Leading Marketplace & Monetization Insights With Adi Thacker | EM Group Chat #193

Welcome back to Everything Marketplaces, where we talk with founders and leaders from some of today's top marketplaces. So this is episode 193, which is a really great group chat I just had with Adi Thakkar, who's the VP of monetization at Poshmark. Prior to Poshmark, Adi was at TikTok, and before that led monetization at Facebook Marketplace. So he has some real incredible marketplace experience. So this is a great chat with Adi where he started off by sharing some of his key learnings from Facebook Marketplace and TikTok, what led him to joining Poshmark and leading monetization. We discussed some of the common monetization strategies for marketplaces. Adi shared some of the biggest mistakes that he sees founders make with monetization, a few trends with consumer marketplaces and AI. We also had a great group Q&A. I really enjoyed this conversation and I think you're going to find it a great watch to the end. So Adi, as I mentioned to you right before we started recording, you know, it's been a while since your past group chat and workshop that you led with us on monetization. So I'm excited to have you join us again here today. I want to start off by saying a huge thanks for taking the time to do so in advance. We have a lot that we're going to dive into with all of your incredible marketplace experience. But before we do, I thought it might be great if you can start off by briefly sharing your background for those here that have just joined us.

I'm Adi Thacker. I'm based in the SF Bay Area. Good morning. Good afternoon, folks. I'm not sure where you guys are. So I've been in marketplaces and ads for nearly a decade now. You know, I think this space chose me and then I kind of embraced it and I've sort of been in this space ever since. I really enjoy this space because of the complexity of matching. It's almost like you're building an economy of sorts. I was, my marketplace journey started with the startup in the education space that I started. That didn't go very well. Transitioned to Facebook, worked on ads, worked on marketplaces, transitioned to TikTok where I built a managed marketplace and then now at Poshmark for the last 3 years and some. as VP of Monetization and Partnerships. So that's my journey. I've seen a lot of dos and don'ts in the space. So happy to share my learnings from all these adventures.

That's a really incredible background again, and thanks for sharing with us on that. So a lot that we're going to get into here, but going back to Facebook Marketplace and TikTok, what were some of your key learnings from there that you've taken with you to Poshmark?

So for context, right? I mean, like, you know, Facebook, I mean, Facebook and TikTok are a billion, sort of billion DAU, MAU consumer companies, right? And Poshmark is sort of, I'd say, mid and small cap-ish and startups are startups, right? So like different orders of magnitude of just users and complexity and sophistication, not just of the product, but also of the talent and the tools that you have to build, ship, iterate, and manage. So I think, you know, when you go from a large company to a small company, there's a lot of lessons you can take in terms of Just talent, product rigor, craft, and just sort of basic principles around building good products and building good marketplaces. So I think the ones that I can rattle off from the top of my head, I've made some notes over here, is sort of product craft, right? I think of when you hire PMs, I've seen a lot of PMs in the non sort of FAANG companies tend to be sort of feature factory, sort of, you know, feature pushers. But The product craft at Meta especially was so sophisticated where the focus really was on the consumer or the user. And there was a very clear mapping of building features that drive metrics that solve problems for the users, right? So picking the right metrics that map to the mission of the product, whether it's buyer or seller or rider or driver, was super, super critical. And then it was very intentional to make sure that every feature that was shipped or launched Moved the metric, an input metric like Amazon calls it, that eventually moved the output metric to the marketplace, not just orders or GMV. So I think that rigor, I've not seen it, at least in my journey, I've not seen it pervasive. So I think that's number one. I think 2 is having a high bar for talent. I think, you know, being mid-cap or startups, you can't hire sort of the best of the breed, but I think that's something that you can probably, you can start with the B team and eventually advance to an A team as the product and the company grows. I think the one lesson that is marketplace specific that can be carried over, you know, irrespective of size is finding product market fit on both sides of the marketplace. And that's super critical, right? So I think that's something that Facebook obsessively sort of measured and managed and said, hey, are we Did we find PMF? Are we still in PMF? Have we fallen out of PMF, right? So I think PMF is a state, it's not terminal. So just sort of being able to understand that your product market fit, not only for the entire supply and demand side, but even the different layers and segments, a few different categories, right? It's super important. I think the one thing we did not worry about at big co that small co's have to really be concerned about is sort of the channel economics, right? I think, you know, at Facebook, I was oblivious to Like, you know, user acquisition economics and just sort of LTV to CAC because everyone is on Facebook every day. And, you know, when you have a billion people, especially on Marketplace daily, there's tons of room to, and traffic to experiment, not the luxury of a small cap company or a startup. So finding channel market fit and we really spent a lot of time on making sure our channel allocation or media mix allocation is optimal at Poshmark so that we have the right contribution margin and to drive the right EBITDA. I think, you know, as you're a startup or a small co, that becomes super critical too, right? What's the right channel to find my audience? You know, what's a sustainable payback period? And all those things become critical. So product market fit, channel market fit, right? I mean, Facebook, you just, they're only concerned with product market fit. And then finally, I think, you know, for folks that have marketplaces that are doing email and push, I think it's really easy to get caught in the loop of sending a lot of email to drive traffic or push to drive traffic, it's a slippery slope, man. That's what I can say, right? I think if you don't, you can only increase volume. Well, you can have reach, but you gotta be careful with frequency, right? And you wanna increase frequency only when you increase relevance, right? So just make sure you hold a high bar. I mean, you know, there are times when perhaps you need more traffic and you can get a bit greedy and do it. It's okay. But on a sustainable basis, you wanna make sure that, you know, you're principled, right? I think that's the other sort of overall lesson that we learned at Facebook is stay principled while building your products. I think they always said that, look, you know, first find product market fit and then find the spark and then forecast when you have the spark, right, to actually grow. And monetization and ads will always step 99, right? That, like, you know, the first 99 steps are for you guys to build liquidity, build liquidity, build traction, get the flywheel going. and then monetize, right? Because I think it's always a trade-off between long-term and short-term revenue, right? When you focus on liquidity, focus on sessions, focus on matching, focus on traffic, you're solving for long-term revenue. You start, you know, over-monetizing, increasing the take rate. Yes, you know, you may improve your revenue numbers in the short term, but you're paying the price in terms of liquidity down the road. So I think that's what I learned. I think Facebook was more sophisticated than TikTok. A lot of folks went from Meta to TikTok and Pinterest and and sort of read it and take those lessons with them into Snap as well. And then raise a product craft over there. And then folks will go to places like, you know, Poshmark, sort of mid, sort of small cap, mid cap, and then startups, right? So mostly the kind of waterfall of learnings you can take as you sort of go from company size in terms of craft and rigor and key principles from marketplaces and products.

