How to build great products—lessons from Amazon, Facebook, Twitter & Deel. | Aaron Goldsmid, Head of Product at Deel

Aaron was Director of Product at Amazon, VP Product at Twilio, PM at Facebook and Twitter. Now he’s head of product at Deel where he reports directly to the CEO. Deel was founded in 2019— now, just 5 years later, it’s worth $12B and raised over $650M. 

We go deep on what it takes to build world-class products, how early-stage founders can balance customer feedback, vision and data, and how the best product leaders are leveraging AI today.

Keywords

productivity, AI, product development, user experience, product management, customer feedback, product market fit, technology, innovation, startups

Why you should listen

• Why you’re better off trying to enhance than replace with AI— at least today.
• Why you need to deeply understand your users to build products they truly love. 
• How to avoid the customer feedback trap and only build features that move the needle.
• How and when to hire your first product manager. 
• What makes a product truly great 

Timestamps:

(00:00:00) Intro
(00:10:25) AI Solves the Informational Retrieval Problem
(00:14:54) AI Agents Aren't Real Yet
(00:18:12) Signals that Could Lead Towards to PMF
(00:23:04) Finding the Customers Need
(00:25:17) Hiring Product Managers
(00:28:00) Difference Between an excellent Product vs an Okay One
(00:34:17) Prioritizing Bug Fixing vs Feature Requests from Existing Users

Chapters

• 10:25 AI Solves the Informational Retrieval Problem
• 14:54 Why AI Agents Aren't Real Yet
• 18:12 Signals that Could Lead Towards to PMF
• 23:04 Finding the Customer's Need
• 25:17 Hiring Product Managers
• 28:00 Difference Between an excellent Product vs an Okay One
• 34:17 Prioritizing Bug Fixing vs Feature Requests from Existing Users

Transcript

Aaron Goldsmid (00:00)

Think about how much of our day is just detritus, right? Like not high leverage work. is making it so more and more of a human's day is high leverage work. And so I think the companies and the products that are doing this well are recognizing who it is they're working for and either A, providing them more leverage or B, making it so it's easier to interact with computing. 

Pablo Srugo (0:25) 

People think of the user as “you have to build for a five-year-old” like they're stupid or whatever. And I think the right way to think about it is : they just don't care nearly as much as you do. You know what I mean? Like they're going through flow and they click on another app and something else comes and it's distracting or maybe just more important to them. And your app is done. It's dead. Like it's the back burner. That’s the 22 % difference, you want someone to get something done, like make it easy and keep them there and keep them focused. Cause there's a million other things that can be done.

Aaron Goldsmid (0:52)

For all those founders deals five, six years old, there is a path to being an enormous tens of billions of dollars of valuation in a real short period of time. i think as we go forward, there's going to be a really interesting lesson to learn from how we approach building products.

Pablo Srugo (1:06)

 Welcome to the product market fit show brought to you by Mistral, a seed stage firm based in Canada. I'm Pablo. I'm a founder turned VC. My goal is to help early stage founders like you find product market fit.

Aaron, welcome to the show. 

Aaron Goldsmid (1:24) 

Thanks for having me. 

Pablo Srugo (1:25)

Dude, so I mean, this is a bit of a different episode than we normally do, but I think we'll get pretty tactical on the product side. You've been- name the big company you've been there, right? Like Amazon, Facebook, VP product at Twilio. Now you're head of product at Deel. Maybe if you can take a few minutes just to kind of go through your background. just so people understand kind of like the context and then we'll dive in on a bunch of product things.

Aaron Goldsmid (1:51)

 I would love to. Look, I've been privileged to work at some of the world's most renowned tech companies. Like you said, Amazon, Microsoft, Twitter, Twilio, and in roles that were sort of prime and center. So I looked after the growth team at Twitter or a Facebook messenger at Facebook or messaging at Twilio, which was their core business. And while the products and like FinTech and while the products were diverse, the real common thread on these roles and its name is the same with Deel is just helping the high growth tech companies make their products as intuitive and human-like as possible and growing and scaling and developing great product organizations. know, at Deel , which is for those of you not familiar, it's an all-in-one payroll on HR platform, but it's really for global teams. So, I mean, yes, it could be, you could be one market, but if you're a truly global team, And, you know, look after the flat-on-share product. but at the moment I'm spending a lot of my time working on AI and also Fintech, and sort of a good deal of time in the intersection of the two.

