuming there is a bunch of potential features one could evaluate various aspects in a matrix, like - fit with core product - fit with general product vision/strategy - demand/amount of user feedback - revenue potential

A note on Minimum Viable Product: I think the MVP concept is great for startups looking to create a new product from scratch, and for companies that already took the decision to add a particular feature now defining the scope of it.


The first rule is keep it simple.

If you only need one feature to make a successful product, why overdo it?

David McClure from 500 Startups said it this way:

"Features are like having sex. You make one mistake and you have to support it for life."

As far as the mix of features, I think you just have to keep your product/service centered around the user's needs.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.