Those are all really great learnings. It's going to be super helpful. Lots of topics too that we're always discussing here as early-stage founders. I'm glad you helped kind of emphasize the point of, you know, at the early stages focusing on getting to liquidity initially. So, you know, so transitioning over to, you know, Poshmark, you know, what led you to joining Poshmark? You know, what are some of the biggest challenges with it as a resale marketplace?

Yeah, so I think my goal was always to work to be an executive leader in a smaller company, and I think that is still my goal. So, you know, I sort of done the big code thing, and for me, You know, I felt like I'd reached, you know, I'm curious and, you know, I'm keen to learn and I felt like at 4 and a half years, I sort of maybe maxed out learnings over there. And I was looking at smaller companies and Poshmark, you know, was a very successful, that still is a, you know, fashion resale marketplace. It's the largest by a mile in North America. And, you know, they'd grown really well. They had a Good bunch of years during COVID they went IPO and then they were almost like over a decade since they were founded. And then the question becomes, look, you know, like, you know, we have a good, we have good liquidity in the marketplace right now. How do we build other monetization streams to diversify revenue? And I think, you know, there's a bunch of things you can do, right? You can introduce tools for sellers, tools for buyers. You know, you can use monetization responsibly. You can do subscription products. That's when they kind of brought me in. I think the company was ready to take on that next step and I was looking. So I think timing worked out and I've been there over 3 years now. So build out the ads business, we're increasing take rate. We are a material driver of revenue now. And yeah, so I think that was the genesis of That's great.

So, and I did want to kind of spend some time discussing specific monetization strategies. And I know you mentioned one to kind of start with as far as like take rates, you know, so when it comes to like earlier stage marketplaces, how do you advise founders to think about, you know, initially setting a take rate? And what do you think some of the key considerations are as well?

