Welcome And Meet Linda Tong

SPEAKER_00 0:02

Welcome to Uplink, where we explore the world of digital infrastructure, uncovering the technology fueling AI and cloud innovation with the leaders making it happen. Alright, I'm here. We're in San Francisco. We're with LT, an old friend of mine, Linda Tong. We first met when you were at Cisco, or actually you're at AppD. You ran all product at AppD for a period of time. And I was in the Thousand Eyes role and we were working together a lot, actually. It was a lot of what the time is. Yes. And there was full stack observability, was this thing, and there was all these components to it. Your background is probably the most fascinating background. I'd love to like bounce through that really quickly as well. But you're currently the CEO at Webflow, which is an incredible company that is, and I'd love to you to share a little bit of this what Webflow was, where it is today, and what's going on with AI, because that's that's also changed. We use your platform, so we're very grateful for that. Yeah. The marketing team love it because they can actually build websites really, really quickly without having to talk to our developers or whatever it is, I guess it's so the web development team. So for all these um events and whatever it may be, they're just punching out um great, beautiful, beautiful content that they can deliver themselves. So we use you a lot for that. Um I know you've got incredible scale on on on the the web uh the webflow uh side, but also um your background is really crazy. So I can't remember what you landed at Google, but were you somewhere before Google? Was Google or the first one?

SPEAKER_03 1:30

I did a very short scent in investment banking, which is sort of like soul sucking.

SPEAKER_00 1:34

Yes. Well, you were at Yale before that, I think. Yeah, I was at Yale.

SPEAKER_03 1:37

Citigroup.

SPEAKER_00 1:38

Yes.

SPEAKER_03 1:38

So Yale, I studied economics. Oh, yes.

SPEAKER_00 1:41

Of course you did.

SPEAKER_03 1:42

My senior thesis was rewriting the Medicare Act of 1965. So clearly like very applicable to my job.

SPEAKER_00 1:48

Totally makes sense from a product side. That's right. Yeah. Uh and then I went But also you you you got yourself through college. You were hustling at that point in time.

SPEAKER_03 1:56

Hustling, doing a lot of jobs. I had three jobs.

SPEAKER_00 1:59

Yes.

SPEAKER_03 1:59

Um, and then I had a poker hobby on the stuff.

SPEAKER_00 2:02

This is one of my favorite ones. Yes. You actually were doing some real high-stakes poker to fund your way through college. Yeah.

SPEAKER_03 2:08

Online poker helped a lot. So party poker, poker stars. It's all the good stuff.

SPEAKER_00 2:13

It's so cool. Yeah. Actually, you gave me a really good book to read.

SPEAKER_03 2:16

Um Thinking and Bets.

SPEAKER_00 2:17

Thinking and Bets, yeah.

SPEAKER_03 2:18

Annie Duke. Yes.

SPEAKER_00 2:19

Awesome book. Um actually gave it to the team. It's so good. It actually used it a lot. It's like, um, oh, you're so convinced of that, are you? What would you bet on it? And so it's like changes the dynamic.

SPEAKER_03 2:28

And more importantly, it's about not resulting. And so it's a it's making sure that when you make a decision, and knowing that that decision is a bet and there's different outcomes that could happen. Correct, it's not guaranteed. Regardless of what the outcome is, making sure that that doesn't influence how you make that decision.

SPEAKER_00 2:44

Yes, yeah. Yeah. It's uh it's awesome. But I got all these little notes that I wrote down from it, actually. I should really look at it. Anyway, I I love that you had a very you must have had a very mathematical brain though, given that you studied um something very landed, so you did investment banking randomly.

SPEAKER_03 3:02

Yeah, did that. I mean, that was sort of the interesting everybody was doing investment banking or consulting. So I did it just to follow the crowd, and I

Poker Bets And Better Decisions

SPEAKER_03 3:10

was terrible. Yeah, that it was not for me. It's interesting.

SPEAKER_00 3:13

It's funny, I deal with a lot of investment bankers now. Yeah, yeah. Yeah, it's you wouldn't have to be a lot of people.

SPEAKER_03 3:16

We talked a lot too, and I respect them so much for what they do. But my only skill that came out of it was learning how to traverse Excel without using a mouse, right? That was like the one if you could do that, you're upset uh all the keyboard shortcuts. CS and then writing Visual Basic.

SPEAKER_00 3:31

But so good, yeah.

SPEAKER_03 3:33

Yeah, I did that and then realized very quickly it wasn't for me, quit that, and applied for any job that would take me. Any job. You know, I applied to Microsoft as a test engineer, which by the way, I don't know how to code. I actually never took a CS class. Then I applied for a job at Google to be a product marketing manager, and I applied for a job at Abercrombie and Fitch to do like men's jackets, like outerwear.

SPEAKER_00 3:57

Of course. This makes complete sense.

SPEAKER_03 3:59

Yeah, very do men's jackets, like uh supply chain for designing men's outerwear and actually like figuring out the supply chain to produce these jackets and get them in stores, which I have zero fashion sense. I have zero supply chain background.

SPEAKER_00 4:12

Um made sense.

SPEAKER_03 4:15

It was a job that was an investment bank.

