TikTok VP of Product on Turning Video-First Feeds into a Full E-Commerce Platform | David Kaufman | E291

TikTok has officially crossed 200 million monthly active users in the US — but the real story is what they've built underneath the surface.

In this episode, Carlos González de Villaumbrosia, CEO at Product School, sits down with David Kaufman, Vice President of Product at TikTok, to go deep on the platform's transformation into a full end-to-end commerce engine. David breaks down how TikTok is using AI to automate everything from creative generation to affiliate-based promotion — so merchants can upload a catalog and let the platform handle the rest.

If you want to understand where social commerce is heading, this one is essential listening.

What you'll learn:

• Why removing steps from the purchase funnel is the single biggest growth lever for modern product teams.
• How TikTok's generative AI tools allow merchants to produce high-performing video creative at scale without a production team.
• The shift from spontaneous live streams to scheduled commercial events driving $1M per hour in sales.
• How TikTok balances feed, commerce, search, and messaging as four distinct monetization pillars.

Key takeaways:

• The most successful products in the AI era put the user directly at the point of transaction.
• Organic creator content consistently outperforms traditional high-production advertising.
• A long-tail creator ecosystem can promote a 100,000-item catalog automatically through affiliate-based commission.

Credits:
Host: Carlos Gonzalez de Villaumbrosia
Guest: David Kaufman

Social Links:

• Find out more about Product School here
• Follow our Podcast on TikTok here
• Follow Product School on LinkedIn here

Transcript

Carlos González de Villaumbrosia | Product School 00:00:41 

Hey, this is Carlos, CEO at Product School and your host on the Product Podcast. Today's guest is David Kaufman, Vice President of Product at TikTok. TikTok has reached over 200 million monthly active users in the US alone, and with TikTok Shop, the platform has built a full end-to-end e-commerce infrastructure directly inside the app — making it the only social platform to successfully integrate product discovery, creator promotion, fulfillment, and purchase all in one place.

David leads the product and client solutions team, overseeing monetization and commerce. He previously built monetization ecosystems at Facebook and Google. In our conversation, we explore how TikTok is redefining the future of social commerce — why every additional click in the purchase funnel cuts traffic in half, using AI and automation to generate product creative at scale without a production team, how live shopping evolved into a scheduled commercial channel doing a million dollars an hour, driving affiliate-based promotion across a merchant's full catalog through the creator ecosystem, and how intent-based search on TikTok connects to the rise of LLM-driven shopping. Let's get into it. Welcome to the Product Podcast, David.

David Kaufman | TikTok 00:01:44 

Thank you. I'm really excited to be here.

Background and TikTok's Four Pillars

Carlos González de Villaumbrosia | Product School 00:01:47 

I'm excited to have you and to go deep into e-commerce and the future of social commerce — as you are the VP of Product and Head of Monetization for TikTok.

David Kaufman | TikTok 00:02:00 

I lead the product and client solutions team, and it's a pretty exciting time for us. E-commerce is an area we're really focused on, and the whole industry is excited about how we can continue to make advances there.

Carlos González de Villaumbrosia | Product School 00:02:11 

Tell me more about how you developed that expertise and what you do specifically for TikTok.

David Kaufman | TikTok 00:02:17 

I have to go back into my background for a bit. I studied physics originally, worked in venture capital for a while — really trying to get a grip on advertising-based businesses and e-commerce — and then I started a couple of my own companies. One was a performance-based ad network. The other was trying to do distributed e-commerce. Then I focused on e-commerce and advertising through Facebook, worked on monetization ecosystems at Google, and then came to TikTok.

What I was really excited about was four pillars of advertising that were all present in the app at the same time: feed-based experiences, integrated commerce in a way that had never been successful across social media historically, a search-based solution, and messaging and business communications integrated into the platform — especially related to commerce. That space just seemed ripe for innovation. I got really excited to join and focus on it, and the last couple of years have been truly transformational. When you talk to our customers, they used to think of us as the short-form video solution. Now all four of those pillars are established.

Carlos González de Villaumbrosia | Product School 00:03:27 

So you said feed, commerce, search, and DM.

