User Avatar

I love making important things better, while glimpsing at the patterns that connect

You preparing to build a new feature ... and you don't want to waste time and money with no impact
Steve Sanderson @stevesanderson

Context

You've got a new feature you're preparing to build. You're convinced there's a problem worth solving.

Problem

You're uncertain about the demand, and you don't want to waste time and money building something that won't have an impact

Solution

Offer the 1st step of your solution and put them on a waiting list

  1. Rough design the user-facing aspects of the feature

  2. Implement the first point of interaction with the user

    • log when a user views the where the first point of interaction occurs

    • log when a user interacts with the first point of interaction

  3. Implement a response to their interaction by notifying the user that this feature is coming soon

If you want to know more about their needs, consider

  • prompting them to describe their needs

If you're concerned about the consequences of not delivering the benefit, consider

  • revising the language to shift away from offering the feature and towards telling them about the possible feature, e.g. instead of "get report on usage" revise it to say "find out more about usage reporting"

  • prompting them to let you know if the need is urgent, and then fullfill their need with a concierge MVP

If you're concerned about the whether they're window shopping or actually have a need, consider

  • prompting them to actively choose to tell you they're interested, e.g. "yes I'm interested" and "no thank you" buttons

capabilities you'll need

  • to decide ahead of time what success looks like, e.g. what % of people who view the first point of interaction eventually interact?

  • to view and analyze the event stream (i.e. the views, the interactions, any optional entries, responses, etc.)

  • to immediately turn off the experiment

  • to serve any concierge MVP work that comes out of the experiment

management systems you'll need

  • team monitors and adapts/decides based on the results of the event stream

  • terminates the experiment when done

  • track and deliver any concierge MVP work

Caution - even the narrowest experiment may test other variables

  • in this case, if you don't have a proven, predictable, method for customers to discover a new feature then you are also testing how customers become aware of a new feature in your product.