A take rate or commission is sort of the marketplace fee for matching buyers and sellers, right? And for processing payment and for managing fraud, right? So that is sort of your your fee from servicing the value that you provide. I think when you're typically starting a marketplace as a founder, it's not an advantage. So you're gonna look around to see like, you know, most marketplaces will have a take rate of a certain amount on the supply side for sellers. Some will have it on the buyer side. I think most commerce marketplaces have it on the supply side. There are a couple of exceptions like Vinted in Europe that has it on the buyer side. But I would say look around and, you know, see what the You know, see what your competitors are doing. I think the take rate magnitude has to be high enough to sustain your business, but it has to be low enough where there's an incentive for supply side to move over to your platform to keep most of their earnings from making a sale or selling a service, right? So I think that is, that's sort of the nuance right over there is there's no perfect number. I mean, lower is better always because then, you know, you just encourage more supply to latch onto your platform. And there's incentive to move over into multi-tenant, you know, if you have sellers on multiple platforms. I'd say, you know, it's just a function of your space, right? Let's take eBay for an example, right? eBay has got different take rates based on category because different product categories have different margins, right? I mean, in apparel recently in Europe, they actually dropped their commission to zero. That's what Vinted did. So Vinted in the UK and Europe, has zero on the seller side. They have a buyer side take rate. And then eBay followed suit in the UK and for apparel globally as well. So that's what they've done, right? So I think they've looked around, they've seen what is working or sort of how the trends are shifting in terms of consumer behavior on resale. And I'd say those are the key principles, right? I think like, you know, high enough to sustain your business, low enough to incentivize the listing behavior, make sure it is maybe, you know, consistent or so with the competition because you have consumer behavior at scale that's been attuned to certain models. So unless you have a novel insight and believe you can do the contrary, then I think, you know, that's a different approach. But, and actually Vinted did that, right? I think they came into Europe saying, look, you know, I mean, we're going to like go against the norm, we're going to shift the fee to the buyer side and it worked well.

That's going to be some great examples as well. So another thing that we're often discussing here is as far as like changing take rates or kind of marketplace fees, right? So do you have any tips maybe for founders to kind of think about this? And we can use an example maybe at like the earlier stages. So maybe they've kind of like used lower take rates to initially kind of like attract supply, but then thinking about like, you know, when they should potentially change it and how to do so.

Look, I mean, Etsy has done that, right? I mean, Etsy has one of the lowest take rates in resale. I think they're currently at around 6.5%, but they actually started off, I think around 4%, you know, when they IPO'd, right? So the take rate was very attractive, extremely low, They've since built other monetization streams to increase the overall take rate between the commission and other streams so that the blended take rate is pretty high. But being a new marketplace in that space, they were able to attract a lot of supply, you know, with that strategy. Over time, they've, I think they've changed their take rate 2 times since they started, right? So they went from 4, 5 something and the 6, 6.5. Each time they did that, there was uproar and pandemonium in the community. And I think the key really has to be is I think being at CU, look around and say, what's the best alternative platform for the sellers still here? And there are not very many, right? I mean, if you think about fashion resale, Poshmark is the biggest and a lot of sellers complain in the community, but eventually I think this is, if they're a professional seller or a small-medium business, this is their livelihood, right? So there's always going to be pandemonium and complaints and they'll be distraught with, you know, with the changes. But I think the lessons we learned is that you have to persist, right? I think what happened with Etsy's selling was, I mean, there was chaos, but eventually I think it did not affect the business too much. But I think what they saw is that the revenue growth exceeded their GMV growth because they were taking more from the marketplace. I've not looked at whether GMV growth has decelerated because they maybe increased the take rate. But it's hard to attribute that directly, the change directly. So they started low in, again, in the eBay apparel example, I think eBay is roughly around 13.5% or so as a take rate. I think they dropped that to zero. So that's the opposite example. And I think they did that as a reaction to competition. I think Mercari did something similar actually, and then they actually went back because it did not work for their economics and that was very painful. But yeah, I've seen Whatnot do short-term holdouts as promotions where they actually say, hey, like, you know, we're, you know, we're not gonna charge a seller fee for, you know, for a few weeks. So I think that's the other level, right? So I think in general, I think you wanna start low to encourage liquidity. I think you wanna build supply typically in most marketplaces and then you attract demand and then that's what What kind of kicks off the floodgates?

Definitely. Those are lots of great examples. Also too, you mentioned the Whatnot example, you know, previously on your workshop as far as like suggesting for founders to make it time-bound as well.

It's also a great test, right? I mean, you know, to see if dropping your marketplace increases supply liquidity, and you could actually test that, right? You could drop it, you know, by a few points and then drop it to zero and see the elasticity of that in terms of supply and see what that means for attracting demand. as well.

