The Founder of Facebook Marketplace Says Customers Will Tell You What Your Product Is—If You’ll Listen

If you can turn your business into a platform, you probably should. There are amazing multi-billion-dollar products out there, but any one product has limits. Companies that run platforms, by contrast, cultivate ecosystems that make it possible for other businesses and many products to thrive. That is a much bigger market.

Deb Liu knows a ton about platforms. It took years of advocacy and strategy, but she literally invented Facebook Marketplace and ran Facebook’s entire platform group, helping monetize the different ways users wanted to leverage the network. Now she’s founded an AI startup helping small businesses go from zero to fully automated.

Like platforms, Deb is multifaceted, so this conversation also goes deep on payments, the trust gaps that have to be filled to make online transactions possible, and some of the differences between running public and private companies.

The transcript has been edited for clarity.

Lionel Foster: Deb, welcome to Catalyst. Thanks for joining us.

Deb Liu: Thanks for inviting me.

Lionel Foster: All right. Well, you’ve had a pretty amazing career. Actually, I have no idea how old you are, but you just don’t seem old enough to have done all the stuff that is on your LinkedIn profile. So I think this will be a lot of fun.

I know you have a new startup of your own. We will get into that. You know loads about the payment space, the digital payment space. You know a lot about AI, and man, are those big topics.

So let’s start with some of your history at Facebook. When I hear Facebook—and I know you were a leader within platform, commerce, and other parts of the business—how do you even think about platform at a place like Facebook? Because potentially it could mean so much. I’m curious how you all defined that and disciplined yourselves in what to focus on.

Deb Liu: Well, it’s interesting. Basically anything with any scale should be a platform.

Because if you allow people to—people use Facebook groups for their own purposes, for example—and it became a platform even though that was not the original intention.

Think about people building businesses on Instagram or WhatsApp or Facebook. You think about what is possible and, any place where there are people, there are people who want to use it to get distribution and customers.

So if you’re a consumer product that doesn’t allow people to build on it, you’re actually missing something.

What Mark did early on was launch Facebook Platform to allow developers to build applications on top of Facebook. This was back in the desktop days. Then on top of that, having APIs for things like Facebook Login, which was an authentication system, and sharing back to the site.

I encourage every consumer company, especially consumer applications, to think about platform strategy. Think about how powerful Shopify is by making it possible for anyone to build applications and share them with other shops. Think about Intuit having a platform where people can build things that might not even be on Intuit’s roadmap.

I think more tech companies, especially B-to-small-business or B-to-consumer companies, should really think about themselves through a platform strategy lens.

Lionel Foster: Is there such a thing as thinking about that too early in the lifecycle of a startup?

Deb Liu: Well, it depends. You need enough users to make it worthwhile for someone else to build.

Success in a platform is that other people are successful on your platform. That’s how we measured it. How many people found success using our platform? Could they build businesses on it?

Ads platforms were the same way. We had agencies that were effectively homegrown digital agencies that emerged because they cracked Facebook advertising in a way traditional agencies took longer to do.

But you need enough scale that people can sustainably build businesses and enough opportunity that investing time into your platform makes sense, even if it’s not venture-scale.

Lionel Foster: Listening between the lines, it sounds like part of this is vision. You need a vision for what the ecosystem could become. But another part is timing and scale.

Deb Liu: Well, it’s funny. When I joined PayPal, I was already an eBay seller, so payments felt like a natural fit. I had been selling on eBay for a long time, and people would literally mail checks and money orders. Then PayPal came along with referral bonuses—$10 referral bonuses, $5 referral bonuses—and suddenly eBay sellers started accepting PayPal as a payment method.

The instinct inside PayPal, from what I heard, was initially: “We need to shut this down. What is this eBay thing?” But it turned out to become their biggest and most profitable business. They were asking, “What are these people doing? This isn’t what we built this for.” Because originally they had ideas like beaming money and other use cases. eBay wasn’t the intended one.

But it became the success story and eventually drove the acquisition by eBay. The lesson is what we call people-led development. Customers vote with their feet, and if you don’t listen, it’s at your own peril.

Lionel Foster: Yeah. Let’s drill down into consumer digital payment behavior because I think you’ve been in the industry long enough to see the early consumer internet.

My recollection is that people used to stop and ask: “Do I really want to put my credit card online? Is this safe? Is this site trustworthy?” We’re miles beyond that now. Could you talk about that evolution?

Deb Liu: E-commerce is impossible without payments infrastructure. We talk all the time about online commerce, but without payment rails none of it works. Are people going to mail checks? Send money orders? Digital payments enabled online commerce.

Early on people were suspicious. Do I save my credit card? What are the chargeback policies? What happens if somebody goes rogue with my card? What if the item isn’t what was promised?

Whenever there’s a trust gap, somebody has to step in and say, “I’m going to take care of you.” That’s how people cross the chasm. Early adopters try everything. But then you hit the next group saying, “I don’t know if I trust putting my card online. What if someone drains my account?”

That’s where trust systems mattered. Is there an intermediary? Is there protection?

If you look internationally, China was very different. Early systems leaned heavily into escrow because trust levels were lower. The mentality was, “We trust nobody, so control the money.”

In the US, credit cards already had chargeback rules, so there was some built-in protection. But somebody still had to build fraud systems and trust mechanisms. That innovation is easy to forget now, but it was foundational.

At PayPal, the joke was that we weren’t really a payments company: we were a risk management company and a fraud reduction company. If we couldn’t manage fraud, the whole business disappeared.

You’re taking maybe 3% on a transaction, but if fraud exceeds certain thresholds your economics collapse because you’re exposed to losses approaching 100%.

Lionel Foster: You’re effectively underwriting every transaction.

Deb Liu: Exactly. Every single transaction.

You had to trust that your systems could identify patterns faster than fraudsters could exploit them. Phishing, account takeovers, fraud detection: that was the secret sauce of online payments.

And it mattered even more in two-sided marketplaces. Macy’s selling directly is one thing. Credit card companies know Macy’s isn’t defrauding customers.

But in a marketplace, you can have seller fraud, buyer fraud, collusion fraud, and then third-party fraud layered on top through hacked accounts. It’s a multi-headed beast. Solving that problem enabled marketplaces that still thrive today.

Lionel Foster: You hit on something I underappreciated until now. Someone had to fill the trust gap so digital payments could scale. PayPal was one example. Who else helped make that turn?

Deb Liu: Intermediated marketplaces mattered enormously. People needed somewhere to go when things went wrong.

Amazon is a great example. Originally Amazon just warehoused books and sold them directly. It was basically Macy’s for books. But then they opened Amazon Marketplace and let anyone sell. Same with eBay.

Those models were incredibly powerful because they enabled both payments and fulfillment. Payment infrastructure plus shipping infrastructure—that was the huge unlock for e-commerce.

Lionel Foster: Another thing that fascinates me is how payments have changed what counts as money.

Take Starbucks. There are huge balances sitting in Starbucks accounts. Consumers preload value, Starbucks holds it, earns interest, and consumers don’t get interest back.

I don’t drink coffee, so Starbucks isn’t my example. Amazon is. If I have Amazon credit, it’s effectively cash to me because I use Amazon constantly.

Camber Creek portfolio company Bilt Rewards operates at massive scale with loyalty systems and points. I’m fascinated by the idea that if enough consumers trust something, it doesn’t need to be dollars: it functions like money.

Deb Liu: I don’t drink coffee either, by the way. I get tea. But Starbucks absolutely nailed something.

They built loyalty in a way others haven’t replicated. Imagine if there were a Starbucks-style app for every major chain. McDonald’s and others have apps, but they haven’t matched the combination of points, stored value, ordering, and loyalty.

Years ago we even explored ideas like that: what if you had a Starbucks-style system across your favorite restaurants?

Starbucks locked in customer behavior. You preload balances, collect stars, receive gifts, get promotions. Suddenly people feel compelled to return. They reportedly have billions sitting on their balance sheet in prepaid balances and gift cards.

Run a promotion—5% off, extra drinks, bonus stars—and you lock customers into coming back. It’s incredibly powerful.

I’m surprised more brands haven’t replicated it. Even companies with strong apps haven’t matched the combination of loyalty, stored value, and behavior design.