SPEAKER_00 4:17

And true, yeah, okay.

SPEAKER_03 4:18

Uh luckily of the of the three, um, you know, uh I got offers at all three, but Google made the most compelling offer because they basically were like, this is a job we think you can do.

SPEAKER_00 4:29

Yeah. What what year what year was that?

SPEAKER_03 4:31

Uh 2007.

SPEAKER_00 4:33

Oh yeah.

SPEAKER_03 4:33

Oh my god, I'm aging myself.

SPEAKER_00 4:35

No, no, no. But yes, um, but at the time I'm trying to remember the time because Google was so hot. And then I remember in 2004 I first came across Google, like in being 2003, probably. I've still my my email's like mick.read. Yeah. You know what I mean? So um yeah, I was like, I know I'm proud.

SPEAKER_03 4:49

I have linda.tom.

SPEAKER_00 4:51

How good is it?

SPEAKER_03 4:52

There's hundreds of us out there, and I got I got that name. I got the name you say it. I feel so proud.

SPEAKER_00 4:56

So was it 2007 or eight? What did you just say?

SPEAKER_03 5:00

2007.

SPEAKER_00 5:00

2007, so just before the world caved in.

SPEAKER_03 5:02

Yep.

SPEAKER_00 5:03

Yeah, exactly. So you joined um Google. And I think you also you worked throughout that time with uh current CEO as well along the way. Well, you had different lot of experiences, but yeah.

SPEAKER_03 5:13

I was like a little nobody at the bottom of the totem pole. Um, but I the team that I joined when I joined Google was Chrome. And at the time, the VP for Chrome was Sundar. And you know, got to like I got to sort of, you know, many steps away, but got to work on that team and help launch Chrome to the world, which you know, for um for Google that's really unique because it was a first launch that was multilingual. We did it in 52 languages. It was also the first launch that was multi-OS. So usually they would launch on Windows first, then unlock like uh Mac and then Linux.

SPEAKER_00 5:45

And we're talking Chrome the browser, yeah.

SPEAKER_03 5:47

Chrome to browser.

SPEAKER_00 5:48

Everything that was just you launched the original. Yeah, yeah, yeah, which I use every day. Yeah, yeah.

SPEAKER_03 5:52

And so, you know, this was a huge launch for Google. Um, and it was like it didn't leak, which is like pretty unique.

unknown 5:59

Yes.

SPEAKER_03 5:59

Um, except it ended up leaking. Um, but it was just this magnanimous launch. Like it was absolutely incredible. And I got to work with some of the most talented people, uh, who many are still actually at Google, but um that like that was probably the cool thing you could work on. And I got to launch it to the world. And on the marketing side, one of the most interesting things was like, what's the campaign gonna be?

SPEAKER_02 6:25

Yeah.

SPEAKER_03 6:25

How are you gonna talk about it?

SPEAKER_02 6:26

Yeah, browser.

SPEAKER_03 6:27

And our original campaign was Chrome, the browser built for apps. Because what was unique about the browser at the time was it had this VA engine that could basically run complex applications really quickly.

SPEAKER_00 6:38

In the browser. In the browser. So yeah, which didn't exist. Yes.

SPEAKER_03 6:41

If you think about apps like Facebook or Amazon at the time, these were complex, you know, web applications that needed a powerful browser to run quickly versus like a simple HTML website.

SPEAKER_00 6:52

I remember it just being so fast.

SPEAKER_03 6:54

It was super fast. Yeah. It was lightweight. Yes. Um, and it's just it didn't have a lot to it. It was just like a search bar and then like basically a browser. And so our that was our claim to fame. But then as we were testing out the marketing campaign, the browser built for apps, um, we realized that no one knew what an app was. Because, you know, it's 2007, everyone's like, oh, everything's a website.

SPEAKER_00 7:16

Yeah, that iPhone really wasn't it wasn't there actually at that point. Exactly.

SPEAKER_03 7:20

And I was like, actually, well, some of these are websites and some of these are web apps. And the difference between the two is sort of lost in the public.

SPEAKER_00 7:26

Well, they don't care either. It's like an interface, however it comes, yes. Yeah.

SPEAKER_03 7:29

Well, the crazier thing was as we went deeper into this user research to understand, you know, what do people understand about websites and apps, I realized very quickly when asking a couple of people, I was like, so what what browser do you use then? Yeah. Right. And most people were like, oh, I use the Google browser.

SPEAKER_00 7:43

Which is Internet Explorer, but they had like, I guess, Google as the default homepage.

SPEAKER_03 7:47

Google was like the default homepage. And so people actually never. There was no Google browser at the time, right? And so they're all like, oh yeah, yeah, yeah. I use Google. And I'm like, really? Like, what is the icon you click on?

SPEAKER_00 8:00

Has it got an E on it?

SPEAKER_03 8:01

Yeah, no, there's a blue E. And I was like, Huh. Yeah. Like, you don't know that that's Internet Explorer. And I remember asking a couple of people, I was like, So how do you spell internet? E-N-T. And I was like, oh god, this is the worst. But that, you know, doing that research and realizing that everyone thought that they were already using the Google browser when there wasn't a Google browser. They're in fact using IE or Firefox.