No, definitely. And then you also mentioned, you know, previously as far as like, you know, value-added services. So sometimes, you know, marketplaces, it could be, you know, somewhat table stakes to, you know, add these value-added services. So going beyond just simply kind of, you know, connecting buyers and sellers or matching. So, you know, when it comes to like monetization strategies in this, you know, how do you, I guess, like advise founders to think about that and how to account for that in a monetization strategy?

Yeah, there's kind of typical streams, right? So beyond your take rate, which should be your primary stream, there's advertising and there's 2 types out there. One is sort of a marketing tool where sellers can promote or boost their listings or products. I think that's most common if you look at eBay, Etsy, Poshmark, DoorDash, they all have this, you know, they call it promoted or boosted. So they can, and it's a seller agency tool, right? I think sellers that are willing to pay more to sell more or sell faster, increase sell-through, will pay the extra premium, which adds to their overall take rate. But I think that's a pretty typical tool. If you look at I saw the Q1 numbers a couple of days ago and eBay adds around 2% to their take rate through ads, right? Around 14 to 15% of the revenue is advertising, which is very material. I think Etsy's even higher because their base take rate or commission is so low. Poshmark, we've been on this journey for 3 years now and it's working out pretty well. So I think that's one core stream, right? The other streams that we've seen are subscriptions, right? If you look at like, you know, Uber One and DashPass and those typically are on the demand side of the marketplace where you see like just like having advertising automotive listings for the power sellers, you also have tools for the power shoppers or buyers or riders, right? They actually want a premium car or they want free shipping, for example. But that's also another revenue stream that is pretty common. And then eBay has got something called subscription for shops. So, you know, some of their bigger businesses want insights, they tooling, they want lower listing fees. So I think that's the other stream. So there's a diversity of things. I think, you know, I've looked at this in the past and, you know, companies like SitterCity, for example, you know, that match caregivers with care seekers have a background check service and other stuff like that, right? So these are the peripheral things, but I think the key over here is sort of timing and making sure that these secondary streams are accretive to Take rate and they're equated to GMV, right? It's very easy to go far down the path of introducing too much advertising too soon because the nature of advertising, even if it's done very responsibly, is not all sellers are gonna advertise. So, and when you show these ads in any experience, it could be search or recommendations, you're pulling from a smaller pool of items to represent. Right? And it's taking up real estate, right? So the relevance is always going to be lower, right? But I think, but you're getting some relevance. I mean, the cost of lower relevance is immediate revenue. It's a click or an impression or what have you, right? So I think it's important to be cautious to introduce especially advertising, especially endemic advertising, which is like, you know, promoted listings at the right stage, right? I was looking at some of my older slides and eBay introduced ads around 20 years after they launched, right? Significantly. Etsy was also, I mean, Poshmark has been over a decade and then you can start using ads. There are exceptions like I think I was looking at Instacart. Instacart launched ads before IPO, right? Because that's a 3-sided marketplace, groceries, very low margin, ads are high margins. Ads are a very inherent part of So yeah, so I'd say, I think, you know, I think the cardinal rule here is focus on liquidity and grow our marketplace as much as you can. I think GMV is sort of the North Star. I think there are sort of many levers downstream and as to increase take rate both via commission and through ancillary streams of revenue. But Marketplaces that get to that point, a secondary stream of revenue is usually a privilege of the successful. It's not something you want to consider early in the journey.

Yeah, that's such a great point. So I know I mentioned to you right before you started recording, you know, when it comes to ads, like a lot of times it's, you know, typically kind of like too early when founders are thinking about that. And, you know, sometimes kind of degrades the marketplace experience and is at the cost of the marketplace.

Yep. Yeah. I think the other example is, I mean, you know, you could argue that, well, you know, if I'm a classifieds marketplace, you know, that's my only source of revenue. At Facebook, again, Facebook has the privilege of being sizable and being funded by, you know, by Meta. But Facebook Marketplace, we were very responsible with ads. We only introduced ads, you know, once the marketplace had product market fit on both sides of the market, right? So in general, you don't want to introduce any elements in the shopping or selling experience that may impede your path to product market fit and get the flywheel going to scale.

Yeah, that's going to be really helpful. And I guess like if you had to share like a few kind of tips for founders on, you know, like I would say like, you know, when and then how to, you know, roll out like ads, you know, what would a few of those tips kind of be?