David Kaufman | TikTok 00:03:32 

Yes — direct messaging and communication around products on the platform itself. It takes a couple of forms. Obviously you can message a shop that you find on the platform directly, but the other thing that's really interesting is live shopping. We see a lot of trends where creators are talking about products they're excited about, or businesses are going live and showcasing products. You see a tremendous amount of engagement in the comments — live discussion in an open forum where someone asks, "Does it come in different sizes?" or "Does it work for this use case?" and they get answers from both the host of the livestream and the audience itself. It's a kind of social collaboration around shopping.

The Monetization Philosophy

Carlos González de Villaumbrosia | Product School 00:04:23 

I want to talk about this in context of how you think about monetization. I remember Mark Zuckerberg talking about growing a product to, say, a hundred million users before monetizing — not trying to do it too early. How does that apply to the work you do at TikTok?

David Kaufman | TikTok 00:04:48 

I'm a big believer in that. A lot of folks who worked at Facebook — Meta now — used to say "the revenue takes care of itself." If you do all the right things for your users to make sure they have a deeply engaging, satisfying experience on the platform, and then build a monetization solution that honors that experience, you will find a way to make money.

One of the things I've really enjoyed about working in advertising is that you're balancing all these constituencies — trying to do right by the users, keeping the experience natural, while also generating value for businesses and serving the company's needs at the same time. The way we think about it here is evaluating the advertising experience the same way we'd evaluate the user experience. Are we doing the right thing for the user? Are they engaging with a format they actually find helpful? Are the ads relevant? Can they get enough information about the product or service? Can they interact and see it from a different perspective?

If you give the user that level of information, you get a really engaging product — but you still need guardrails around monetization. We want to make sure we're honoring the user experience, keeping people excited to use the app, and not creating an overwhelming or intrusive advertising environment.

Onboarding New Merchants and Addressing the Scale Problem

Carlos González de Villaumbrosia | Product School 00:06:20 

I want to address the elephant in the room. A lot of companies think, "My customers are not on TikTok." How do you help new vendors find value on the platform and expand the portfolio of solutions?

David Kaufman | TikTok 00:06:43 

The way I think about it is with those four pillars — we should be able to serve every single use case a business might have. It's a jobs-to-be-done framework: what are you using your marketing dollars to achieve? What is your objective?

In a lot of cases, that objective might be reaching an audience they've never reached before. TikTok has a very successful brand with youth culture, but we've expanded aggressively. We have a wide demographic on the platform now — people moving for the first time, people in my demographic who've come on more recently. Even my parents use the platform sometimes. We've reached 200 million monthly active users here in the US.

Once you move past reach, the next thing you might want is engagement. We have solutions that optimize for things like a six-second focused view — we call it Brand. Our creative team has worked very hard to bring out these deeper engagement formats. That matters for measurement: when you're trying to understand the value you're deriving from digital ad spend, you really want to confirm that your message is landing.

After engagement, you start thinking about measurable outcomes that drive a direct business result — sell a product, drive traffic to a website, get an app install, drive deeper app engagement, or collect a lead. That last one is essential in industries like automotive, where local dealers want more information from potential buyers. We want to bring solutions that cover all of it.

What gets me really excited is how many dimensions affect whether an ad is successful. What's your targeting? What's your optimization objective? What automation do you want to apply? This is an area the industry is exploding right now, and we think we're on the leading edge of AI automation-based solutions, especially related to commerce.

If you think about what it would take even just ten years ago to open a store — you needed a website, a product catalog, fulfillment, stock photography of your entire inventory, messaging, an ad ops team managing budget across multiple channels, monitoring campaigns, pausing, adjusting bids, going after new audience demographics. There's a whole array of decisions involved in driving a single business objective like selling a product. We've built all of that into automation.

If you open a TikTok Shop and upload your catalog, we'll handle everything else. You want to sell more products — we can do that. You want creators to promote what's in your store — we can do that too. We bring the creator ecosystem right into the shopping experience, allow creators to promote your products, and handle all the commissions and affiliate mechanics automatically. That's our GMV Max solution.