Pablo Srugo (2:56)

 Perfect. Well, let's talk. I need to talk to you about like, because you've seen product from so many different angles every what? 15, 20 years now with maybe shift platforms, right? The last one was like mobile. And for the first time recently, we kind of saw what the next wave would look like. You know, we've been talking about VR, you know, MR or whatever for so long, but that Orion released to me as a non-product expert at all, who, you know, in 08 was like in high school or whatever. So I just wasn't thinking about these things. That seems to me like the first thing I've seen and certainly like hearing people who've actually tried it that can replace the smartphone or like be the next like true computing paradigm and whatever that means from a product perspective in terms of apps and whatever. Like, I'm just curious about your impression on that. And if you, and if you agree,

Like does that take over the smartphone? Like where are you at? What do you think?

Aaron Goldsmid (3:46)

 So I think if you look over sort of the longitude of media, right? Like I'm gonna get sort of nerdy and Marshall McLooey. 

Pablo Srugo (3:53) 

Please man, please. Yeah, yeah.

Aaron Goldsmid (3:54)

The way that humans want to interact remains the same. There's no distinction. There's hot and cold medium. The thing was that like what the mainframe was to the PC meant I could do computing and I could access and do work outside of an office and inside my home, right? then to the laptop, I could do it in a coffee shop. And mobile now was, had sort of the entirety of the world's history anywhere I am. But those were still relatively clunky affordances. Like you had to train to learn how to type, you had to train to learn how to query, you had to train to learn how to sort of do these. And they weren't very intuitive. Now, you know, I think with the smartphone, we saw a really big step forward of that. But what we're seeing with sort of the current generation, both on the AI and some of the new visual affordances. They are becoming more and more intuitive. So they're mashing people where they are versus people have to train up to them. Now, when that happens, one, you see a sort of a shift, but where you see a sort of limiting factor is, you need the killer app, right? You need the like, this is the way that not just this is my life getting a little bit better, but wow, I am doing something in a way that I did not think I could do before. For VR, I think gaming was the best use case. When we look at generative AI, it's, you know, being able to just interact in a very conversational way and get incredible results. 

Pablo Srugo (5:17)

Well, what's your thinking on - the two approaches I see now, you look at what Meta is doing and how Zuckerberg approached it. And then what Apple's doing. I'm curious on, on your take from like a product strategy specifically, because what I see is to your point of like use cases. Zuck seems to have gotten it right in the sense that he's got like he started with the VR, which was gaming and like that's been great with the Quest, right? And then the Ray-Bans were just, you can take pictures. And then AI really helped him out, but it was very specific, like, you can take pictures and videos like with a click of a button, very specific use case. And I know people love it for that. The Apple Vision Pro felt like a frigging awesome product that you don't really know what to use it for.

Aaron Goldsmid (6:00)

I think when you think about building great products, you have to think about what the DNA of a company is. So like for us, it's about speed and global expansion and like being really sort of going deep into the core infrastructure. At the end of the day, Facebook is about the social interaction, right? So everything is in service of that social interaction, right? And so the Ray-Bans or even the Spectacles, which was the Snap version that predated that, was about how do you create the content in a way that doesn't take you out of the moment? Why is that great? It creates more content, right? Now, I think what you're seeing now is how I consume social content in a way that doesn't take me out of the moment, right? Like how do I not have a screen in front of my face, but I'm still interacting. If you look at the origin, like what's the DNA of Apple, it is about the knowledge worker. It is about the knowledge of the creative version of that, about the media creation and consumption. That's kind of been where they have innovated at every step of the way. and they are a platform. Apple's not seen for their apps, Apple's seen for building the platform. And I think what you saw kind of fell short on that Apple Vision Pro is that they didn't have a killer app, right? Like I was at Microsoft when we launched the first Xbox. If we didn't have Halo, it wouldn't be something we talk about anymore, right? And so when you see this as like, think maybe that they, and I'm not saying they're not gonna get there, but when… The first phone came out we didn't know- people were billing it because it was really, really cool. Now there's a very, very mature ecosystem of developers for apps and a really mature way is sort of like a business model behind it. And it's really hard to say, I'm going to build a killer app for a $3,500 device that's going to be owned by basically just me and my friend. And I do think that to be perfectly frank, like what we've seen in terms of what Med has been doing is still the early versions of it. We haven't lived with it.