There's 2 types of ads, right? I think the first type is just sort of endemic, which is like, you know, you're boosting listings. So it's organic, it feels organic where, you know, in commerce you click on an ad and you go and buy something within the marketplace, right? And then there's non-intrinsic ads, which is inserting third-party advertising into the consumer experience where a click takes the user offsite, right? I think the latter is more dangerous than the former. You want to be careful with both. I think the cost of introducing third-party ads is low because you just plug in an ad network. The cost of introducing first-party ads is higher because you need some sophistication in-house to Well, not only build it, but also do it responsibly. I'd say push it out as much as you can is my advice over here, right? Because as I mentioned before, I think this ancillary revenue stream is something you want to consider when you're at scale, when you have a high-growth marketplace and you want to boost the take rate. I mean, I always call them take rate boosters and not GMV boosters, right? I think at this stage you want to be focused on liquidity and GMV. me as a founder.

Those are great tips. So I want to spend some time discussing AI. I know we could get kind of bogged down into monetization. It's a really exciting time in the world of marketplaces right now with AI. So how are you seeing AI impact on marketplaces and how do you think this will kind of, I would say, like change the future marketplaces kind of moving forward?

The main changes that we are seeing at our end are on the listing journey. I think listing can be a bit tedious now, especially if it's not coming through a feed. I mean, it's a manual consumer-based listing. So there's a lot of tooling to list better, more efficiently where you can just take a picture of the product and, you know, the different attributes auto-populate. I mean, it's not perfect, but it definitely sort of brings down, it reduces the tension of creating a listing. And, you know, the more listings you have, the more accurate they are, the better sort of better the description is, it's better for your search and matching engines to actually pick up attributes to show them to the right shopper, the right user at the right time. So that's one area. I think the other part that I'm seeing a lot in the marketplaces is a lot of PDPs or listing detail pages now have review summaries, right? So that's pretty common, right? So it's a very easy and very simple use of AI that they're sort of reading through every review. You get a So I think Amazon has it and a few other marketplaces have it. I think the other part is sort of this using transformer models in matching is key now. So a lot of ML-based matching has just been based on ML, right? I think like there's recommendation, there's search, but now with new transformer models, it actually introduces new dimensions and attributes to create more liquidity in matching demand and supply. So I think these, these are the main things that we're seeing. I think we're also thinking on the fashion side, on the shopper side, we're seeing the use of AI as a shopping and sort of a styling assistant in the apparel space. So, you know, say like, you know, if you say you like certain things, you could, you know, the AI shopping agent is gonna recommend different styles that they think might sort of match your taste. So that's what we're seeing mostly. So mostly I think what we've done at Poshmark is we have a smart AI listing tool that has worked out pretty well. It definitely eases, sort of reduces the friction of listing. It's more accurate, it's speedier. And then on the ad side, we are introducing customer-based models to predict clicks and conversions.

And another topic that I wanted to chat about kind of briefly right before we get to the group Q&A is, you know, some trends that you're seeing as far as like with the social commerce and consumer marketplaces. So I guess like, are there any, you know, specific trends that you're seeing right now?

Yeah, I think, you know, we've been paying close attention to just styling and sort of agents, right? I think that's one space. So we're actually watching a couple of AI-based startups that help with styling. So I think, you know, as the modality of interacting with apps changes from being very search-driven or to kind of being interactive and sort of chat-driven. You know, it's very likely the hypothesis in this case that I think that's gonna become maybe the predominant or significant share of behavior as we interact with apps in general. So, I mean, I spend as much time now on ChatGPT as I do on search, right? So, and if you carry that into different sort of marketplaces, experience as a buyer or seller, your modality of selling or buying something will be very different than it is today. You're not gonna be searching or browsing. I think you may be a bit more direct. So I think that's something that we are watching. There's a bunch of startups in the fashion space that are doing this, you know, both on, you know, both in terms of selling easier and even in terms of being shopping assistant. So I think that's a trend we're seeing. What else? I think from a macro perspective, I think the tariffs are definitely helping resale in the fashion space. So we don't know how persistent that advantage is gonna be. But I'd say those are the 2 main trends. I think on, again, on a macro basis, I think, you know, it's not been the easiest time for commerce marketplaces. There's very few that, you know, the other big ones, Walmart, Amazon, have been doing really well. If you look at eBay and Etsy, you know, they're not doing as well. I mean, relatively. So I think overall we're sort of watching consumer spend on commerce and retail. So I'd say, yeah, so AI and sort of the macroeconomy have been key areas. I think broadly, I think the other thing that's really interesting is I think since IDFA has happened, right? I think it's changed the user acquisition economics significantly for DTC companies and performance marketing teams. And that's a very dynamic space, right? I think, you know, Apple and the different, I think the French regulator now actually has said, hey, like, you know, maybe, you know, you should pass an IDFA to offer better targeting. So I think how that space evolves will change the acquisition economics and even the channel market fit that you talked about earlier for many startups. It might make the acquisition economics more We'll have to watch that space.

