How do you find new product features to build?

Michael Ludden / November, 2017


Michael Ludden is the Director of Product at IBM’s Watson Developer Labs & AR/VR Labs. Previous to that he worked with product and developer marketing at Google, Samsung and at HTC. In this video, he explains IBM’s methodology in developing new products.

Video transcript:

How do you find new product features to build?

The process by which we vet products is pretty simple one from the start. Everything we do we put through the prism of is this good for developers, does it solve a real problem for them that wasn’t solvable adequately before. Is IBM uniquely situated to solve that problem using our technology versus the competition in the marketplace?

If it passes those tests and we think it’s a good idea we kind of slide it into our backlog or to our queue. Then we do some technical feasibility vetting with our engineering team. We have a very large, because IBM is a large company, cross-functional stakeholder list (or I do) of people that either me or somebody on my team goes to insanity checks essentially; are we in the right track, is this something that you find useful for your products etc., and try to look at it from really across IBM perspective.

Even from just a sort of platform-agnostic perspective, is this a problem that developers need to be solved and what developers are the target audience. For us, are they python developers, are they R or are they swift or is there something that we’re not focusing on as much.

In terms of prioritizing features the method that we’re going about with that is really sort of around there’s a metaphor that people probably heard, a cupcake and then like something larger and more elaborate, like a wedding cake – so a cupcake, a birthday cake, and a wedding cake, something like that. So you ship a cupcake, but you have an idea for what a bigger larger better version would be right, those will usually not make it into the launch.

So, we will build an MVP focused on delivering what is essential in the first edition and iterating on that. Those features which we probably have already had in the back of our minds go into a backlog and then they get sorted and prioritized based on our strategic priorities, the organizations largest strategic priorities, and then the companies. Like we might be going after, this is just an example, maybe we’re going after FinTech this year. So something perhaps about auto insurance, a feature that would be great for a chatbot. Maybe that doesn’t get prioritized ahead of a feature that pertains to FinTech for the next version of the launch.



Next Up

What is the best way to prioritize a list of product features?

How often should I ask for customer feedback in the early stage of my startup?

How do you stay on top of your product roadmap?

How should startups think about their mobile strategy?

What have been your three major challenges in building an AI startup?

Join the Startupedia Community

Connect with VCs and unlock vital startup advice!