SPEAKER_00 8:26

Means you don't have to convince people that you're a safe bit.

SPEAKER_03 8:28

Exactly.

SPEAKER_00 8:29

You're I already do use it. No, you just need to click to actually one. Here's the new version.

SPEAKER_03 8:34

And so we changed our entire campaign. We stopped talking about apps because we realized it was a little bit too complex. And the entire marketing campaign was Chrome, the browser built by Google.

SPEAKER_00 8:43

That's it. It's so simple. I'm pretty sure I remember that. And it was far. Oh, there was like quick for fast or something like that.

SPEAKER_03 8:48

The icon with like what spin really fast.

SPEAKER_00 8:50

Yeah.

SPEAKER_03 8:50

It was uh, you know, marketing genius.

SPEAKER_00 8:55

And I used to run my parents' real estate IT. The first thing I'd do is delete uh Internet Explorer and download download Google, and then start downloading all the other stuff. So yes. Uh but was that the one that you told me some of the story? I don't even know if you can share this publicly or not, but the one that that I think I can't remember the link but leaked the marketing team accidentally sent.

SPEAKER_02 9:13

Yeah.

SPEAKER_00 9:14

Was like sent something to analysts or something. Was that the one? Or was that was that Android?

SPEAKER_03 9:17

Or was that it was Chrome. We we had this comic book to explain how complex this browser was and how special it was, and a lot of the the story behind why we built it and that comic book, which was meant to go out sort of the night before the launch to all these like major tech reporters. I somehow got to this German reporter uh weeks ahead of time, and person got it on Labor Day, scans every page, uploads it to like the verge. Yes, and suddenly I'm getting a call on Labor Day, and it's like, what's going on? What happened? Why did we leak this? And like all of us are just driving down to Mountain View and war rooming it and like launching, we're like just launch the browser. So we're just turning everything on, which is absolutely wild. But um, you know, luckily it was pretty well baked.

SPEAKER_00 10:09

How early was it before you were in a couple weeks? That's so funny. So everyone's prepared for this perfect launch.

SPEAKER_03 10:14

Yeah. And then Nemo were like um probably got the hustle.

SPEAKER_00 10:20

As you told me, I think you were trying to get people in because like you had to um everyone's on holidays or whatever.

SPEAKER_03 10:26

Yeah, yeah.

SPEAKER_00 10:26

So a lot of you would have been like last month, this Monday.

SPEAKER_03 10:29

Uh no, no, Labor Day, that's September. Oh, yeah. You're thinking Memorial Day.

SPEAKER_00 10:31

More day, sorry. That's American holidays. Too many, yes, yes.

SPEAKER_03 10:35

But yeah, Labor Day, literally Labor Day. It's the day to take off because of all the labor. Um, but yeah, so a lot of people are coming in on a day off. But uh it was honestly, it it was my first time being in the trenches, like truly feeling like in the trenches with a team. And it's the most addictive feeling you can ever have. Like you, like you chase that feeling all the time because being in the trenches with folks, it just you feel like a team and you feel your impact right away. Yeah. And it just it makes you so grateful for everything you can do. And um, whenever I work with any other team, I'm looking for that. I'm looking for these moments where we can come together and achieve something.

SPEAKER_00 11:16

Yeah, it's funny, like in these moments of utter chaos or

Landing At Google Without Coding

SPEAKER_00 11:20

panic or whatever, everyone it or the ones that some folks lift and then pressure, some tope folks just show up in a way you never thought. And then it actually you end up building a team of people around you that you know will always show up like that. So if you if you see someone hustle like that, the next time you have a problem, you just bring that same team together and you know you're just gonna be able to execute through it.

unknown 11:37

Exactly.

SPEAKER_00 11:38

Which I suspect you ended up because you you also did you did Android didn't you as well. Is that yeah?

SPEAKER_03 11:42

So after Chrome, I got to migrate to a new like I was in a rotational program because I'm you know, I'm like little junior of the junior. So I'm in this rotational program, and they're like, okay, you have to move. And I got my top pick and I got to move to Android as my next rotation, which was out of this world.

SPEAKER_00 11:59

Um Android hadn't launched that time. I had not launched it.

SPEAKER_03 12:02

So you know, Google had acquired Android. Uh separate company, any Ruben had started on the side, and they had acquired Android, and they were building uh the operating system itself uh and the first device and the app marketplace. And all of that was launching at the same time.

SPEAKER_00 12:18

And was that what year was that?

SPEAKER_03 12:21

That was gosh, I think it was 2009. Yeah, so iPhone was like 2008, early 2009. Because the iPhone had launched first. Yeah. And then we launched Android basically like early 2009 or late 2008. Um, and basically the T-Mobile G1 was first. Uh, and then it like in quick succession, we had devices across the four main characters.

SPEAKER_00 12:47

Of course, because it's so it was open source, like source.

SPEAKER_03 12:50

Anyone could start to do it. Yeah, and it was amazing. Like it just took a still have probably the most Yeah, it's I think it's the most popular smartphone at this point in terms of distribution of the operating system.

SPEAKER_00 13:01

Yes, yes, yes, yes, because every Samsung and every Exactly. Yeah, and all the non-Apple folks out there, which it's probably hard to do these days, but still, yes, they've all fought for it. And then a lot of um uh country developing countries were massive in that space.