Lionel Foster: We can workshop this in real time because there’s a company called inKind in restaurants. Do you know them?

Deb Liu: Absolutely. I use it all the time.

Lionel Foster: Great. So I’m just amazed at the way—so their CEO, I know he spent some time here. I live in the Washington, DC area. He pitched Camber Creek. We’re real estate technology, but hospitality has a strong real estate component, of course. We did not invest, but that was my introduction to the company.

Part of how they work is they have this national network of restaurants, and you, the consumer, can load credit on InKind, and they’ll guarantee you a certain percentage bonus above that credit, so your spending power is, say, 1.25x or 1.33x the dollars you actually put in.

The way that darn thing has locked me into spending in their network because I want to use those bonus dollars is amazing. They’ve got me. And part of how it works is the deal InKind makes with partner restaurants is, “Hey, we will give you, say, $100,000, but let us treat it like that’s $150,000 worth of spending power.” So they have negotiated what seems to me to be a pretty nice arbitrage.

All right, so that’s the model. How does that line up with what you’re thinking about is possible with payments and rewards points? Is that in line with what you’re thinking?

Deb Liu: Well, it is and it isn’t. First, I use inKind. I really like it. I’ve used it at several restaurants, and I just loaded up credit because I’m going to New York and there are a couple restaurants I saw on their list. So I’m very excited to use it. Free advertising for that brand.

But the bigger thing is what they do is they are actually driving demand to local restaurants, especially those that can’t sustain their own ecosystems. Most restaurants, at least in our area, are independent. You can’t have an app for every restaurant you go to.

Lionel Foster: That’s right.

Deb Liu: If you’re Domino’s or Shake Shack or somebody you visit more than once a month, sure, do an app. Starbucks does it. My kids like boba and we go to the same couple boba places, so yes, absolutely do an app there. But for independent restaurants, that isn’t sustainable.

What they’re doing is harnessing aggregation, the aggregation of independent restaurants, so you have collective spending power.

I was introduced to InKind by a friend. He said, “Costco has the gift card. You should get some.” So I picked some up. It was on sale at Costco. And now I’m locked in because I got maybe 20–25% off. Now I have to spend it. They bought my loyalty.

So they’re buying demand from me, and now they have to match supply. At first there was only one restaurant in our area, which happened to be the restaurant my friend had taken me to, but I liked it and now go pretty regularly. More recently they added ten times as many restaurants. Clearly they’re expanding because there are lots of people like me looking for more options—higher-end restaurants, different experiences.

What they’re doing is classic marketplace supply and demand. You buy demand first. Once there’s all this stored balance out there, people are looking for exits. Then you tell the supply side: “Hey, I have all these people waiting.”

Lionel Foster: Yes.

Deb Liu: “You need to come because I’m going to drive demand.”

And restaurants already spend 20–30% on marketing to get people through the door. This becomes another marketing expense. If you categorize it as marketing rather than meal economics, you think about it differently.

Lionel Foster: I love this. And yes, neither of us has any formal relationship with InKind outside of using the platform. But when someone is doing something smart and it’s working, why not use them as a case study? Some of the best moments on Catalyst are when we have industry-leading experts like you reacting to examples in real time. This is red meat. I queued that up for you. I’m glad we talked about it. I learned something, and I think our listeners will too.

So you’re a startup leader now, but you have roots in publicly traded companies. You’ve got all these ideas and probably more opportunities to roll out products than you know what to do with. Is life different? Are your decisions and what influences them different in a public-company context versus startups or private companies?

Deb Liu: I’m on both a public company board and a private company board, and you can see the difference in how people lead. I also ran Ancestry, which was private-equity owned. I was at Facebook before and after IPO. I was at PayPal after it was public and inside eBay. So I’ve seen a lot of versions of this.

There’s tremendous pressure at public companies. I remember at eBay there was a point where they said, “We need to figure out how to spend X dollars in the next two weeks.” Everybody was like, “What?” But there are ways to manage earnings. You don’t want to overshoot because then next year you have to lap it.