And so it's a little bit unfair,you know, one's in a perfect environment and the other one now, now is the, mean, let's be clear, the demos for the Apple vision pro were truly amazing.

Pablo Srugo (8:07)

That's true. But even if you talk about the products that are out, at least they have clarity, right? Like if you think about the RayBans, like they are a product that's live. It's definitely not Orion, but it's like, Hey, you know what to do with it. and then Quest on the other side, you know what to do with it. So they have that least use case clarity, which I find is maybe missing on the other side.

Aaron Goldsmid (8:25)

I think that's correct. But also if you think about, if you rewind to Apple in that moment, what is the obvious killer apps for the phone? No one knew. and mind you, they had the war chest to be able to insert the time, they look at time rise and be able to do this, which for a zero to one founder, this isn't something that you can live with. They'd be like, someone is gonna come up with a killer use case for this and we just need to make it an open platform and the right thing will emerge. versus with a Quest or something like that, it's like, this is a gaming device. You can do other things with it, but it is a gaming device. And so the pantheon of an app developer's imagination is more constrained. But it's also, “I know what to do with this. I know how to make games. I make games. This is a thing.”  Plus the price point was a pretty significant delta. So I mean, I still think that having used both, I think the experience of the Apple Vision Pro, that like the fundamental ability to interact and how we interact with computing is truly, truly impressive. But you know, like the games on the Quest are pretty fun.

Pablo Srugo (9:32)

 And then I guess maybe just shifting gears, right? Like I think that's, that's a top of mind stuff that, to me, it's interesting, maybe from a more VC side, like where are things going? But you know, at the end of the day, as a founder, you either have a product right now, or you're thinking of building a product right now. You kind of got to build, like you want something that's going to be relevant in the future, but it's also, it's gotta work today. Like it's a weird time because I think when - let's talk about the last one, right? When mobile came out, like the time to build a winning mobile app was that like, wait to 2012 period, right? You look at Venmo, Snapchat, Square, what at Uber, like they're all like in that window. What are you seeing today in terms of AI? Like what are some of the, there's like right ways and like wrong ways, I think to build this stuff into your product, to leverage it. Like what are some of the best ways you've seen this implemented or like even just maybe, or on the other side, like mistakes you're seeing people make when it comes to building with AI?

Aaron Goldsmid (10:25)

 Sure thing. I think, well, I could also say what we're doing at Deel. I think is a good representation, partially because I have a big hand in it, about how to do it right. Like we're using it everywhere possible, internally, externally. But the way that I think about it is twofold. One, is, it doesn't replace a human. It gives human a lot of leverage. Or as I joke, it's like giving everyone a graduate research assistant, which means they do a lot of legwork. not always right, but they do a lot of leg work. But what it does is it gives me language or gives me leverage. And then the other is a new, it's a new user experience. It's a new way to access and access computing, right? Like the little glowing, the phone in my pocket gave me infinite stuff. I now no longer have to think about, how do I structure this query? Or, where do I go to do this? I just ask the question in the way that I normally would. And so you put those two things together. How would you- When we talk about product market fit, it's usually universal. It's like unrequited human needs. could be, you know, a thousand years old, a connection or efficiency or knowledge, the speed of information or money, all of those things. Those have not changed over hundreds, if not thousands of years. The tools had made it all easier and faster. And, and it has become the barrier to entry has gone precipitously down over time. Right? And think as we're seeing AI, like I can just ask it. So for example, for us, you want to go hire somebody in a foreign market, like, what are the parental leave policies in the Netherlands? Or like, how do you do time off reporting in Argentina? It becomes, as a human, you're just like, nah, I'm out. I can't think about it. Speaking of YC founder who had a friend in France, they're like, nope, too hard, not going to deal with it. But being able to access that, it's an information retrieval problem. One, we built a ton of, I actually, I think one of the ways I would think about is when search engines first came out, there was a burgeoning market of SEO, writing content for a search end. I think there is a similar moment right now, which is writing content for an LOF. I think there's a big problem on the fundamental training of them, but also in the sort of more the application of them. Like how do I get this content to be retrieved and give the right answer to the human? In summary, we're talking about a new information, like an affordance, like a UI. Like it's a new way to interact with computing that is way, way, way more natural, right? In which I'm not having to constantly think about how is the computer going to think about this? How is the search engine query going to think about this? Or where would I put the button if I was this app developer and that's why I will go look for it. Instead, I'm just going to say, I'm having a dinner party. Here's the things this person doesn't need this, make me a menu. That is just the underlying human needs. So we're getting closer and closer to that sort of human nature. On the graduate research side, which is that other part, which is like, how does it give me leverage in time? Like we're doing this across the board. So for example, there are, you know, 120 countries in which we're operating in. Every one of those has probably four or five government agencies that have some impact over work, right? It's the tax agency or, or maybe it's labor or finance keeping up to date on that could be an enormous human problem. Instead, we are automating that process and being able to sort of say like, look, it's not making the final decision saying like, look, there's something new here. Here's what it is. Here's the raw. Here's the summary. Then our humans, like the experts, are able to sort of go through and do the analysis and sort of create the right content and make the right decisions. But that one human is the force multiplier where I would need to have 10 people. I can now have one or more importantly, that one person can do so much more. and they are so much more efficient. And really what we're doing is we're thinking about how much of our day is just detritus, right? Like not high leverage work. is making it so more and more of a human's day is high leverage work. And so I think the companies and the products that are doing this well are recognizing who it is they're working for and either A, providing them more leverage or B, making it so it's easier to interact with computing. I think the companies that are not going to be as successful are trying to replace humans. 