SPEAKER_03 13:16

Because you can create sort of like any spec of a phone that mats matches sort of like the minimum structure Android and then download the OS and run it on there.

SPEAKER_00 13:25

Make it a smartphone of sorts. Okay. Um tell me the one story, I don't know if you share it, but the the one story where you ended up on a plane in an Android outfit.

SPEAKER_03 13:34

This is you probably can't share. But we you know, one of the coolest things about Android was that there's a unique culture to the team, right? So we named every OS version uh after a different dessert. And it just went in alphabetical order. And then at the end of it, we would essentially um print these massive versions of the dessert and put it in the front lawn in front of our building in Mountain View. So the first, you know, the first one that most people are aware of is cupcake. Cupcake was like sort of the first episode. So we printed this giant plastic cupcake and like plopped it in front of the building. And then, you know, there was like a donut and like uh an eclair and then fro yo and um like gingerbread. Uh like gosh, what was H? I can't even remember. But like it just was every letter there was a thing. Uh, when we got to K, we actually did Kit Kat. So we actually had to get like a branded thing. Um, and you know, we would celebrate by giving everyone these dessert treats uh when we launched. So you have like cupcakes and donuts and and whatnot.

SPEAKER_00 14:36

Basically getting everyone like whites and creasted all there. Yes. Exactly. Yeah, yeah.

SPEAKER_03 14:42

No, I'm just kidding. Um, but one of the things was we we wanted in addition to all these like sort of um desserts on the lawn, we were like, we should get a big Android.

SPEAKER_01 14:50

Oh yeah.

SPEAKER_03 14:51

And we should also have like androids everywhere. Yeah, we should get like a cool Anthro band. And there's some in Korea, and they're like, hey Linda, go to Korea and get one. And the thing is, these things are so big that you'd have to essentially pay for them, print them, put them on a ship, and ship them over. What is it?

SPEAKER_00 15:07

Like a sorry, it's like a it's like an app.

SPEAKER_03 15:09

A big plastic like Android suit, a suit that you can sort of wear.

SPEAKER_02 15:14

Okay.

SPEAKER_03 15:14

Um, and they're like, Well, you know, you could just get one and and wear it onto the plane, um, which we attempted to do. And I was sort of the I was the nominee. Um I I volunteered as trivia. So I went I went to try and attempt that. Um, but well, we ultimately ordered a bunch and just shipped them to the United States and in a shipping container. But um it was a really the culture of the team was really incredible. Just like the things that we did, it was you know, it was um it was almost like a subculture inside of Google that did all these like really funny, silly things, but it it brought us together to launch all these different versions of of the operating system.

SPEAKER_00 15:55

So you yeah, I think you went to I know I don't know where you went for, you went to a startup actually from there, wasn't it? Or and you built you built some bars or something.

SPEAKER_03 16:03

My my claim to fame inside of Google is like every everybody did this thing called 120% time or 20% time. It was really just extra work on top of your full-time job. And my hobby was sports. I had proposed sports.google.com so many times and it got shut down every time. But I had found sort of a community of people who also love sports, like me. And so we'd worked on things in my 20% time of like the sports one box. So when you search for sports scores, what shows up there is something I yeah, yeah.

SPEAKER_00 16:33

Now okay, that's cool. Yeah, I use that all the time. Like, what's the score? Right? What's the sport right now?

SPEAKER_03 16:37

Like what's the next game?

SPEAKER_00 16:39

It's actually very it's so accurate for random things, it's always they score a point or something.

SPEAKER_03 16:46

You know, and so Google's top one box was always the sports one box because that's just it gets the most traffic. Um, so worked on that, also worked on like a fantasy, like you know, fantasy football app and a March Madness app and a scores app. So did all these like kind of fun things on the side. And like what was key to that was actually the people I worked with, um, this guy Ben Lewis and this other guy, Stephen McCarthy. We essentially did this 20% time project together. And Ben ended up leaving Google to start a company called Tap Joy. And he left with and and built it with his best friend Lee. And as they were starting to uh pick up momentum, he called me and was like, come over, come B RPM. And Steven as well. And so me and Steven went over to Tap Joy. We were, you know, the the first two like sort of full-time employees uh beyond the founders. And then uh we brought over one of Steven's best friends, and so the five of us, as well as another guy from Google that I was close with named Kai, we all went over and really like kick-started Tap Joy. Um, and it was a it started as a mobile gaming company. We were building these mobile apps for iOS and Android. Um, think like Farmville or you know, whatever, and you know, Sudoku and Tap Defense, like all these sort of like fun games, and we weren't very good at it. Okay, like our games sucked. Uh, you know, one was a hit, 26 were like a miss. But we realized really quickly with the one that was a hit, we weren't making enough money on it.

SPEAKER_00 18:07

And so we ended up in like pay a dollar to download it or something.

SPEAKER_03 18:10

Yeah, because it was it was a free game and we're just trying to monetize it with ads. And we weren't making enough money, and so we ended up building out like a monetization mechanism to like show more ads. So we want to maximize the number of ads that people were seeing because we were getting all these empty ads. Uh, so in the advertising world, it's like we didn't have enough fill rate, right? There weren't enough advertisers. I don't know. And so, you know, sometimes we just have no ad. And then the other thing was we realized that there was no at the time for both iOS and Android, there wasn't this concept of in-app payments. So if

Launching Chrome And Learning Users

SPEAKER_03 18:40

somebody wanted like a virtual good in our game, yeah, like uh Blender pops up.

SPEAKER_00 18:44

I want that Blender blank. They couldn't really bucket.

SPEAKER_03 18:47

That's so bizarre.

SPEAKER_00 18:48

What do you think about it? We can go deep on that.

SPEAKER_03 18:50

Let's definitely go deep on that. Um, but yeah, they couldn't buy stuff. Yeah. And so we created this incentivized ad network where you could actually take an action, like download an app or sign up for Netflix, or you know, do some action that an advertiser would pay you money for, and that money would be silver or whatever it is, yeah.

SPEAKER_00 19:07

Exactly.

SPEAKER_03 19:07

I think we would give you the blender or give you a sword or give you a shield. Uh so we created these like really high value ads that were incentivized through action. And we realized really quickly overnight that this game that was really popular but not making a ton of money is suddenly started making a ton of money.

SPEAKER_00 19:22

Wow. And we turned that's because of the way the ads that you were.

SPEAKER_03 19:25

Yeah, because the ads themselves could generate more money, you know, striving action, and people really wanted these virtual goods and had no way to really pay for them at the time because they're who's paying for the ads.

SPEAKER_00 19:33

Is is it you taking Google ads in that, or is it?

SPEAKER_03 19:36

We went directly to advertisers. Oh, yeah. So we would talk to like um Yeah, like Google could be like, Hey, I'll pay you 35 cents to get me a download of the Google app.

SPEAKER_00 19:46

Okay.

SPEAKER_03 19:46

Or Netflix would be like, Hey, I'll pay you five dollars.

SPEAKER_00 19:48

Oh, so now it's not about the game. Now you oh the game was using the app. The app sorry, the ad ad the app would would click to something, it might be download this. They will pay you 35 cents if you get from your game the app that gets you down. Oh, okay.

SPEAKER_03 20:01

So a good example is you're playing a tower defense game and you're like, I want a super like arrow and I want to like buy it. And normally you would buy it for two bucks. And we're like, look, uh, there's no way for you to pay me two dollars because there's no payment system in this in in iOS. So you'd be like, you know what? Instead of paying me two dollars, why don't you go sign up for Netflix? If you sign up for Netflix, like click on this ad that I have to sign up for Netflix, sign up, and if you do it, Netflix will pay me five dollars, and I'll take that five dollars and transfer it to you as like this you know, set of sweet arrows. So we created a ton of ads that basically did that.

SPEAKER_00 20:36

You basically invented the world that my kids live in where they're constantly playing on something and they have to watch an ad to get to the next thing and make it fast. Yeah, okay.

SPEAKER_03 20:45

Yeah, that was so you could like watch an ad.

SPEAKER_00 20:46

You've created this world. I know I'm so sorry.

SPEAKER_03 20:49

But yeah, you would watch an ad or you would interact with some sort of like an offer, though. They're called offer walls at the time. So you do an offer, and after completing the action, you would get virtual currency and then you could buy, you know, whatever thing you wanted. And it was already, you know, we didn't invent it, we just did it on mobile first. So at the time, Facebook was huge, Farmville was huge, right? So people were playing all these Zynga games and they were doing this already. They were like signing up for Disney cruises and magazine subscriptions and getting like coins in in Farmville and like buying, I don't know, blueberries and whatnot. Um, and so we just took the same model and tested it on mobile, and it worked. And we we turned that into an SDK. We talked to our friends who are also developers building these games, and we gave them SDK, and they suddenly started making money. And we went from being a gaming company to becoming an advertising company. And then we built this network of publishers, so these game developers and advertisers, and we just matched them together via like our one innovation, which was the offer wall, which is a list of all the offers you could do. And we figured out how to close the loop on the monetization, and suddenly developers are making, you know, millions of dollars. Of course, yeah.

SPEAKER_00 21:53

That's why Fruit Ninja guy made millions and whatever. Yes, yes, yes.

SPEAKER_03 21:56

Exactly.

SPEAKER_00 21:56

Yeah.

SPEAKER_03 21:57

And so that that startup went from being, you know. Tens of thousands of dollars in revenue to millions of dollars in revenue in, you know, like ten months. And it was wild.

SPEAKER_00 22:06

That's I mean, that's wild back then. Today that's not wild.

SPEAKER_03 22:09

I know. Today it's like you can go zero to a million like five hours.

SPEAKER_00 22:12

Five hours with one person and chord code and a yeah. Yeah. It's funny, just before you, I had chatted to some other folks who've done exactly that. Like companies of like 10 people. Um had four people and went to a million in ARR.

SPEAKER_03 22:26

That's absolutely wild.

SPEAKER_00 22:26

And out of 25, and I don't know what they're at there.

SPEAKER_03 22:28

It's like probably at like 10 trillion.

SPEAKER_00 22:30

10 trillion. Yeah. In infinity. Yeah. Approaching infinity.

SPEAKER_03 22:33

Infinity.

SPEAKER_00 22:35

Yeah. So then you went to NFL, I think.

SPEAKER_03 22:38

Uh I actually did another startup with all my Android friends. So all the people I worked with on the Android team. Went back and built another phone, built another OS. Okay. Uh it was called NextBit. It was basically we took the concept of Android and said, what if we make a cloud phone? A phone that like literally, it doesn't matter. The hardware doesn't matter anymore. Your operating system lives in the cloud. You could take the phone, log into it, and all your stuff comes down. You drop the phone in the toilet, get another phone, log in, everything comes back.

SPEAKER_00 23:02

It's like um virtual machine. What do that what do they call that? Like citrix for phones or something like that.

SPEAKER_03 23:07

More or less. It's like it just carries it, it carries you in the cloud. And then phones are just sort of like this sort of like lightweight hardware. It doesn't matter what you log into, it's just it's all there. Yeah. Um, so we built that. I did that for a couple of years. And the company sold to a Singaporean company called Razor, which is actually built like um bought a bunch of Android companies. Um, and I uh went and joined the NFL go run product uh because I'm a sports fanatic. Yes. I love football. I love the Bears.

SPEAKER_00 23:38

Nice. Um what is product at an NFL team?

SPEAKER_03 23:41

It's funny, it's uh because actual product at the NFL is the game, it's football, right? The products that I worked on were digital products. So the websites like NFL.com, the mobile app, the team apps, the team websites, yeah, all of our streaming services. Uh I work closely with the network team. We also had all these like side products that, you know, like fantasy football. The fantasy football app is part of it. You know, we were closely uh so the NFL had its own uh own fantasy game that we work closely with, like a bunch of our partners on. We had connected TV apps that's streaming the game. You know, we stream the game in our mobile app all the time. We also built a streaming service so you could actually watch the video on demand, like games afterwards, you could watch all the games that have ever happened. Uh, and then we also work really closely with um the the league itself on like all these things that have come out now that you see now. So, for example, Nextgen Stats was a product that was born in the NFL and is now a sponsored product. So you can see positional data on all the players. Uh, it's like, you know, they have chips in each of their shoulder pads, there's chips in the ball, chips on the flags, and you can actually see positional data as they as they play. So it's just a bunch of the digital experiences. It's all the digital experiences around the game that enhance the game. Yeah.

SPEAKER_00 24:57

Yes. Uh, we we have a customer of ours, um the major league baseball team. They actually built a streaming app for MLB.

SPEAKER_03 25:06

Exactly.

SPEAKER_00 25:06

That streaming app became Disney streaming. Yeah. Oh, yeah.

SPEAKER_03 25:10

They built the top video streaming service. Yeah.

SPEAKER_00 25:12

It's MLBM.

SPEAKER_03 25:13

Yeah. Uh MLB Advanced Media.

SPEAKER_00 25:15

Yes. Or was the platform under them or something? But yeah, they ended up all the Disney streaming was actually from Major League because they had so many games.

SPEAKER_03 25:24

Yeah.

SPEAKER_00 25:24

They had to figure out how to stream them all or what have you. Yeah, and they solved this really hard problem.

SPEAKER_03 25:28

It's super difficult.

SPEAKER_00 25:29

Yeah, it's a really interesting problem. Um anyway, so then you went to App D, I think.

SPEAKER_03 25:34

Yeah.

SPEAKER_00 25:35

And that's where we met. Yeah. So you were an application performance monitoring.

SPEAKER_03 25:38

I know. That's a mouthful. Yeah. It was my first time moving into enterprise. So this whole time I'd been on the consumer side. And then uh I was a customer of App Dynamics when I was at the NFL. And I think I was the most hated customer. Yes. They uh I pretty much told them everything was broken, the UX sucked, it like I didn't like the product, it was too expensive.

SPEAKER_00 26:00

You seem to have an opinion here. Come and work for us and fix it. Is that how it means? More or less.

SPEAKER_03 26:04

I think they're just like, we just gotta get this girl to shut up.

SPEAKER_00 26:06

Yeah, we can let's hire hire her, stop.

SPEAKER_03 26:09

She can't be a customer anymore. So I ended up joining Ap D to run their innovation lab. And um and then very quickly I found that uh, you know, we were inside of Cisco, a lot of change was happening, and they're looking for seasoned leaders to step in and help operate um the organization as you know, we'd just been acquired by Cisco.

SPEAKER_00 26:28

Yeah, it was like a $3.2 billion acquisition or something crazy. It was huge. It was taken off just before it went public, which is one of the biggest acquisitions. So yeah, it was a huge deal at the time. Week in before, yeah.

SPEAKER_03 26:38

Exactly. And so a lot of people were really um, yeah, there are people who want to be a part of a like a standalone public company and they chose to vote with their feet and and left App Dynamics, and other people want to stick with the company.

SPEAKER_00 26:49

And so there was this transition that was always a moment in time, yeah. With always uh the acquisitions, yeah, particularly when an acquisition is different, i.e. Um Cisco is a giant network, it had compute, it had cybersecurity and so forth, but application performance monitoring was like a totally different yeah, it was like a little pimple on the side. Yeah, yeah. Um

Android Culture Desserts And Team Bonding

SPEAKER_00 27:10

it was actually, you know, it's funny, it was when SaaS and software was like so cool, hardware was sort of frowned upon back now. So funny that if you're gonna be able to do it.

SPEAKER_03 27:17

How much? How much the times have changed now?

SPEAKER_00 27:19

Oh my god, so funny that yeah, hardware's now in vogue and SaaS is the SaaS pocalypse. But at the time, Cisco was trying to get to all their SaaS revenue and actually recurring revenue, and um, there's a huge focus for the company. So this was a big foray into it. So, yes, so you they put you in the seat to run it basically, or to build out all the product and sc and take it, I guess.

SPEAKER_03 27:38

Um try and turn it into like a tip of the spear for for Cisco's like SaaS movement and software movement.

SPEAKER_00 27:46

Well, it would have been the largest SAS revenue they had, probably at the time. Meraki, if you consider that a SASE, sort of but um pure SAS, it was the the largest. It was the largest one. Yeah. When was that acquisition? I can't remember.

SPEAKER_03 28:00

It was uh 2017.

SPEAKER_00 28:01

Oh, okay. Because Thousand Eyes was 2020. Yeah. So then yeah, so it was already three years in by that stage. Um, yeah, okay. Lots of interesting folks. Actually, a lot of successful folks from App D went on to do all sorts of interesting things.

SPEAKER_03 28:14

Oh, yeah.

SPEAKER_00 28:15

Um, amazing leaders. CRO went off. Um, I think he ended up at Z Scalar, then now Wiz, and I guess now Google. I guess that closes. Yeah.

SPEAKER_03 28:27

Oh, it's not close. That's still good. No idea.

SPEAKER_00 28:29

I think it's the world's biggest acquisition. Was that 20 or 30 billion? I can't even remember.

SPEAKER_03 28:32

Something crazy.

SPEAKER_00 28:33

So insane, yeah. Yeah. But that doesn't even seem insane anymore. That was so crazy.

SPEAKER_03 28:37

I know that was crazy, and now you're like, it's so small.

SPEAKER_00 28:39

So nice pennies. It's silly. Um, so Webflow, CEO of Webflow, Webflow is a really cool company. So explain what Webflow is for folks on the call, and then um, yeah, and then what's happening in that space? Four years, it's lots of change, AI, blah, blah, what's going on there? Yes.

SPEAKER_03 28:56

Yeah. So Webflow, it's funny. Our original mission was to empower everyone to build for the web. And it was a product that took the complexity of having to know how to code to build a website and turned it into or rather democratized how you build websites by creating a visual experience that people could actually vibe coding for websites or something like before vibe coding existed.

SPEAKER_00 29:16

Yes.

SPEAKER_03 29:17

It's like vibe coding by like, you know, visual coding. It's like visual coding for websites before vibe coding happened. And it allowed people to assemble these amazing, powerful, like performant experiences that translated to clean semantic HTML, CSS, and JavaScript. And so it sped up how quickly you could build a website. It changed who could build a website, and it changed who could then actually operate that website going forward. And so it was a huge transformation.

SPEAKER_00 29:43

That's why we use it because the amount of times one of spins, the amount of random things. We run 50, like a real simple one. We run 50 uh events in cities called the World Tour. And everyone wants to build a single page for each one with a little landing thing with this and all the detail, and they want to optimize it for blah, blah, blah, blah, and they want all those things. You either need to talk to a web developer to go and create, or you literally just go on this and create it yourself. And then it's like, well, it needs to be optimized for mobile, it needs to be optimized. It just does all that, just takes all the plan away.

SPEAKER_03 30:07

Makes it so easy.

SPEAKER_00 30:08

Drop some images in it, just makes it work. Um, and the marketing team's got no skill set. Well, the the marketing folks um don't have coding ability, obviously. So this sort of marketing skills, they know what good looks like. Yeah, they don't know how to make that, and you don't want to translate that to someone who's literally coding the machine behind it. So yeah, it's perfect.

SPEAKER_03 30:29

Yeah. And it what it does is it's a it leverages your marketers for the skills that they have because it's the website's just one piece of your strategy. Then you're like, oh no, your marketer now needs to run a performance marketing campaign against it. They have to drive traffic, they should optimize for signups, they should make sure the language fits, they should think about you know, SEO and AEO. And so what we've built, so beyond the website where we started, which is just building, yeah, we're saying, hey, once you hit publish, there's so much more. And so we've really uh shifted to own the entire life cycle of that web experience and going from just building sites to actually managing and optimizing them. So running analytics. We have an analytics platform, we have an optimization platform that allows you to do everything.

SPEAKER_00 31:08

Site load fast, stuff like that. Yeah, that's one of the big ones that then you rate higher in the searches.

SPEAKER_03 31:14

We help you with SEO and AEO. We support you with experimentation. So if you want to do A-B testing, we support you with personalization. Let's say you want to do an account-based marketing campaign. You're gonna target 50 different customers.

SPEAKER_00 31:27

Imagine web megaport to get that. Exactly.