Carlos González de Villaumbrosia | Product School 00:10:56 

That's an interesting nuance. The evolution I've observed is: first, people had to set up their own websites, then CMS solutions like Shopify made that easier. Then social ads would drive traffic back to that website. But what you're describing is hosting a store on the platform itself — not sending ads outside, but enabling purchases from within. And to do that, you need automation for fulfillment and shipping, essentially building an e-commerce platform inside a social network.

David Kaufman | TikTok 00:11:40 

That's exactly right, and that's what we've done. We have TikTok Shop — if you swipe to the right on the platform, you'll see the Shops tab. It's a catalog across many merchants of products that are available, and people like to browse through it to see what's trending. Then in the feed itself, some of that content comes through. We're seeing a lot of success where the creator ecosystem finds products they're excited about and starts talking about them.

There was a trend probably starting about fifteen years ago where unboxing was really popular — videos on YouTube of people showing you the first-time experience of opening a product, setting it up, reviewing it. That content was popular. Then we got more into people using the products. Creators would show off items — in my case, because cooking is what got me into TikTok, I get a lot of kitchen gadgets. A creator will say, "Hey, I'm trying out this new knife, let me show you what I like about it."

But as a merchant, there was a common misconception I saw as an entrepreneur all the time: people thought the hard part was building the website. Shopify's done a fantastic job making that easy. But the hard part isn't launching the website — it's getting people to care and come to your site. If you have a great brand and wonderful organic strategies and ambassadors, you can build some of that content. But the search landscape is evolving heavily, and if you're using paid media to drive people to your website, you need automation. You need the right creative, the right content, and the right offering. And doing that at scale is really hard. If you sell 2 million items, even getting stock photography of all of them is a formidable challenge.

Our creative team has done a lot of work putting really compelling AI-generated content around product images — adding frames, motion, and in some cases using the creator ecosystem to promote the products. It works really well.

Shortening the Conversion Funnel

David Kaufman | TikTok 00:14:44 

Here's the thing I want to come back to about on-platform commerce: if you look at historical trends on the internet, every time you make the user click again, you lose half the traffic. So if I show you an ad — which is already low-engagement relative to other content — and then you have to click through into the ad, then to the website, then to the product detail page, then to add to cart, then to create an account, then verify your email and password, then check out, then add payment information — the funnel is brutal on new merchants.

One of the easiest ways to optimize that is to put the user right at the transaction. That's something the Shops platform does incredibly well. You discover the product in feed, you click, you're already a user. On your first purchase you add your payment information, and once it's there, it's there for future purchases. You can get that serendipitous discovery and purchase to happen at the same time. From a business's perspective, you've shortened the conversion funnel aggressively. That's something really unique about the platform that we're excited about.

LLM-Driven Shopping and Conversational Commerce

Carlos González de Villaumbrosia | Product School 00:15:53 

I want to talk about the next frontier. Last month Shopify announced an integration with ChatGPT, so users can literally buy things within an LLM. I imagine that's going to continue happening. What's your perspective on how that partners with — or impacts — the business you've built, as some of that traffic may move off TikTok?

David Kaufman | TikTok 00:16:24 

There are a couple of ways to look at it, and there are clear parallels to existing solutions. One is intent. Search is still a massive portion of the advertising industry and a huge driver of commerce. It used to be that you'd type your intent into a search box and get some links. Now you're conversationally describing your intent — sometimes to a search box, sometimes to an agent, sometimes to an LLM — and you get an interpretation layer. In some cases, that experience drills down: what do you actually want? You say you want something for a night out — how formal does it need to be? A restaurant or a formal event? Will you wear this again? The LLM helps refine what you're actually looking for. It's the same as a category-based drill-down on a shopping website, but it accelerates the journey and is much smoother for the user.

We see that same behavior on the platform. There's a lot of search behavior where people are typing things in, and a lot of entity pivoting — once you've described what you want, you watch video and get engaging short-form content with people describing it in detail, giving their perspective. If you don't like that, there are other search terms you can pivot to. You can ask questions directly into the AI and get responses. Conversational commerce is finally having its moment. This was one of the areas I worked on in the business messaging ecosystem at Google, and it's exciting to see LLMs rise and drive this category.

Carlos González de Villaumbrosia | Product School 00:18:09 

I want to dig a little deeper on that. More commerce platforms are deploying their own bots where you can have an interactive conversation before results are shown. How are you thinking about this, especially in a video format where the prompt isn't always text?