Pablo Srugo (14:44) 

you don't think like this whole AI agent thing, like take a whole workflow or set a workflow. That AI agent is kind of the buzzword today. You know, that's what everybody talks about. Does it mean anything to you? 

Aaron Goldsmid (14:54)

I think it is not– If I was doing a zero to one startup right now, that would not be where I'm going. Right? Because you've got two things going on. One is Waymo can have an infinitely better driving record than your average human. But if it gets in one crash, it's the front page of every paper. I don't think humanity instinctively is rooting on the agent to succeed. And so unless you can guarantee perfection and outperform, you're not gonna create a product people love. And if you're making a zero-to-one product and you don't have deep conviction that you can make someone's favorite thing, then you're probably not solving for the right things. And so I also think that the… are AI agents going to eventually get there? for sure, right? But I think they're going to get there for relatively low leverage work. And I think that if they are interacting with a human, I think the human's gonna get pretty annoyed pretty quick. So, I mean, what's the common case? Like a sales development rep, let's generate that email and replace all SDRs. I think that's a, maybe you can send the first email, but what happens when you respond and… you know, the AI is there. Like at some point the prospect is going to go, I don't think this is good, this doesn't feel personal. This feels cold. This feels like the same form letters, but in a different way. Right now, could you say, hey, for the same thing, like, can we give you all of the information? So in our case, we have a big immigration practice. So a founder needs to get a visa in a new market. they want to sort of relocate or, or they're transitioning. some of their work staff across global lines. I would not trust an AI generated visa application with no one touching it because the cost of getting it wrong means you can't relocate the place where you wanna go. But I do feel really comfortable with saying, hey, look, we can automate most of the homework, all of the research and give the expert all of the things that they need so they can be super high leverage and they can actually go deeper. They can do better work. because they say I can spend this much time on this project. And so that extra bit of research, that extra bit of work here doesn't happen versus being able to go super deep. i think those are the types of products that I have a high degree of confidence, especially in the near term, being able to be successful.

Pablo Srugo (17:21) 

You mentioned something earlier, you said building products people love, which I think is the core thing on the wave, like on zero one, when you go into course product market fit, like that is the kind of your first stop, like building something that people truly love. And I guess I'm curious, because you've seen so many products that people ultimately did love. you saw many at scale, but you also work with early stage founders, especially in those early days. I mean, there's a point at which you have so much, so many customers, you have revenue up into the ride. Like there's so much signal, you know people love your product. But in those early days, when you're kind of shooting in the dark, what are some signals that people actually love your product versus it's kind of good and it's just always gonna be like, you know, you might get to a million a year or you might get to two, but like, you're always going to be pushing a boulder up a hill because you're never going to have that pull that comes from people, you know, really loving your product.’