Or somebody says, “Top-line growth is soft. Can we put more into marketing?” There are dials in ad businesses. You can literally change how many ads people see. I’ve heard about this across companies. There are levers you can pull because you’re managing operations and expectations simultaneously.

Lionel Foster: Yeah.

Deb Liu: Public markets react to tiny signals. Right now someone announces great earnings and their stock still falls. Cloudflare did well recently and the stock dropped. I’m still trying to understand why. People are constantly reading tea leaves.

One of the advantages of private companies is time. You can invest in things that may take years rather than quarters. You can take different risks.

So when joining a company, you need to understand the risk profile. Have you planted enough seeds? Different initiatives operate on different time horizons. I’ve seen people come into companies promising turnarounds and I’m like: do you know how long it takes just to hire people, let alone get them up to speed? Three months in and people say, “Why hasn’t this turned around?” It’s incredibly hard.

Before I joined Ancestry I was offered some turnaround CEO jobs, and everyone warned me: you need two years but you get six months before people write you off. Ancestry was on a growth trajectory, so it was different.

The same issue exists for startups too. Do you invest in the product people want for the future or chase revenue today? Some companies rush revenue and neglect product. Others build forever and run out of runway. Some products genuinely take years. Even with AI, all the pieces still have to fit together. So are you chasing short-term revenue, perfection, or solving a real need?

Lionel Foster: You mentioned Ancestry, which is fantastic. I want to do this in two phases because both matter. First, you mentioned Ancestry and that you had other offers. How did you evaluate the options? Why did Ancestry win?

Deb Liu: People think decisions work like college applications: you apply everywhere, get answers at once, compare options, then choose. Life isn’t like that. Jobs are more like dating.

Lionel Foster: Depending on who you are, but yes.

Deb Liu: Outside of The Bachelor, you meet someone and ask: “Is this where I want to stop? Plant roots? Grow a garden?” Or do I keep looking? Jobs are like that too.

You might be happy where you are and suddenly an opportunity appears. You compare it against today and also against every unknown future opportunity you can’t see yet. Humans are bad at that. Dating apps have this issue. People always think someone better is around the corner.

Back to Ancestry: it was a forty-year-old company, and I loved what they were doing. Helping people tell their family stories and preserve family history matters. If we don’t know where we come from, how do we move forward? I loved the mission and the people. It was an amazing four years and I’m still very close to them.

Lionel Foster: Nice. That’s interesting because I assumed there might have been some strategic reason I couldn’t guess. But helping people preserve some of the most important stories of their lives, that’s hugely compelling.

Another thing that occurs to me: you just seem to think and explain things very systematically, very fluidly. When I come across that, I notice it. So I went to law school, and some of the best thinkers and performers in law school happened to be engineers, because they could dissect everything and turn what felt like a messy word problem into something that looked more like math.

Where does that tendency come from for you?

Deb Liu: Well, I am an engineer. I have a civil engineering degree, and I wanted to be an architect, so I took architectural engineering classes.

I really think more people should think in systems. I am not a very intuitive thinker. I’m very much a systems thinker. I actually question my own intuition.

One of the things that’s important is that systems make sense if you look at them the right way. Everybody has a superpower, a thing they do that seems so easy to them but is very hard for other people.

My kids love these YouTube videos where someone tastes a cheap ice cream and an expensive ice cream and explains the difference. They’ll say, “This one is $15 for a tiny container, this one is a half gallon for four dollars,” and then explain all the nuances. They do this with all kinds of foods. My kids ask, “Is that real?”

And I tell them that when you get so good at something, it becomes almost intuitive. You understand subtleties other people don’t see.

Lionel Foster: Mm-hmm.

Deb Liu: For some people it’s taste. Think The Devil Wears Prada. I have very simple clothing taste, but clearly there are people who understand cuts and lengths and all these things I barely comprehend.

So what is that superpower for you?

For me, it’s seeing something and seeing the through line in it. That’s why I was in strategy consulting. I can see a problem, envision a solution, and run the through line through it.

When I advise startups, people often tell me, “You explained it better than I could explain it to myself.” Part of that is because I have distance. But the other part is that I can see the throughline.