Yeah, definitely an interesting time right now. And you bring a unique perspective as well. So it's going to be great. We'll probably get to a few of the questions during the group Q&A that are related to it. So, you know, right before we do though, because I definitely want to make sure we can save some time for founders to come on here, do you have a few tips maybe for early-stage marketplace founders that you could share with us?

I just say like, you know, focus on product-market fit. I mean, liquidity, liquidity, liquidity, right? I think that's the key. I mean, look, I mean, there's 3 ways to increase orders, right? I think you increase demand, you increase supply, and and you improve matching. I think, I mean, it's just like physics, right? I mean, those 3 things do not change, right? And I think depending on where you are constrained, you're going to have different sensitivity of these different levers to improve or increase GMV or orders, right? If you are demand constrained, you bump up demand, you're going to see an uplift. If you're supply constrained, same thing's going to happen, right? And matching is always a constant lever, right? You can always optimize algorithms like constantly to get those, hit those signals and continually up. And I think that's the other learning from Meta as well. I think at some point you become demand constrained either because of the acquisition economics or you become supply constrained because you have all the key sellers on the marketplace, right? But the gift that keeps on giving is optimizing and matching the algorithm. There's a lot of signals you can pick up, you know, from about shoppers and about buyers on the demand side. And you can continually improve the way you show the right item to the right user in the right moment. So that they can click and convert and buy something or book something, right? So I think that's something that we learned, that even though the impact of this endeavor is small, but it's almost like you can hit these singles quarter over quarter and the impact of it compounds over time.

It's going to be helpful tips.

So cool.

So we're going to get to the group Q&A. Hey, Molly, do you want to come on?

I'm Molly Morse. I'm founder of Recast, which is a marketplace for in-person kids' classes and camps. I really am curious about matching and advertising. So how you relax, if you relax some of the matching, sort of the matching algorithm to surface paid ads. I've seen places that relax that too much, and you see an ad that's completely unrelated to your search. It's a very aggravating consumer experience. But then I imagine if those restrictions are too tight, you can't make enough revenue on those ads. So just curious if you have found the sweet spot, what that looks like.

So good question. And there is no sweet spot. I think the question becomes sort of the give and the get, right? So I think there's actually a way to measure this and arrive at the sweet spot that's palatable to you. And I think we usually do it, we create a holdout. It's actually called a no ads holdout. So maybe like 10, 20% of Or if you're a consumer, you don't want to show them any ads. And then what you'll see over time is that you'll see the difference in GMV between the segments that see ads and don't see ads, right? So, and then you'll see the net revenue as well. So eventually the segment that sees ads, you want to have incremental revenue. So I think your goal with ads is incremental revenue. So if the segment that sees ads, the ads are so disruptive that the revenue gained offsets the revenue you get from your take rate or your primary stream of monetization, then it's a bad idea. So then maybe you wanna curtail ads a bit more until you get to a point where advertising is relevant enough for it to be incremental to your overall business. So there's no exact point. I think the exact point is something that you arrive at and it's very unique to a marketplace based on the moment you are in terms of liquidity and in terms of, you know, buyers and sellers or I guess parents and kids in your case, right? So I think, you know, understanding the framework to arrive at that point is more important than knowing what the point is. And I think that is sort of the framework, right? I mean, it's gonna summarize partition your audience, Measure diligently, see if ads are incremental. Are they incremental at all, incremental enough? And then you make the call to relax constraints or not.

That's a great question. That's going to be super helpful. Hey, Adi, I actually had a related question because it's come up a few times in the community, and it's, you know, for marketplaces that are a little bit earlier stages, so they might not, you know, have experimentation, you know, how do you suggest maybe testing when it comes to ads?