Aaron Goldsmid (18:12) 

I think there is- it's a couple of ways. If you're, you know, A VC and you're writing a check, you need to have conviction that whatever that company you're writing your check can potentially return your entire fund. And if you're not, you shouldn't be writing that check. And I think if you're a founder, if you don't have conviction, it's like, this would be a thing that will, that there will be a product that someone loves, keep trying. Now, in most cases, if we go back to those books, it's products that the founder themselves like, this would change my life. Now the problem is that's N of one, which is woefully inefficient. And so, if you look at our situation, we had two founders from different parts of the world who were here as students and being able to sort of figure out how remote worked. To make remote work work? Like it was really hard. And so they solved a problem that they themselves experienced. Perfect example of when I was looking after Messenger, we lived in America where SMS was free. WhatsApp was built out of Europe where SMSes were expensive or Uber was built in San Francisco, not in New York where taxis weren't. San Francisco taxis never worked. So the first version is - like most founders have lived it. This is not the McKinsey arm length, I see there being a market here. It's like, this would change my life. This would solve my problem, right? It's the Airbnb, like I need a place to stay during Dreamforce. That's your first customer. But then it's like, how do you make sure that you're just not, it's not just building for you. And, you know, think the first thing is, look, you have to go, go back, you know, you go to that customer, you're spending time with them, you're, they're a design partner. And at the risk of being trite, like you'll know it when you see it, because it won't just be like, this is helpful. They'll be like, how do I get that now? Right. And that is like, when you see those moments, you see their eyes light up. But the only way you do that is to be really close to them and be really, really tight. I think what that then introduces is like, okay, now N is two. Could you still just have found the one person who thinks the same way you do? Which when I talk to a lot of early startups, and I think we did this very successfully at Deel, the conventional wisdom is like, do the narrowest thing possible, delight that one customer, and build as little, rent as much infrastructure as you possibly can. Which is great to get to say, okay, yeah, we may have found product market fit, but it really doesn't tell you whether I found a TAM or a SAM that gets back to that returning that VC's fund. And so the way that I advise is like, yeah, do that, but do that like three times in three different segments. So if you're doing something in regulation, do like, like for us, it's like, okay, we need to do something for a small sort of tech-centric company, something that's like a mid-size, more traditional company and maybe like a large, you know, regulated enterprise or something that is, you know, solving for different use cases. And by going and going deep on more than one case and then investing in that infrastructure to be able to say, I know I need to have enough optionality that as I pivot, not necessarily on product market fit, which could be product market fit, but a lot of that pivoting is on how you, the interaction, the nuances, the details, you need to build that optionality. And that is your infrastructure. Like for us, it was building all our own financial rails. It was opening up 120 entities and markets so that we didn't have to worry about someone else. It's building our own payroll engines. because then you had that continuity of service where you're like, no, this is the way that someone really wants this. I can do this thing that has fundamental human product market fit, but if I do it this way, they'll love it. If you're renting that from somebody else, you now need to convince them to go change their product in order for you to do the right thing. And so it's a bigger swing. It is a riskier play, but, and so I think still maybe early, early run the same play, but once you think that you start to find traction, I think you need to diversify and you need to look about How do I build optionality? Because even if I have the right product market fit, the way that it works is going to change dozens of times before I'm at scale.

Pablo Srugo (22:02)

 Honestly, one of the mistakes I see probably most commonly on the product side is - trying to describe it in a succinct way- But it's this idea that people, they start with a product, they go to customers, customers start saying something along the lines of like, it's really cool. But if you had this thing, then it would be really cool. And then you get into the cycle of like one more feature, one more feature, one more feature. Have you ever, and then you end up with this really convoluted, and I've seen this really build out products that could do a lot of things that have like very little to no traction. Does that ever work? Like my experience is it ultimately in the early days is about getting one thing really right. And it's almost never the case that if you just add another thing at some point, you'll like, you know, you'll hit it, right. And you'll be like, yeah, now I've got the 15 features they really wanted. And so.  Now I've got that customer love. Have you ever, like, is that your experience as well?

Aaron Goldsmid (22:57) 

Yes, but I wish there was an easy recipe book for this, but if there was, you know, our-

Pablo Srugo (23:01)

No clay book here, man, 100%.