People are actually very bad at identifying their own superpowers, because they’re so good at them they assume everyone else is too. I always encourage people to ask: what comes easily to you that everyone else is baffled by?

The people who recognize it are often not the ones closest to you, but the next ring out. They’re the people saying, “Lionel is amazing at X. I don’t know how he does it.” Whatever that X is, that’s probably your superpower.

Lionel Foster: Yeah. I’ve heard that framing before, and it was a huge unlock for me. I was like, “Oh wow.” It’s like asking a fish what the water is like.

Deb Liu: Exactly. The fish says, “What’s water?” Because everyone is in it.

I really encourage people to identify and understand their superpowers because workplaces often focus on the opposite. We tell people, “Here are the three things you need to fix.”

You’re too aggressive. Not strategic enough. Not technical enough. Not enough of a team player. Whatever it is.

The world would be better if we also let people play to their strengths. If someone is exceptional at X, why don’t we unleash them on X? Why force them into Y?

How do we create environments where people can amplify their strengths—not destructively, but constructively? Let them mentor others in the thing they do best instead of constantly criticizing everything else.

Lionel Foster: Yeah. You said a few minutes ago, “I question my own intuition.” That sounds incredibly wise. It sounds like something a behavioral scientist would say. But I’ve literally never heard anyone else say it that way.

Why do you do that? How did you learn that? It sounds very healthy and productive.

Deb Liu: I think your first inclination is not always your best inclination. So I believe in interrogating where you are. Humans have very visceral reactions to things.

One of my kids has very strong reactions to everything. I tell them, “Before you say anything, count to ten.” Just breathe.

Whenever something happens, I’ll say, “Go to your room. Take a shower.” You’ll come out with much more clarity after cooling down.

One thing that’s dangerous in the workplace is how reactive everything is. On a Zoom call, you’re processing the presentation, the discussion, the notes, so many things simultaneously that people have no time to think.

Lionel Foster: Yep.

Deb Liu: Then the person who speaks first and fastest dominates the conversation.

One of the best product managers I ever worked with slowed that process down in an interesting way. We had executive reviews, and she would present but never answer questions. People would ask something and she would just sit there, so I had to answer.

People asked, “Why doesn’t she say anything? Is she not good?”

Finally I told her, “You have to answer. It’s your presentation. I’m just your manager.”

And she said, “I’m a processor. By the time I’ve processed and know what I want to say, the moment has passed.”

Lionel Foster: Wow.

Deb Liu: And I realized we were missing out on her brilliance because she was somebody who takes her time. Everyone else just fills the gap.

One of the things I did—and I did this with leadership teams at Meta and later at Ancestry—was require everyone to send a pre-read. Not only that, but people had to vote ahead of time on what their decision would be and write it down. Then we put it in a shared document where you could technically read everyone else’s responses, but we’d white them out by turning the font white.

We did this because it slowed people down and forced them to think. Too often you’re trying to make a decision and it’s the first time you’ve heard about X. You have no context, and suddenly you’re making a major decision. I think it’s dangerous to do that without taking time to think and process.

Lionel Foster: Yeah. That’s great. Tell us about Ember AI.

Deb Liu: Yeah. So we started Ember. I had stepped down from Ancestry. I had breast cancer last year. My mom had passed away. I just needed a break after four wonderful years. I thought, “I’m going to take the rest of the year off.”

Lionel Foster: Well, let me just say—it’s really good to see you upright and happy.

Deb Liu: Thank you. I’m doing great. Really blessed.

But I thought, the next revolution is AI. We had brought some of it into Ancestry and explored a lot, but I wanted to understand the technology deeply, so I took a boot camp class. Suddenly I realized: this is incredible.

At the same time, I think we in Silicon Valley live in a bubble. I went to dinner recently with people from San Francisco and everyone was talking about agents. One person said, “I don’t go on a walk without three agents running so I have work done when I get back.” Another was building agents for different workflows.

I remember thinking: this is not what the rest of the world sounds like. We’re inside this bubble where everyone feels behind because they learned one AI stack and now there’s another one.