You know, this is a problem, right? I think, you know, everyone talks about moving fast, breaking stuff, A/B testing, speed of testing, but it's a privilege of the rich, right? I think you have to be a certain scale to actually do it, right? And, you know, the speed of testing at Poshmark is significantly lower than the speed of testing at Facebook for a reason, right? As, you know, because your audience is limited. So I think there's a couple of ways to do this, right? I think you don't need to have a perfect statistical setup at this stage. You can't even do it, right? And if you do it, it'll take you months, maybe a year to run that, right? So I think it's okay to do a before and after and On-off test, like go cold. Like, you know, it won't be perfect, but I think because your scale is so small and the treatment of the insertion of the treatment is so different from the baseline, you'll see a large enough effect where before and after, and we don't before and after institutional testing because of the effects of seasonality, right? But seasonality effects are small, right? So I think at your stage, you've been doing, hey, I'm going to run ads, I'm going to run this new feature for a week, I'm going to take it away for a week. If the change is large enough, you will see a difference right away, and that should give you confidence to be able to do it. That's one. The other one is sort of more principle framework, kind of leap of faith. Well, that company in our space has it, so we should do it because they're big enough and they've done a bunch of testing at their end and they're more advanced, so we should do it, right? So I think if you can triangulate your way to to being experiment-oriented without being really statistically rigorous, given the scale.

It's going to be really helpful. So yeah, you also mentioned that seasonality, right? So I feel like there's so many different kind of variables that come into play.

Yeah, yeah.

But I think, you know, when, you know, for example, if you insert ads and you insert a lot of ads, right, the change is going to be significant enough where any seasonality change, whether it's weekly, is going to be sort of minuscule compared to the impact you're going to see. But it would be an introduction, yeah, ads or some major feature, right?

Awesome. Hey, Connor, do you want to come on? I know you sent me a question in advance.

Heidi, thanks for jumping on. My name's Connor Leydon. I'm the founder of Trust App. We do end-to-end transactions for marketplaces, so we run kind of payment, shipping, dispute management for all sorts of different types of marketplaces. And I had a question kind of going back to your time at Facebook and Just curious, you know, why you think Facebook never pursued kind of a strategy to go transactional or to, you know, to let their users transact with each other. Like, yeah, that's kind of the main question. Facebook Marketplace is majority classifieds, right? It's local, it's, you know, and they do have shipped products, right? I mean, if you go through their listings, they do have items that, but that's a minority of of listings. They did, well, they had Facebook Shops and Instagram Shops for a while. So I think they did have an intentional strategy, which TikTok has one right now for TikTok Shops, right? So they tried it. I was there when they tried it. I think I left and I think, I don't think they were a success because if you look at Facebook Marketplace, right? Majority of the commerce that happens is discovery driven. It's not intentful search driven. So it's people that are coming in, they're just like looking around, hey, this is a good deal, I'll go buy it right now. It's the same behavior on Facebook and Instagram. So when you create a commerce marketplace over there, you know, and you bring in all these brands, you know, you get a lot of discovery but not enough conversion. So I think what I, my observation when I was there is when they tried doing this, they brought in a bunch of brands across different categories, fashion, et cetera, et cetera, just the conversion yield was, it was just, you know, for that scale was just not optimal. So which is why I think they, that I don't think they found product market fit with any category that they had tested from a commerce perspective. And yeah, and I think just knowing Facebook, because if they don't find product market fit with something, they basically choose to terminate. initiatives pretty quickly. So I think, yeah, I think because the audience, the intent of the audience when they open a Facebook app or an Instagram app is not commerce. They want to be inspired, they want to be entertained, and that's it. TikTok Shop's a bit of a different story. I think TikTok audience is younger, Gen Z, and they tend to start a lot of searches on TikTok. So they're coming in with intent and then they also have a creator-based approach in terms of shopping and their shops. So I think it's a different audience with a different intent. And I think it seems to work better with TikTok and they have the scale to attract the right kind of creators and brands to create shops.

That's a great question. It's come up a few times in the community now, kind of discussing. So that'll be helpful. Hey, I actually had a question from a founder that sent me in advance and it was around reduced fees for power sellers. And is that something that you've seen work well as far as reducing fees to attract power sellers. And then somewhat related to this was as far as when power sellers might bring demand and as far as reducing fees for that. So do you have any maybe tips or I would say advice on that?

Yeah, I've seen that happen where larger sellers get lower fees because they are pursuing operating margin is sort of their optimization, right? So they're very margin sensitive. Yes. So I think that is a tactic, especially if it is a large seller that has a key bunch of listings or supply that is strategic to the marketplace. Or I think, as you mentioned, I think, you know, if in the case of live streaming, like, you know, if you have a key influencer that brings in an audience, which is great, right? Because then you're saying, well, you know, you're going to save me user acquisition costs because you're going to bring me all these shoppers or users on the demand side of the platform. So I'm actually okay, like, giving you a lower sort of taking a lower take rate from you. You know, to, in this experience. So, so the answer is yes. I can't put a finger on when and where and who has done this, but I think it just makes sense, right? I think the challenge over here is I think if you start doing it across the tiers, then it becomes a bit sensitive in how you communicate this to the community at large of sellers. So I think if you have key strategic deals with large influencers, then it probably makes sense. And nothing that's been done before.

Definitely not. I'm sure too, if you have a, you know, you have those on like a per-seller kind of basis, then it becomes hard to manage those as well. Cool. Hey, Maxine, do you want to come on?

Thank you for such nuanced, deep answers that make my brain like consistently go deeper. I'm Maxine. I'm the founder of a company called Pick My Brain, where we sell marketplaces to communities that want to facilitate more conversation and connection. But my question is related to percent of your decisions that are based on data versus intuition at the different companies you've worked on? Yeah, when do you jump outside of the models? When do you stay within the models? How much range do you get depending on the size of the marketplace? Comment on that.

When you're starting a company, it's based on intuition, it's based on a whim, right? I think, you know, most product decisions are just based on instinct, they're based on intuition, you know, I've seen this work well, This is a consumer pattern. You know, I saw this work in that industry. I think that can work well over here. So I think, you know, as you go from having no data to less data to having, well, having no traffic to some traffic to a lot of traffic, you can gradually shift your sort of the construct of what informs your decision from intuition to data. It can never be data-only experiments. I think you have to do even some, you know, your instinct, the qualitative research, talk to the users, triangulate patterns from other marketplaces or the consumer apps, and then arrive at the decision. But I think you're able to create, you're able to build conviction the bigger you are, right? Because you have multiple sort of sources of information. But even when you start off, I think, you know, you can use your instinct, you can talk to shoppers and talk to users, and sort of at least build a fair degree of conviction. I think, you know, when you're building a product, the key always is I'm gonna invest all this time, all these resources to build something. How do I have confidence that I'm gonna have ROI on this feature or the service that I'm building? And I think the more conviction you can build, the better off you'll be, obviously. I mean, you know, I think it's sort of almost like risk minimization in terms of getting to the outcome that you want, right? So I think that's sort of how I think about it. You know, When we built a couple of zero-to-one experiences, I think, you know, we said, look, you know, here are the key problems we're trying to solve for. We know these are problems because they're true. We've talked to these users. They currently solve the problems using methods A, B, and C. They'd be better off with our solution because the problem is a fact. The solution is creative, right? So you wanna make sure that the solution that you arrive at is solves the problem better than the current alternatives, right? And that's why you need to keep iterating and moving fast. And yes, you can A/B test, but you can quickly ship the solution to solve the problem better.

Well, cool. So we're almost out of time here, but, you know, I appreciate taking the time to join us for the, you know, the follow-up group chat. This is fun to kind of dive into things again and, you know, hear some kind of updates and also to the latest with AI here. So I know it's an exciting time and also appreciate, you know, taking the time to answer some of the questions that we all had for you. So Actually, I had one last question before we wrap things up, and I know I kind of asked it before, so I'll kind of like slightly reframe it here, which is, you know, if you could go back to kind of, I would say, like that kind of pre-AI, or sorry, right before that, the kind of AI era for our marketplaces, you know, what would you tell yourself about them?

I mean, it sounds like a cliché, but to me, like, you know, like, you know, massive change is coming. I mean, if you just, I mean, ask yourselves, right? I mean, the way you consume information has been transformed. I mean, You know, I plan my trips and seek advice and, you know, learn stuff primarily now on AI. So I think what's interesting is, I think what's changed is not just the chat agent experience, but I think we're starting to trust the responses a lot more and have that guide our decisions in life, right? To me, that is sort of, I think it's not just a modality of interaction. I think it's the impact of that interaction that I think is significant that we don't talk about enough. You know, you can search for stuff and you get a lot of options, but then you do a lot of digging. But I feel like the depth of trust and the impact of that research or the information is not as profound as it is when you're using AI. So I think to me, that's the key transformation. So the impact of the experience is far more severe, positive or negative, hopefully mostly positive in how we make life choices, right? I mean, you know, you're buying a car, you're shopping, going on a trip, seeking advice on some issues, personal Or otherwise you have an agent right there, like $20 a month, and it does a pretty darn good job.

Well, that's going to be great parting thoughts for us all here. And once again, really appreciate you taking the time to join us here today for the awesome follow-up group chat. And I'll also be sure to include links to the past workshop resources and also your socials as well so everyone can follow along. So thanks again for joining us for the awesome group chat, Adi, and thanks everyone for the questions.