David Kaufman | TikTok 00:18:36 

The way we've had a lot of success is around the content itself — recommending products that are relevant for that content. We have a solution called Spark Ads. What it does is take content that's already on the platform and relevant for your business — maybe a user talking about your product completely organically — and allows you to promote it, transforming it into an ad seamlessly and automatically.

That's an area where we're seeing a lot of growth. Because it's organic content, people discover it in multiple ways — through a search term that surfaces it, because they follow a creator who's thematically focused on their area of interest. When you start talking about using an LLM in this context, a lot of it is about digesting and understanding the depth of the product catalog — all the content, descriptions, and categorizations — and making sure we use that intelligence to surface the video that is most relevant to each user. That's an area where we focus quite a lot.

AI-Powered Creative Tools

Carlos González de Villaumbrosia | Product School 00:19:54 

I've heard you mention AI multiple times and I want to learn more about the automation you're describing. Creating videos — let alone videos to sell a product — has a real learning curve. How are you applying AI in ways that let more people automate that process?

David Kaufman | TikTok 00:20:12 

We have two platforms here. The first is how we use AI to help you figure out what to do. If you want help deciding what your script should be, what your ad should look like, we have tools inside TikTok that will generate the script for you. You put in prompts just like you would in an LLM, and you can refine it inside the platform. If you want to use a standard production team with videographers, camera operators, and actors, that helps — we can support the idea-generation side.

We also have a solution called Content Suite, which shows you all the relevant videos for your brand based on the search terms you care about — right on the platform — so you can see what's trending and spark your imagination.

Then there's the actual production side. We have generative AI, so if you don't have a production team, you can type in what you'd like the ad to be and we can generate it completely. We can start from nothing, or start with product images and add motion, animations, and expanded assets.

We've done a lot of internal focus on dogfooding this technology because we want the onboarding experience to be really good. We want small operators to be able to come to the platform, have an idea, and generate an ad without going through the classic production process.

Then there's ads ranking, which is honestly one of the open secrets of AI — it's been using machine learning forever. The terminology evolved from "machine learning" to "AI," but ads ranking is one of the most robust areas AI has ever been applied in the tech industry. We have a long history there, and we continue to apply AI in ranking to make sure we're hitting whatever the business objective is.

So if you come with a product, we can add motion to it, serve it to the right people, and drive that sale — because ultimately we're here to drive revenue. Particularly in commerce, if you're building advertising solutions, what we're trying to do is sell you revenue at a discount and take care of everything else.

Carlos González de Villaumbrosia | Product School 00:22:38 

What are the leading indicators you're paying attention to as a way to track progress toward that ultimate revenue goal?

David Kaufman | TikTok 00:22:49 

As a business, you really want to make sure all of your products are getting the exposure they need. There's a scale problem that's really important here.

One of the things we look at is creative density against the product catalog. Take an advertiser on our platform with a hundred thousand items — are there enough creatives to ensure that any user who might be interested in a given product is going to find something engaging? AI is a great way to provide that kind of scale. Generating compelling creative at that volume would otherwise be formidable.

The second thing we look at is penetration — the depth of that catalog's exposure. Are we actually helping that business sell everything they offer, or are they only selling their three most popular products? How can we drive additional volume and depth across the full catalog?

And then there's engagement with the creator ecosystem. Do they have fans generating content about their products? Are those fans getting the exposure they want? Can the business help by promoting what their fans and creators are making? Those are all ways to apply scale to this problem.

Beyond that, it's about speed. TikTok moves at the speed of culture. If you rewind even five years, you'd see digital advertising campaigns running for an entire year with the same creative, reaching approximately the same audience, generating very predictable ROAS. The industry has evolved well past that. You have to move quickly to stay on top of what's trending, what's driving engagement, and how to drive category expansion and new experiences that get customers excited to come back.

Live Shopping: The Scheduled Commercial Channel

Carlos González de Villaumbrosia | Product School 00:25:11 

That brings me to live shopping — a major trend, especially in Asia. How are you integrating new channels like this to help merchants make the most of the platform?

