Becoming a Machine Learning Product Manager with Opendoor Director of Product

This week, Product School hosted Sam Stone, Director of Product at Opendoor, for a special #AskMeAnything session. Sam has some great insights on how to enter into machine learning Product Management.

About Sam

Sam Stone, Director of Product at Opendoor

Sam is passionate about building products at the intersection of finance and machine learning.

He is currently the Director of Product for the Pricing and Machine Learning Products at Opendoor, a late-stage startup that uses algorithms to buy and sell homes instantly, saving homeowners the hassle and uncertainty of listing their homes and hosting open houses.

Prior to Opendoor, he was a co-founder and Product Manager at Ansaro, a SaaS startup using data science and machine learning to help companies improve hiring decisions. Earlier in his career, Sam worked at genomics software startup SolveBio, investment firm TPG, and consulting firm Bain & Company. Sam holds degrees in Math and International Relations from Stanford and an MBA from Harvard.

What’s your advice to non-tech PMs joining the ML space?

First, learn about the technology side of the industry. Take a CS 101 course and/or a statistics or ML 101 course. Get a sense of whether you find the tech side energizing or draining. ML is definitely a trendy topic these days, but like any trendy topic, that can lead people into it who actually aren’t that passionate about the fundamentals. As a result, some may find the work draining and eventually churn out. 

person sitting outside on steps with a laptop on their knees

If you find you do like the technical side, think about how you can marry user research/insight and technical knowledge. There’s a temptation to treat lots of user problems as “datasets” that can be solved by black-box models (for example, “let’s just throw in a lot of data, and the model will figure out the best solution”). This approach has serious limitations, but if you can figure out ways to gain low N, qualitative—but deep—user insight, and then use that to guide model building and testing, that’s a powerful combination. 

Also, think a lot about data quality and coverage. If you have a great data science team, but bad data, that data science team isn’t going to be able to do much.

Read next: Cracking the AI/ML Product Manager Interview

You co-founded Ansaro, and have been a part of other startups. How helpful has your Harvard MBA been in startup journeys and at your current role? 

Great question. I like wrestling with (reading about, discussing, writing about) tough business and leadership problems. For that reason, I really enjoyed earning my MBA while I was in school. However, the learning I got from those two years has to be compared against the learning opportunity from spending two years working in the industry. And the amount you can learn in two years at an early-stage startup or high-growth company is pretty incredible and invaluable. 

Much of what I learned in my MBA is very high-level and focused on becoming a c-level executive. It’s not as applicable in my current day-to-day. I certainly don’t think an MBA is a requirement or even that much of an advantage for being a PM or a Senior PM.

During an Senior PM interview, what questions do YOU ask of the interviewer?

First, I’d make sure to understand the “logistics” of the role. For example, what’s the org structure, what’s the functional team composition today, what will it likely be in 1 year? I’d also ask about major wins the team has shipped. This will be helpful to see if those types of wins and/or projects excite you. Lastly, I suggest focusing a lot on who your manager is and if you’re truly, deeply excited to work with this person. I find the old saying very true: “people don’t leave jobs, they leave managers they don’t like.”

two people sitting at a table with coffee and notebooks. they are looking at each other and one person is writing on a digital tablet

What resources/books do you recommend?

Here are some of my favorite recs, which I share with all the new PMs on my team:

Product-Related Books

Product Management-Related Blog Posts

How do you go about improving your skills as a storyteller?

person holding open book entitled A Storytelling Workbook

First, lead with the answer. That means, if there’s a key decision to be made, make it clear from the start what your recommendation is. Then zoom out to the details. Second, combine quantitative and qualitative insights. In most groups, different audience members will respond to different styles, so have a bit of both.

How do you go about gaining more influence in the C suite?

Influence (or “managing up”) has a lot to do with building trust. Building trust with C suite executives stems from demonstrating:

  1. An understanding of business needs, and
  2. The ability to execute consistently and meaningfully against them.

You have to make some bold bets and have a decently high winning percentage. Small bets are still important, and typically look good to peers and more junior folks, but C suite execs tend to think bigger. At the same time you have to keep up eng/DS/design velocity as a company scales and inherently gets more complex. And that’s something where “paying down tech debt” is extremely valuable—and the C suite generally needs the education to fund resourcing/time there.

Read next: Presenting to Executives by Google Product Leader

What is your take on involving internal stakeholders on the details of a B2B product initiative pre-development?

I think broad pre-launch involvement of all stakeholders is critical. I believe the key question is HOW you involve different stakeholders (not whether you involve them). With technical stakeholders (eng, data science, and even design), I try to share the enough-but-not-overwhelming amount of business context while already completing much of the pre-work with business stakeholders. This way, what I’m sharing is relatively clear. 

With business stakeholders, it’s a bit of the opposite. I want to make sure they understand the end result, or “what they’ll get/see” when the product ships, but not go too deep into the technical elements of how we’re going to get there. However, the information should still go deep enough that they understand the reasoning behind the timeline to ship it. Many business stakeholders assume projects will be easier or faster to ship than in reality, so it’s important to give them the context to understand why a project may take three months and not three weeks.

What’s your favorite digital product?

I’m a machine learning and data products PM, so I love a good algorithm. I particularly like TikTok, because the UI is simple, and the algorithm behind it is so good at detecting the nuances of what videos I like and showing me more of those. It definitely leads me to waste a lot of time, but it’s well-designed for that purpose!

two teens crouched on the floor and looking at a phone screen, smiling

How do you approach scaling a product that has a very diverse set of customers and requirements for each?

Good—and tough—question. I don’t think there’s any magic one-size-fits-all solution. One approach I take is the classic ICE framework: impact, coverage (or sometimes c = confidence), effort.

In other words, if we ship a new feature these three components should be addressed: for users that benefit from it, how much will they benefit (impact); how many users will benefit (coverage), and how much time and/or money do we need to invest to ship it (effort). I tend to chronic under-estimates in terms of effort. For example, most projects take 50% or 100% longer to ship than originally scoped. I also skew towards “bolder bets,” or projects that have uncapped or less-capped upside potential. I’d much rather take on one project that has an expected value of $100M and is +/-50%, than two projects that are each $50M and each +/- 20%.

Learn more: The Many Ways to Scale a Product: A Guide for Product Managers

Any final advice?

Here are my parting words of advice for aspiring Product Managers:

  • Work on a product that you believe in as a user. In the short term, it’s possible to do good work on a product that you understand, but aren’t passionate about. However, in the long term, a lack of passion is draining. And if you want to succeed in the long term, you need to find alignment between your passion and the product(s) you’re working on. 
  • Continuously find ways to understand your customers better—that includes both internal and external users. Users are continuously changing and evolving. Even if you understood them yesterday, that doesn’t mean you understand them today!
  • Think long and hard about how to be a good partner to other functions, from engineering and design to data science and operations. Try to “walk a mile” in their shoes on a regular basis. This will help you better understand their perspective and become a better partner.
productcon banner

Enjoyed the article? You may like this too: