Saam Motamedi (General Partner at Greylock)

Saam Motamedi on AI and ML, advice for students, and more.

Sep 28, 2020

If you are interested in reading more interviews, just like this one, with founders, operators, and investors from today’s leading companies / startups, feel free to subscribe below!

Guest Profile:

Interview Guest: Saam Motamedi (Twitter: @SaamMotamedi)

Role: General Partner at Greylock Partners

Previous: Prior to joining Greylock, Saam founded Guru Labs, a machine learning-driven fintech startup, and previously worked in product management at RelateIQ, one of the first applied AI software companies. There, Saam drove the development of data products before and through the company’s acquisition by Salesforce.

Quick Note: This interview was recorded via a Zoom call between Saam and Roshan (that's me) in mid-August.

Roshan’s favorite quotes from the interview:

On investment decisions at Greylock: “If you look over our history, some of our best investments have actually been the most controversial ones.”
On the future impact of AI and ML: “I'd say I don't think there's an industry that AI and ML will not have a big impact on.”
On COVID’s changes to the VC/tech ecosystem: “When I meet new entrepreneurs today, and on the boards that I sit on, it's not just, "is the value proposition still relevant in a post COVID world?" It's really, "how do we make our value proposition even more relevant in a post COVID world?”
Advice for aspiring student founders: “Take your time. When I went to Stanford, and I certainly think this is true at universities across the US, there is this almost anxiety around, "I need to start a company, and I need to start a company as quickly as possible." I think what ends up happening is people pick suboptimal things to go work on because they just want to start a company.”
Advice for aspiring venture capitalists: “There are a lot of firms, and there are a lot of investors at those firms, and as such, the business is more competitive than it's ever been. So if you're thinking about building a career in venture capital, then the question becomes, "how do you set yourself up to be successful when you are working in venture capital?”

Roshan: Thanks for coming on, Saam. To start, can you discuss how you came to join Greylock and what type of companies you focus on?

Saam: Thanks for having me! I've been one of the Partners here at Greylock for the last four years. I had the good fortune of getting to work on the startup side before joining Greylock. My first job in technology was working as a product manager at a company called RelateIQ, which back in the 2013-2014 window was building and bringing to market one of the first intelligent CRM companies.
I had a chance to join them and work on a number of data products. That company was ultimately acquired by Salesforce. Parts of that product became what's now called Salesforce Einstein, which is Salesforce's ML platform. I left Salesforce with a few friends from RelateIQ and we started a company called Guru Labs, which was building offline commerce automation software. Through those two journeys, I got very exposed to the early days of startup building. I fell in love with the first part of the journey where you go from an insight around a customer problem or pain point, and some new technical solution, and then the whole process around product productization, landing your first set of customers and then scaling go-to-market.
I'd gotten to know folks at Greylock while working in different companies. There was an opportunity for me to join the team and bring that skill set to our enterprise investing practice, so I did that in 2016 and haven't looked back since.
To the second part of your question, what types of companies I focus on. At Greylock, we're primarily focused on enterprise and consumer software. We have a team of folks on the enterprise side and a team on the consumer side. On the enterprise side, we think of the stack in four buckets: applications, data and app services, infrastructure, and security. We flex up and down the stack; all of us on the enterprise team have knowledge and interest in different areas of the stack. I invest in security companies, companies doing AI and ML (either applications or infrastructure), and then SaaS application companies. I'd say there are a couple of themes that companies I've invested in share.
One is companies that want to target the large enterprise and build software that, over time, can sell into, say, largest 20,000 accounts globally. The second is companies that are either using ML or data as a core part of the product advantage to go after an existing or new application or security market, or infrastructure companies that are making it easier to stand up AI and ML projects at large enterprises. The third is just from a stage perspective, we have the flexibility to move up and down stages. I love being the first investor of a company very early, either at the seed or Series A stage.

Roshan: Can you talk about the most recent investment you were involved in, and what made you so excited about it?