David Kaufman | TikTok 00:25:29 

Live shopping is a really interesting space. It starts with user behavior. If you rewind a bit, you'd observe people going live and sharing their phone number — "contact me if you're interested." That's a pretty good signal there's a monetization opportunity surrounding the live experience.

So we started building tools that enabled better connections to our Shops infrastructure and advertising infrastructure, so live could be used as a true commercial channel. A couple of observations from that:

Originally, people thought about live as a spontaneous experience. But look at any other platform — live sports, for instance. The big game is not a surprise; everyone knows it's coming and plans for it. Some of the businesses at the forefront of high commercial volume through live started telling their audience in advance: "Next Friday, this creator you love is going live on our platform to talk about our products." They'd send that to their community, put ads behind it to build anticipation, and work with creators who already had large followings on the platform. That amplification effect is powerful — you're working with a creator whose embedded audience is already deeply engaged.

This idea of scheduled lives was really important. For larger businesses, the other key was designing an internal workflow that actually supported it. Typically in the US, there's an approval flow for promotional content — you go through creative studios, get VPs to sign off, get legal approval — and it takes a long time. What we saw is that businesses who were extremely nimble were much more successful. You'd go live, talk about maybe three products, and the one performing best is what you'd start to focus on. You'd shift your ad campaigns to complement the live stream and that product. The result was tremendous volume.

We have examples from Black Friday and Cyber Monday where stores were doing a million dollars an hour off their livestream sales — unbelievable levels of commercial success. I'm personally really excited about it because I remember watching QVC as a kid and being fascinated by the pitchman — and now that's been fully democratized. Any creator can do it, and any business can work with a creator who's excited to capture that audience.

Prioritizing New Revenue Opportunities

Carlos González de Villaumbrosia | Product School 00:28:40 

As you think about expanding channels and opportunities, you're at a scale where the bar is probably a hundred million — maybe a billion — in opportunity. How do you go about identifying and prioritizing those bets?

David Kaufman | TikTok 00:29:06 

In advertising, it's really important to think about the incrementality of the revenue you might drive.

The company has been very focused on expanding the industries and verticals we support. There are some fundamental questions to ask. Let me stick with the automotive example. In the US, auto works on a three-tier system: manufacturers, regional dealer groups, and individual dealers — each with different advertising needs. Manufacturers want to introduce new models and generate excitement. Dealer groups want to drive excitement about a particular brand in, say, the Pacific Northwest. Dealers want leads and foot traffic.

TikTok has its roots in short-form video. We were really good at getting a message out that served manufacturer needs incredibly well. As we think about driving leads and foot traffic, the question becomes: are we going to unlock a new area of budget? And as we think about what to build, that's really important. Are we improving an existing use case — adding value to an investment advertisers are already making with us? That's worth doing. Or are we opening up an entirely new use case that the platform doesn't currently serve well? That's where there's a lot more incrementality in the revenue opportunity. When we benchmark what to do next, those are the kinds of things we look at: is it a new use case, is it incremental, what is the extra advertiser value we can deliver?

Carlos González de Villaumbrosia | Product School 00:31:06 

And the focus remains video-first, I imagine.

David Kaufman | TikTok 00:31:10 

Yes. We're very much a video-first platform — that's our heart and soul. Where we've been really successful is also embracing the creator ecosystem, helping businesses work with creators to produce content people find compelling. We have a major focus this year on commerce and AI automation. But we're going to be video-first. We think it's a compelling, engaging product. That's why users are on the platform, and we want to build solutions that honor that user experience around video.

Matching Creators to Merchants at Scale

Carlos González de Villaumbrosia | Product School 00:31:47 

Another unique element is the creator ecosystem. How do you tap into your community to amplify the message of merchants — especially given that many creators are also active on other platforms?

David Kaufman | TikTok 00:32:10 

There are a couple of approaches. If you're a business that wants to work with a creator community, you could certainly search across social media and the web to find people you think are interesting, send them an email — they all have contact info on their profiles — and form a commercial relationship. That's a lot of work. We want to make it as easy as possible.