Aaron Goldsmid (23:04)

 So I do think the needle to thread is, am I solving an unrequited human need? Like am I basic, you know, sort of neurological, emotional, you know, am I somewhere in the hierarchy of needs? The random walk. A lot of people, you'll hear some of the most successful companies on the planet. They'll be like, we knew this was going to be the killer feature. I don't know that I think that was, probably. So I think great teams can recognize when they've found it and they know that when they found it, this is the thing. This is our thing. We are laser focused on this. You still have to have that, that like hierarchy, the need, the human need, but then when you find the thing that that. So I think there is a sort of a random walk that's there, but It's not, oh I just need this. If you added this extra feature, it would work for my workflow. It's here is how I'm going to interact with it in a different way. Here is how I'm going to solve problems for me in a different way. Not just, well, you know, I need to export a file here. That's, you know, and, a lot of companies can get, and I'm, picking a very easy, obvious, you know, red herring, but a lot of companies, a lot of early founders can get lost in that and effectively find themselves waking up and being a custom development shop. And you know, wasting a lot of money and time. But I do think it's the, you can have a thesis about why this is gonna be fundamentally different. And I think call it the aha moment. i think, you know, in chemistry, you need to have all the ingredients, but then you need to have some activation energy. So it's like, what is that moment that sort of gets them over the hump? That's overwhelming. And if you can't convince yourself that this has the chance of that, then probably not worth trying. And, but also don't hold it precious. But once you do stumble upon it, have the humility that you may have stumbled upon it by accident, but when you do, great teams recognize it, that they found it, and then they execute on it with extreme focus.

Pablo Srugo (25:03)

The only thing that comes up in that, and usually probably once you've hit that mark and you start feeling that customer love, is hiring that first PM. What's your thoughts on that? When should you hire your first PM? What have you seen there? 

Aaron Goldsmid (25:15)

This is- I mean- having seen a lot of first PM’s, it's a tough gig, right? because at some point the first PM is always the founder, even if they're building for themselves, they're the ones that no one's going to sell your product better than you are. They are that. And so in your average sort of setup, you know, there may be a technical founder, a non-technical founder. You know, folks have had some software experience in the past. They've worked at some good shops. You know, what is the job of the first PM? It really is giving leverage to the founder. And so that's tough, right? Cause that's not the most exciting PM job. On the flip side, I've seen founders bring in amazing seasoned product leaders and three months later, they're no longer there because the founder was not in a place, nor should they be in a place to give up that product vision to someone else. Right. And so, you know, you really should think about that first product manager. And sometimes it’s not even a product manager. Sometimes it's a non-traditional role, like someone who will give you a lot of leverage. Like, could be a designer, it could be, but they are able to take your voice and amplify it.

Pablo Srugo (26:22)

 Is it like managing requirements? Is it like working with, you know, kind of being that glue between, you know, design, engineering, sales? Like, what do you find is like the main thing that that first PM tends to do?

Aaron Goldsmid (26:33)

I think those are necessary, but not sufficient. So the mechanics of product is required, right? Like, be in the details. making sure that everyone's aligned, being really thoughtful about prioritization. But I think it is the make or break is finding someone that you can get shorthand to really quickly. And what does that mean? Like, I could be like, I just need this over here and do the thing over here with the thing. And they know exactly what to go do with that. Right? When you have those types of relationships, you know, they kind of can just intuit what you're trying to do and fill in the gaps. That's when you have a really successful first product hire. And they're still bringing you ideas and challenging you and pushing you. But, you know, recognizing that their sort of job is to be, is to sort of, you draw the sketch, they build the full building.

Pablo Srugo (27:22)

 I typically see PMs, like the first PM be hire number 30, number 40, something like that.

Aaron Goldsmid (27:27)

I think that's like the right sort of time to think about it because you've probably raised enough money, you've seen enough traction, you're starting to, you know, your constraints as a founder is, you know, focusing internal and operations or externally on growing your business, also managing future rounds of investment. And that's where they start to say, look, making an okay product is pretty easy. Making a good product takes a lot of work. Making an excellent product takes an enormous amount of labor. 

Pablo Srugo (27:57)

What's the difference between an excellent product and an okay one?

Aaron Goldsmid (28:00)

I'm going to borrow the apple, or any time Dieter Rams right? Like it's those moments of design, like how did they get that? That bezel, like how does it feel in my hand? How does it interact? I mean, I can't imagine how much time someone spent on these scrolls, mechanics and physics on TikTok, right? Like, cause whatever they did, they just got it right. The filters on Instagram. I'm picking easy consumer products. I think there's an easy way to connect to it. So like once you did that, that takes that level of tuning to create something.  in games. Tuning games is enormous amount of work, right? Or look at what Pixar does. The delta between good and great is exponential in terms of effort. And so, when you're a founder, hit that 30 to 40 mark, for you to take that level of attention to detail and push that forward means you're probably neglecting other parts of your business. And so you need that human that has shorthand to be able to sort of like take your vision and get as much of that right as possible. And so you can come in and You can do the tweaks at the end. 

Pablo Srugo (29:09) 

How much of product these days is about the vision? There's like the vision of where things should go. There's like the design of how things look and feel. And then there's like this growth element. Like you mentioned the TikTok thing. That's where my mind went. Like a lot of this stuff is just test, you know, the whatever million blue, blue shades of blue that Google tried or whatever that thing was, you know mean? Like just like test out a bunch of different things and whatever, whatever moves that KPI the most, that's the best product. Where do you, where is product these days? Like what kind of viewpoint matters most? 

Aaron Goldsmid (29:40)

I think where sort of I am. I gotta do all three. You know, if you don't get, you don't get to choose anymore, but someone really smart gave me the best advice about strategies, they said it’s something you do twice a year, the rest of the year, you just do your job, right?, but, honestly, every great strategic insight I've ever had and every great strategic insight, I've seen people who are much better than me have has come from being in the detail, has been in that nuance, has been in the weeds. Why did someone hit that versus that? That's interesting. I wonder. That is, you know, but what all that means is that the difference between the growth hacking and the strategy is not as much as you think. In fact, it's more of a circle than it is a sack. And, you know, when I think about sort of the new features versus optimization versus extension, you know, I guess kind of, you know, prioritization, managing engineering. Here is how I think about it. Every time you introduce some, like, let's say you've gotten the product market fit generally correct. You're solving a need that's real and matters. Then it's really about, in sales they call it objection handling, but it is every time you cause a little bit of friction, life gets in the way. I think of the spectrum, right? On one side you got a tax authority, and on the other side you got a game, right? The tax authority, you could build the worst website, it literally could punch you in the face every time you opened it, you'd still finish it, because you have to do your taxes, right? Utility trumps everything. That game, the moment it's not fun, the moment the level's too hard or it's too easy, you're onto the next thing. You're competing with Netflix, you're competing with whatever. And so what is the difference, and you've seen this with a consumerization of SaaS over the last probably decade, is how do you get moments of friction where people are gonna go out of the flow cognitively and life gets in the way? If you just make it easy, you'll see that traction. And so that's like the injection hand way. Just like take away the reasons we're not to finish this. I think Google 40 shades of blue. The only people who have enough traffic to do that is Google. So it's irrelevant to anyone else. But for the rest of us mere mortals, like it's, I have to upload a document. Where's my scanner? Right? Like, or I could go look up something or that's something, you know, I'm having to on my phone, copy and paste a large text or string or something. I mean, just think about it, I'll give you one example that we learned from Twilio is that remember when Apple, that 2FA, the SMS for 2FA, when that little sheet came up and said, just add this, 22 % increase in conversion, small move, but significant. Why? Because I didn't have to switch out an app to go to another app, which is my, I got a text message from someone else. I got like, but what movie's playing? 15 seconds later, I'm somewhere entirely different. And so by keeping it in context in situ, it can be a really, really convenient flow. And so you, just didn't give them time to have life get in the way. 

Pablo Srugo (32:28)

People think of the user as, you know, you have to build for a five-year-old, like they're, they're stupid or whatever. And I think the right way to think about it is like, they just don't care nearly as much as you do. You know what I mean? Like they're going through flow and they click on another app and something else comes and it's distracting or maybe just more important to them. And your app is done. It's dead. Like it's the bag burner and that's the 22 % difference is just like. You want someone to get some done, like make it easy and keep them there and keep them focused. Cause there's a million other things they can be doing.