Saam: The most recent one that we've announced publicly is a company called Snorkel, (Website: Snorkel.ai). To take a step back, I think one of the largest trends of our generation in software is the transformation of software and business workflows with AI and ML. Every enterprise in the world across every function will find ways to use AI and ML to automate, augment, and make different processes more effective. The question is, how do you make it possible for these enterprises to get AI and ML products up and running? What we began to learn as we talked to customers in 2017, 2018, 2019 is, on the one hand, people were accelerating their investment in AI and ML. Now, on the other hand, they're dissatisfied with what they have to show for that investment.
Despite a lot of advancements happening in underlying algorithmic techniques, and the availability and decreasing cost to compute, it's still very difficult for people to get high-quality machine learning running in production workloads. So then you ask, "why, what's the blocker?" and the thing we kept hearing was that it all comes back to the training data: the data that you actually feed into these machine learning models, which they learn from and then make a decision based off of. So we got very attracted to the idea of, "can we find a company to partner with in the machine learning infrastructure space that understands that we need a data-centric approach to solving the problem?"
We had the good fortune of meeting the Snorkel team in early 2019. At the time, they were at Stanford, where they had worked on the Snorkel open source project at the Stanford AI Lab for four years. Essentially, when you build a new machine learning model, one of the very early steps in the pipeline is that you have to label your training data sets. Conventionally, this is done with manual labeling effort. You have a team of labelers who go through, look at data samples, and label them.
The Snorkel team had invented a new paradigm for labeling, which they call programmatic labeling or programmatic training data management. The core insight here was instead of labeling each sample by hand, domain experts can express labeling logic as code. Then, Snorkel can take that labeling logic, which is often called labeling functions, and programmatically label the data set. So you can complete labeling tasks much more quickly, much more cheaply, and with higher quality than relying on manual labelers.
So they'd built this open source project. This project had gotten significant adoption at leading organizations like Google and Stanford on the healthcare side. They realized through this process that there is not just an opportunity to change labeling — there's an opportunity to change the entire ML workflow and everything that happens after the label. So they spun out of Stanford to start Snorkel AI, which has brought to market a product called Snorkel Flow, which you can think of as an end-to-end platform that's data-centric, solves the training data problem, but actually orchestrates the entire machine learning pipeline. We were lucky to partner with the company at the seed and then again at the Series A several months later. They now have a scaling business around the data-centric end to end ML platform.

Roshan: What does the internal investment decision-making process look like at Greylock?

Saam: At Greylock, we have nine Investment Partners. Those Investment Partners lead investments, and they have the ability to lead investments across areas and across stages. Though, again, different partners focus on specific domains. So, for example, I focus on enterprise, my partner Josh focuses on FinTech, and my partner Reid does a lot of marketplaces. We are very much a collaborative team, and we make decisions collectively, but at the same time, we also allow for non-consensus-driven investments. We certainly make investment decisions where there are disagreements among the partnership.
If I intersect a new company, let’s say you've started a new company and I intersect you, and I get excited about what you're doing, the first thing I'm going to do is invite a few of my Partners, those closest to that domain, to meet with you. We then collectively will pressure-test the company, both advocating the upside, “why this company can be massive and market-defining?”, while also considering the downside, “what are some of the company's core risks?”
We're looking to do our best to answer some of the questions around the core risks as quickly as possible. For an enterprise company, that could include doing customer references, talking to customers who are using your product, and introducing you to new customers in our network –– who both can be good leads for you and helpful points of view for us. That's part one.
Part two is, who are the experts in this market? What's their point of view on what you're doing? If you're starting a new application security company, who are the people in our network who really understand that market, and can we go get their point of view on what you're doing? That helps us answer the product-market fit question. The next piece that's really important is, how do we understand you as an entrepreneur and your team? Again, there are two parts to that. One is, let's spend a bunch of time together, both just getting to know one another, but also testing different areas around alignment on company building. How do you want to build out the team from here? How do you want to build out go-to-market? Then we’ll also do some referencing, and let you do some referencing on us.
We'll talk to people who you've worked with in the past. Then we try to get to conviction as a sub-team, that this is something that Greylock should be a part of and should partner with. That typically means two or three Partners in addition to the sponsoring Partner having excitement around the investment opportunity.
Then the last step, after all that diligence has been done, is a full Partner Meeting, where you come in and present to the entire Greylock partnership. After the meeting, we discuss, and we make a decision. I'd say on the investments we end up making, there are at least a few very supportive partners. There may be some who weigh the risks more heavily and have questions, and that's okay. We can make those investments.
If you look over our history, some of our best investments have actually been the most controversial ones. But you see both investments get made when everyone thinks it's a great idea, and similarly, investments made where there are strong advocates and strong detractors.

Roshan: Looking ahead, what areas/industries do you think AI and ML can have the most significant impact on?

Saam: I'd say I don't think there's an industry that AI and ML will not have a big impact on. If you think about large enterprises and just go function by function, let's take sales, we're seeing a number of companies that are reinventing sales and CRM workflows using AI/ML.
Whether it's predicting the probability that a deal is going to close based on looking at the sentiment in the conversations you're having with the prospect, to recommending to a sales rep what's the next best action he or she should take to maximize the probability the deal closes; you're seeing total reinvention there. If you look in the back office, say finance teams, you again see AI and ML automating things like expense report auditing and procurement workflows. And then, if you look at specific verticals, take healthcare, on the diagnostic side you're seeing new companies get started that use AI and ML techniques to augment physicians in making diagnostic decisions. Both across work functions, and across verticals, AI/ML is going to impact every sector and area.

Roshan: What impact has COVID-19 had on the venture capital and startup ecosystem?

Saam: It's a good question, and it's obviously something that's on top of all of our minds. I'd say a lot for us actually still looks the same for us. There are new entrepreneurs that we have the privilege of meeting every week. Today we meet them over zoom instead of in our offices. But they've identified important problems and are starting new companies, of which some will go on to be very significant. So I think the fundamental mechanics of our job around meeting entrepreneurs, picking the best markets, earning the right to make investments, and then helping build companies — a lot of that actually looks very similar.
It's worth noting that when we invest in companies, we invest very early, and we invest with very long term time horizons. Many of my partners have been on the boards of companies for more than a decade. So we're less thinking about "what's going to happen next quarter, the quarter after, or even the next year?" We're more thinking about, "is this company going to be on the right side of history? Do they have the chance to define a market that ends up mattering?" So I'd say at the highest level, not a lot has changed.
Now in the tactics of the job, of course things have changed, and the biggest is just the way we work every week. We're now running Greylock entirely in a distributed way on Zoom. We run our weekly meetings on Zoom, when we meet entrepreneurs, we meet them on zoom. We can certainly take socially distanced walks as we try to get to know a founder and make an investment decision. Still, I'd say the default interaction modality is Zoom. So that's certainly one thing that's changed. I think there are positives in that. We spend less time in cars traversing the Bay Area, and so we have more time on our calendars and can meet more people. We're less restricted to the immediate network. We can meet with a more diverse set of entrepreneurs because people can come in to us from any direction, different segments, and different geographies.
I'd say the other core thing that we think about is given that COVID has impacted everything, what are the companies that will benefit from COVID? When I meet new entrepreneurs today, and on the boards that I sit on, it's not just, "is the value proposition still relevant in a post COVID world?" It's really, "how do we make our value proposition even more relevant in a post COVID world?" As an example, we're investors in a company called Abnormal Security that's focused on next generation email and communication security. The Abnormal product is much more important post COVID, because in a world where you have distributed teams and people working from home, all your critical communication is happening on digital channels. As such, the integrity and security of that communication has paramount importance. We have some newer projects still in stealth that are all about — now that you have engineering and product teams working remotely — the tooling you need so that they're just as productive as when they're in the office. In summary, we certainly have that lens when we evaluate companies, around a pandemic proof value prop that's going to be even more important given the new working and living conditions post-pandemic.

Roshan: Last question before we move into some advice related questions, what is your favorite part of your role at Greylock?

Saam: That's a hard question. I think it's a very privileged role, and there are many aspects that I really like, but I'd say there's nothing I like more than rolling my sleeves up and just working with entrepreneurs really, really early. It's the early sessions when there's no idea. There's a whiteboard, some wraps from a Mediterranean restaurant, and we're just ideating and thinking through early product concepts and early product-market fit. That's the type of work we love doing. And, you know, I wish I could spend more time just doing that.

Roshan: Now moving into some advice related questions for our followers, most of whom are college students, either undergrad or grad, what advice would you give to a current student interested in becoming a founder?

Saam: I have three primary pieces of advice. One is to take your time. When I went to Stanford, and I certainly think this is true at universities across the US, there is this almost anxiety around, "I need to start a company, and I need to start a company as quickly as possible." I think what ends up happening is people pick suboptimal things to go work on because they just want to start a company. That can be suboptimal either in the market they picked or in the team they assembled to build that company. I think we're in an environment today, where there's a lot of capital — venture capitalists fund lots of companies — and candidly, many of these companies will not amount to being very successful in the long run. So what you don't want to do as an entrepreneur is, as a student, graduating and starting a company that you work on for 5, 6, 7, 8 years, but that doesn't ultimately get to becoming a successful company. So I'd say, first, take your time and realize you don't have to start it right out of school.
Then the question is, "what should you go do?" and that ties into the next two pieces of advice, which is to go find your network and tribe of people that you want to go build with. I think the best way to do that is to join an early-stage company that is just hitting product-market fit and early momentum. There are two reasons to do this. Number one is that those companies attract excellent talent. The people you go work with will end up being co-founders, employees, people you work with down the road. The second is that the reason you go join that company versus a very large company is the DNA of the people who have elected to work in early-stage companies is inherently more entrepreneurial. So the likelihood of finding people who are going to want to be your co-founder two or three years from now is much higher. So I'd say that's the second piece of advice.
The last piece is right now, you have a particular set of skills, and you should think through, "when I start my own company, what are the set of skills I need that I don't have today and how do I fill those gaps?" Often I find for younger founders, a lot of this is around go-to-market, which is, “I can build products and recruit my friends who can help me build products, but then I need to match that against a really quality market opportunity. Then I need to do learning around how to map an initial idea and technical insight with a 1.0 product that I can deliver to five customers and make those five customers successful. How do I go build out a sales function and a marketing function that can scale up from five to 50 customers?” So on and so forth.

Roshan: What do you think are the most important traits and skills of founders?

Saam: I think there are a couple of dimensions. One is having a fundamental market insight, which is whether you're going after an existing market that has incumbents or creating a new category, you have to have an insight into what the customer's problem is and how you can architect and deliver a superior solution. The second is a learner's mindset. We very much invest in slope and not y-intercept. The best founders I work with are just learning animals. They're learning from their customers about the product to build. They're learning from their teammates. They're learning from people outside the company who are best in class at running different functions on improving themselves operationally. The third thing we look for is customer-centricity. Company founders who want to build highly customer-centric organizations focused on solving customer problems.
I'd add two other things. One is iteration speed. Let's say you're starting a new company, and it's in a good market. At that point, the determinant of success is how quickly you can iterate on a product based on customer feedback, deliver that software to the market, and then scale your go-to-market execution? The final thing we look for is storytelling ability in recruiting. When you're building a company, you're constantly selling, and you're persuading different sets of stakeholders to buy into your mission. One set is customers, one set is certainly investors, but in many ways, the most important set is future hires and recruits. So we ask ourselves the question, "is this person able to tell a story that's compelling to those different stakeholders, and are they a recruiting magnet? When people meet them and hear their story, do they want to go work for them?" When someone has that set of criteria, we get very excited.

Roshan: What advice would you give to a current student who's interested in going into venture capital?

Saam: We live in an era where there are a lot of venture capitalists. There are a lot of firms, and there are a lot of investors at those firms, and as such, the business is more competitive than it's ever been. So if you're thinking about building a career in venture capital, then the question becomes, "how do you set yourself up to be successful when you are working in venture capital?" There are different things you can do to differentiate yourself. One example is building a fundamentally unique network. So if you go become a part of the next Stripe or the next Airbnb or the next Facebook, you're going to end up working with a bunch of people who will go on to start great companies. As a venture capitalist, that's a very valuable network to have. That’s one form of something unique. The second form is getting really good at a specific dimension of company building. Whether it's founding a SaaS business and scaling that business and then learning a bunch of things that you can help the next generation of SaaS entrepreneurs, or exploring, joining an existing company that's killing it in product management function, and just becoming world-class at that. Then someone will choose to take money from you because you're world-class at that and you can help them with that function. I'd say you want one of those two, either network or functional differentiation. I think less about, "how do I go become a venture capitalist right away?" and more, "how do I build the skill set and network so that when I become a venture capitalist, I can be hyper-successful?"

Roshan: So last few questions, a little more fun. Outside of work, what hobbies occupy most time, and how do you stay physically and mentally fit given the high demands of your job?

Saam: I could certainly be doing a better job. There are three things I could call out — two of them are very COVID related. One is, spending a lot of time home, so more cooking time. I used to be the person who never cooked anything, and I actually find cooking is a hyper-meditative activity. We try new recipes every week, especially on the weekends, and we just set up a pizza oven and have been cooking fresh pizzas. It's a really fun thing; you get to build something from scratch, hopefully it tastes good, and it's a nice way to take your mind off other things. I got a Peloton when the quarantine started, and I'm a huge believer and convert now. And then I love playing tennis.

Roshan: Last question, do you have any favorite books, movies, podcasts, etc. that have been a big influence in your life as a whole or in your work?

Saam: There are many, one I would call out is Thinking in Bets by Annie Duke. Annie is just a phenomenal thinker and also a phenomenal poker player, and… it looks like you're smiling. Did you read the book?

Roshan: I have the book sitting right next to me.

Saam: Nice, yeah, so you may know, but she's written this phenomenal book about how to think about decision making, sizing bets appropriately relative to the risk, and then improving the process of decision-making by staying away from resulting and looking at outcomes, and more looking at process. I think it's really important for any investor or operator. That book has certainly been quite foundational for me. One I've been reading more recently is a book called The Outsiders, which tells the story of eight different CEOs who have just been phenomenal capital allocators, and compounded shareholder value. It's a really good book because it clarifies and reminds me, and I think the CEOs out there, that a core part of the job as CEO is being a capital allocator and thinking through where you put resources and what the yield on those resources investment is. So those are two books that have been pretty fundamental for my thinking, and I highly recommend both of them.

Roshan: Thanks Saam!

The Takeoff

Discussion about this post