I realized AI is amazing, but it’s also like a kid in a candy store. You can overdo it. As I learned AI, I started asking: what if we brought AI to the people furthest away from it? That’s where the value would be greatest. Silicon Valley doesn’t need more AI. They’re already building everything. But what about companies that aren’t even on the cloud yet?

You could help them agentify workflows, think through AI strategy, modernize operations. That’s what we do. We help enterprises think through AI strategy and build applications that smooth workflows and modernize operations.

It’s incredible what’s possible now. Things that previously would have cost millions can now be built, maintained, and amortized over years.

Our philosophy is not replacing people. Every job has 20–30% of work people wish would disappear. That’s what we target. We ask: what parts of this job should AI do so humans can focus on things only humans do well: taste, judgment, creativity, customer relationships, complex problem-solving?

People say AI will do everything. But that’s like saying Excel would replace CFOs. It didn’t. It gave them tools. AI should make jobs richer.

Instead of moving rows between spreadsheets, reading PDFs, cutting and pasting, synthesizing repetitive information, what if AI did 80% of that and humans focused on final judgment and feedback? That’s where the magic is.

Lionel Foster: Yeah. I want to touch on a few things. First, you left Ancestry, AI was accelerating, and you took a boot camp class. You, a longtime technology executive, basically went back to school for this topic. That stands out to me because some people might be too proud to do that, but you were curious and made it happen.

Deb Liu: Look, we’re only about three years into generative AI. Generative AI and LLMs really hit the mainstream in November 2022. So think about that. Three years into the internet revolution was 1998.

What companies existed then? Which hadn’t been founded yet? Who was in the S&P 500? What did the world look like?

Or think three years after the iPhone platform launched. We’re just at the beginning. Massive technology shifts require people to rethink and retrain. I was part of the internet revolution and mobile. This is the next one.

Everyone is anxious, but this is a thirty-year cycle. There will be winners and losers. Think about Pets.com. The ideas weren’t wrong. The evolution just took time. Eventually stronger companies emerged—Instacart, Chewy. The ideas survived.

Lionel Foster: You mentioned Ember works with companies that may not have engineering teams and could otherwise be late AI adopters. Putting on my VC hat, I immediately think: onboarding. How much work is that? How labor-intensive is it?

Deb Liu: Our ICP, ideal customer profile, is companies with few or no software engineers. They have IT teams, but no engineers.

It is more work, but the opportunity is bigger because they haven’t wired everything up already. The gap can close much faster now.

There’s this sci-fi story my husband watched where one generational ship leaves for another planet. Sixty years later, technology improves and another ship overtakes it because it waited. That’s what’s happening now. Companies that lagged can leapfrog.

We work with private equity firms and VCs, and what we’re seeing is that it used to be “big eats small,” then “small eats big.” I think the next phase is “fast eats slow.”

The companies that adapt fastest will take share, grow revenue, and become much more efficient. Those who learn and adopt intelligently will unlock huge amounts of human capital.

Lionel Foster: You advise a lot of leaders. You’re very sharp on the how—how to solve problems and think about solutions. But there’s also the why. Leaders lose touch with why they started. Do you help people reconnect to that? Because you seem very connected to your own why.

Deb Liu: Before I advise founders, I try to understand what brought them there.

VCs often ask about the origin story because it matters. That’s the why. What was the burning platform? Why this problem? Why spend five or ten years of your life on it?

That’s the rocket fuel.

Lionel Foster: Nice. Is Ember VC-backed?

Deb Liu: Yes. We just raised a seed round.

Lionel Foster: Congratulations.

Deb Liu: Thank you.

Lionel Foster: Wow, we covered a lot. I’m nearly breathless trying to keep up, which is a good thing. Before we wrap, anything else you’d like to cover?

Deb Liu: No, I think this was a really broad conversation. We actually went deep into some areas, which I enjoyed. I love geeking out on payments, marketplaces, and business models because it’s fascinating to understand how companies attack problems creatively.

This is where humans matter. If you asked ChatGPT years ago whether grocery delivery could work, maybe it would have said no. But humans create the future by approaching problems differently.

AI should enable human creativity. I’m optimistic because I think it frees us to spend time on the things that matter.

Lionel Foster: Fantastic. Thank you so much, Deb.

Deb Liu: Thank you.