We have a platform called TikTok One, which brings the creator ecosystem and the business community together. It makes it easy to find relevant creators, gives you information about their followers, their interests, the communities they serve, and makes contracting easy. We also have solutions where you can put out a spec on the type of content you'd like, and creators can generate videos for you — content that you can then promote on the platform. We want to be the best place for creators, and we want to make TikTok the easiest place for businesses to find creators who are going to drive real business value.

Carlos González de Villaumbrosia | Product School 00:33:14 

When we talk about the best creators, following is the obvious metric — but how do you think about the long tail? Creators who might still provide a ton of value but don't have a massive following yet.

David Kaufman | TikTok 00:33:33 

This is really important because, especially in commerce, we want to be in a scaled game — helping businesses get their entire catalog out there. There are different niches of expertise, different creators relevant for different products.

With our GMV Max solution, which automatically promotes the full TikTok Shop catalog, you can turn on creator promotion. Creators can come to your business, make videos, and earn affiliate commissions based on the products in your catalog that speak to them.

To make it personal: I'm really into cooking, and I've been looking at knife sharpeners recently — it's time to freshen up my kitchen. If I had an account on the platform and I found something I liked, I could choose to promote it and earn a commission. That authenticity matters. If it feels overly promotional, it doesn't land.

So the two approaches together — a platform that connects creators and businesses, and opening up every purchasable product to creator promotion — that solves the scale problem and really brings in the long tail. You don't need to be a top-ten creator to be successful. What helps is following your passion and finding the products you actually use, because that authenticity is really important.

Unlocking Enterprise and B2B on TikTok

Carlos González de Villaumbrosia | Product School 00:35:22 

You know, I never considered myself an influencer or creator, but as we started hosting credible guests on the podcast and running conferences, I started getting approached by B2B brands — and honestly, it's been interesting because I already use most of those products. I want to ask about enterprise — how are you thinking about bringing large B2B organizations onto TikTok as you continue to grow?

David Kaufman | TikTok 00:36:14 

What you just described is so interesting, because one of the patterns we've seen is creators talking about how to become a creator — content about which camera to use, what your gear setup looks like, how to handle sound and editing, and how to apply AI to the workflow. Editing takes three times as long as actually creating the content, so that's a real topic. And you get automatic authenticity because you're hearing it from someone who uses these tools daily.

On the enterprise side, this is one of the classic challenges of digital advertising. When you're doing B2B, you're trying to reach the budget owner, the decision maker who's going to select a vendor. That's been challenging on social media historically because people show up with their personal experience rather than their corporate identity. Those lines are blurring. LinkedIn, which was probably strictly business when it launched, increasingly has people celebrating birthdays or sharing personal milestones. But LinkedIn had the advantage of knowing everyone's title and career history — very useful for reaching someone whose title is VP of Marketing.

On our platform, what's interesting is that people sometimes have multiple accounts — a personal one and a corporate persona where they establish themselves as an expert in their field, talking about how to be a great product leader or a wonderful mentor.

The auto dealers I follow on our platform are a great example. When the showroom is empty, the sales staff makes TikTok videos — about the latest model, or humorous content about hanging out in the bed of a truck. That's how you reach a commercial audience based on their interests on the platform.

For truly enterprise decisions — like which CRM your company is going to use — the content itself is already on the platform organically: "How to select a CRM," "What criteria should I evaluate?" All of that is available, and you can advertise against those niches to reach people who are actively researching in short-form video.

It's really about being able to identify the niche of interest at the exact moment someone is looking for it. The other piece of this is the sales cycle. Buying a knife sharpener for my kitchen is an impulse buy. Selecting a CRM might be a five-year evaluation as you rotate off an old contract. So it's about finding the decision maker at the moment they are actively trying to get information that helps them make a good business decision. Enterprise use cases are opening up more now.

Carlos González de Villaumbrosia | Product School 00:40:23 

I guess that's also why CTAs have to be different — I can't imagine a "buy now" button for a CRM.

David Kaufman | TikTok 00:40:31 

A three-million-dollar click-to-purchase, right!

Carlos González de Villaumbrosia | Product School 00:40:34 

David, it's been awesome to have you on the podcast and geek out on the future of social commerce. I really enjoyed learning about all of these details — thank you for your time.

David Kaufman | TikTok 00:40:45 

I really appreciate you having me. This was a lot of fun.