Example Product Interview Answer: Next Killer Feature

Editor’s note: in the interview prep post, we reference product questions as a key area to practice for product management interviews. Meka from Sote has done us the kindness of laying out his thought process and answer to a certain product question. Enjoy!

Q: Imagine you were a PM at Google.  If you were to build the next killer feature for them, what would it be?

Background (The Setup)

If I were to answer this, there are a couple different paths you could go.  You could go with a new product line within an existing revenue stream.  But since so many of Google’s revenue streams are mature, it would be hard to imagine a single new feature making a big impact.  You could try to create a new revenue stream from an existing product line, but that is more of a strategic question than a product vision question.  That is why I would look at both: a new product line that creates a new revenue source. But even that is a super open-ended question, so I’m going to try to come up with some constraints that will give my creativity a target to hit. So one lens of looking at this is something that I got from Stratechery: what types of tech business models are capable of generating billions in revenue?

Google has already begun to diversify their product lines across these different business models, but clearly the vast majority of its revenue is generated by ad-supported consumer products – namely Search. So already, I’m thinking one constraint would be to beef up one of the other business model types.

I really like to use interesting intersections as ways to get my creativity going. Another “vector” of constraints that we could consider is something I really liked from the Stratechery article’s suggested reading, Escape Velocity, where he talked about the different horizons.

When I think about Google, I think of Google X, which is providing a lot of bets in Horizon 3. I think of Mothership Google, which has a number of interconnected product lines around its ad business. And while you could argue that Google has a few Horizon 2 bets, my gut feeling is that this is the area that Google needs to strengthen. So far, our target looks like this:

Now we need a way to determine what things might be Horizon 2, and Escape Velocity provides another interesting vector we can use to think about this: category power. Emerging categories are ones that will fall into Horizon 3, Growth categories are ones that haven’t been won but are on the cusp of generating significant returns for the winners and often fall in Horizon 2, and Maturing categories are in market share battles but the pie is no longer growing and fall in Horizon 1. We will want to focus on Growth categories that Google has some core competency in already.  From following Google, I would say they are well known for AI and machine learning (Google Home, Search), Cloud services competing with AWS, and OS’s for IoT (Android Auto, Pixel Watch).  So our target now looks like this:

Looking at that Venn diagram as a guide, the intersection of these three categories are where we’d like to develop that next killer feature.

Goals and Constraints:

So our target is both the goal and the constraint: to enter a growing category in a way that leverages Google’s current strengths, but only if they can make it material to the business in 2 years.


Often personas are used at this stage, helping to put further constraints on the problem you are trying to solve.  But in this case, I think it is more instructive to use combinations of business model and category, which builds off the work we’ve done so far:

Now we can hone these down even further by trying to write a requirement story for a product that sits at each of these intersections.


I will write the story from the perspective of Google, and why it would want to be in an even more specific product space with their core competencies:

A) As Google, I want to enter the AI executive assistant space because our AI technology is best in class, digital assistants is a new and growing product field, and we can bolster our portfolio with more SaaS revenue.

B) As Google, I want to enter the field of logistics support software, because our processing of real-time data is best in class, many big players are increasingly looking into the category of logistics, and it gives us a chance to gain hardware margins possibly mixed with SaaS.

C) As Google, I want to enter the field of commerce computer vision AI. Again, Google is best in class when it comes to AI, there are many small entrants in this field creating a market for it, and it would be a way for us to take a piece of the growing ecommerce market without having to build our own marketplace.

Then, let’s prioritize these:

You can see my thought process here. I care about how much we could satisfy users with each story, how material I think it could actually be to Google, and how difficult it would be to execute. Based on that, I decided applying AI to logistics, despite being the hardest was the biggest opportunity.


Now, let’s look at more specific solutions to the story we created:

  1. One solution would be to build IoT devices and software to achieve the holy grail of real-time fleet management for trucking companies across the globe. There is a serious sales and adoption challenge here. But if you can prove that you can impact margins through real time data and algorithmic optimizations based on that data, there certainly is money to be made as the unified provider of this hardware/software/service.
  2. Another solution would be to extend self-driving car technology to “self-driving freight”. It would be easy to be the leader in realtime fleet data when your OS is in all the cars.
  3. You could extend your Google Auto OS to optimize for freight trucks, a potentially easier way to get some of the real-time location data that you would need to pair with software for fleet management. 
  4. You could build smart chassies, and smart containers, that maintain a virtual 3D understanding of the container, it’s weight, and how much volume is left to be filled.

# 2 is a horizon 3 solution, so I wouldn’t examine that. #4 is my favorite, because it is a combination #1 and #2, plus much more. Way harder to execute, way harder to sell. But if Google could disrupt the container market with smart containers, there’s money to be made from that hardware/software mixture. Logistics a huge market. But, if you could use the data collected to become the most horizontally ubiquitous, real-time data source for how full each container is, where it is, etc., etc. that would be huge. Impacting margins, using data to remove friction from global logistics, that’s huge. I’m not sure if this is actually a horizon 2 thing, logistics is super slow-moving, but damn it’s cool!



One big takeaway from this one is be creative in coming up with visualizations of your thought process. Second, don’t be rigid in using a framework. I completely co-opted the “Background” portion of my framework, normally reserved for asking questions like “who, what, why” to better understand the question. But, the point of the background section is to drill down to a goal, and have material from which to build your personas. I wanted to show some outside of the box thinking, and ultimately I accomplished the same thing: I found a clear goal, and material from which to build personas.