Aaron Goldsmid (32:55)

 I mean, I would say my team's a deal. We do an immense amount of work. So our clients or our workers don't have to, we will go and look and do a ton of automation and a ton of  looking at data sources and a ton of complexity so that the user doesn't have to lift a finger. And I think it is a cultural thing that you need to instill in your teams that you're willing. That's really hard. That's really complex. but you're actually just asking your user to do the work that you were unwilling to do.

Pablo Srugo (33:25)

The other thing I see a founder struggle with is prioritizing between bug fixing new features and then feature requests from existing users, right? Like I just find often, cause you're taught to listen to your users, right? And there's a right way in a, I think that's the right message, but there's a right way and a wrong way to do it. And I have seen teams get too caught up in this long list of things that need to be built into the product because customers are asking for it. And you just kind of have the sense that you're spinning wheels. You're fixing, you're doing all this stuff, stuff is happening, but none of it is really making your product much better. None of it is getting you new clients. I don't know that you could ignore all of it either, but it's taking all these resources and then not really moving the needle. Have you seen that? What's your take on how to optimize there?

Aaron Goldsmid (34:17)

it is rare a team doesn't get stuck there. i think that is like all entropy leads towards, you know, the external customer commits. Let me get through that list and with whatever, that becomes the 80 % and then I'm done with 10 to 15 % left. I think the way to look at those and assume there's certain clients where you don't have a choice, right? Like if they're a big logo and they're like, this is a make or break for us, like it's just the cost of doing business. But for everything else, I challenge my teams to go, why are they asking for this feature? What is the need they are trying to solve and take something stupid, right? Like I need to download an invoice. Well, why do they need to download it? What is the thing they like, cause they need to upload it in this other system. Like, well, they don't need download anymore. They need an API to go integrate with that. And so like what you just kind of, and to be fair, I think maybe the right answer about what a five-year-old should be is about keep asking the question why until they run out of answers. Because you're solving for the underlying need. And then that list of commits or feature requests, you're seeing through it to what's the underlying problem. And when you do that, you're unlocking a lot of folks' problems. And I think that's one version of it. And then I think other version, and this is gonna sound a little schizophrenic, because we talked about this at meta lot, it's like solving for the marginal user, being calculus. Like what is the differential, like if this is the human who is most theoretically most likely to be successful at this task, but failed. And if you just fix their one little thing, right. And then that compounds. And then all of a sudden you can go from linear to exponential growth. a lot of that, again, classic growth hacking type stuff. i think when you marry those two and the way that I look at it for teams, I actually talk about like a stock portfolio, right? Like you, you're going to plan your financial thing. You have like this much in fixed income, this much in equities, this much in cash. And that changes over the lifespan of a product or a lifespan of a human. It's like, what is your portfolio for where you are? how much you need to spend on optimization, the growth hacking, how much of this is fundamental new build, how much of this is like learning experiments or crazy stuff that would like low probability, but if it hits, it's gonna be big. And if you think of it that way, over multiple turns, you're gonna have a higher outcome, just like investing. 

Pablo Srugo (36:22)

Perfect, man. Well, listen, Aaron, it's been great to have you on the show. We'll stop it there, but I think we've covered a lot of topics, so we'll be really helpful for founders that are kind of going through that zero to one phase and just thinking through what they're doing on the product side.

Aaron Goldsmid (36:47)

I really appreciate the time. It was a great conversation. For all those founders, deals five, six years old, there is a path to being an enormous tens of billions of dollars of valuation in a real short period of time. I think as we move forward, there's going to be a really interesting lesson to learn from how we approach building products here.

Pablo Srugo (37:11)

Crazy. you guys were worth like, last was like what? 12 billion valuation or something? And when did you guys start? What was the year founded?

Aaron Goldsmid (37:17)

I don't know, but I think we're, I think we maybe just had our sixth birthday, fifth birthday. 

Pablo Srugo (37:22) 

That's crazy, That's insane. 

I just gave you content that you liked so much, you actually listened to the end. And guess what? You didn't pay a single dollar. Not only that, I didn't even put any ads in your face. So you just got a bunch of content for free. And now that I've delivered that value, I'm asking for something in return.

Open your app, open Apple Podcasts, open Spotify, open whatever app you use to listen to this and hit that follow button. It's actually going to help you because it's going to help you make sure you don't miss out on the next episode, which you like so much that you listen to the whole thing.