SPEAKER_03 31:30

And so we we have a platform that allows you to run personalized campaigns across your key customer targets. We have localization. So you're a global company, you you need a version of your website in French, German, Spanish, English, whatnot.

SPEAKER_00 31:42

Which is really hard.

SPEAKER_03 31:43

So hard.

SPEAKER_00 31:43

So hard, yes. But even with all the AI stuff, it's still hard.

SPEAKER_03 31:46

Yeah.

SPEAKER_00 31:47

Um, it's gotten a lot better, but yeah, it's a really tricky one. We're in, oh, I can't remember how many languages, but I want to say like 10 different languages: French, Spanish, Italian, English, German, Jap, Japanese, um, Portuguese. Yeah, like it, I'm just trying to think what we uh what else we would have lots more. Anyway, every time we do anything on the website, we need to deal with every single one of those things, which is tricky. Yeah, exactly.

SPEAKER_03 32:12

Yeah, it's not just about translating the words, you want to translate the content as well as like your imagery. Yeah, yeah. So you want, you know, if if your American-based site has a picture of, I don't know, like the White House, yeah, your your French one should probably have like the Eiffel Tower or like something right like Arc de Turiom.

SPEAKER_00 32:29

What the hell is that? Oh my god, right.

SPEAKER_03 32:30

My French is terrible. Yes, but like you you want to have you want to have like you know, media that is also contextual to you know the the language that you're using.

SPEAKER_00 32:41

Totally. Yes. Um, yeah, if it's all just an American thing in the j Japanese, uh, it's gonna make zero sense. So yeah, well, I think we must use that for that as well. But one thing that is really hard, so uh the A B testing I love by the way, our team uses A-B testing, and it's weird. They go, Hey, do you know we changed the colour of the um like the the action button or something?

SPEAKER_03 33:04

Oh yeah, yeah. The called action.

SPEAKER_00 33:05

Yeah, called action button from green to pink, and it performs 35% better. Or some weird, it's like so weird what you can A B test and optimize. It's so bizarre what they can. Yeah, but it's weird what actually makes a percentage or change. I'm like, oh, they really like pink. Yeah, um, or whatever. Maybe it's pink because of the background, I don't know, but yeah, they're doing it.

SPEAKER_03 33:27

Maybe it's the contract, maybe it's the color preference, maybe it just captures some of the attention.

SPEAKER_00 33:31

Our CMO is so technical on that side. He has this, it's funny, like it doesn't really have um an opinion on lots of things, in that his opinion is we'll test it. Yeah, like I don't know the answer, I'll test it. So I was like, Oh, I think this video on the homepage is better if it does this. Maybe he's like, Yeah, let's test it. It's like I think if you put a human face, people prefer it to whatever. It's like, maybe it doesn't agree or disagree, he's like, I'll come back to you and just comes back with data, it's tested. But yeah, the A B testings, I love, love that it's amazing. And the thing is, many enough people use it.

SPEAKER_03 34:02

I mean, maybe it is very actually a lot of people don't. And I'll tell you why. It's not that they don't want to do the testing, it's actually the hardest thing, and where we're innovating next is the hardest thing is thinking about what do I want to test.

SPEAKER_00 34:12

Well, yeah, and then I want to pick or agree. I had to think to do that.

SPEAKER_03 34:16

Yeah, think about it. Why would I even think that's what I think? So what we actually where I think the next world, the sort of the next phase, right, is AI has really transformed. AI has transformed how people create, but where I think AI is gonna really move the market is how AI augments your team and partners with you to operate. And so, you know, where people are like, I don't have the time to dream up tests, I don't have the time to go build it and implement it, I don't have time to see how it performed. We're building agents that will work with you to actually create those directions.

SPEAKER_00 34:44

Figure that out for you. To run it to run a million things, to tell you to come back to make suggestions to change it for you. Exactly. It's worth mentioning. I learned what A B test, but let's explain it. It means that when I access your website today, I will get the I potentially website to a

Tapjoy And The Birth Of Offer Walls

SPEAKER_00 35:00

person sitting next to me who accesses the exact same website. And it could be called action, a color, it could be a complete different landing page, it could be anything. Yes.

SPEAKER_03 35:07

Yeah, it could be anything, it could be a multivariate test.

SPEAKER_00 35:10

And then you see how long I stay on the page, where I click, where I go to, how you know, how do I just close it down, which sites do I go to, blah blah blah. It's like exactly it's so I love it because it's so data driven. It's fascinating. Absolutely.

SPEAKER_03 35:20

And based off of sort of the the number of tests, because it's not just uh, you know, A B people think there's only two tests you're running, right? There's like one variant and a second variant. But we actually can do multi, uh, multiple variables. So it could be like a hero image, it could be content, it could be color, and we could test them across like hundreds of thousands of variants. So you can actually run so many tests at the same time, and we'll actually look at that data and look at what's performing and start to organize it for you and help you win.

SPEAKER_00 35:52

As it just constantly doing those things.

SPEAKER_03 35:54

Exactly.

SPEAKER_00 35:54

Yeah.

SPEAKER_03 35:55

And so we actually have it's continuous optimization that is testing all these different variables for you. So you don't have to dream up the exact test and say like this one versus this one, but rather we say, let's test it all, and we'll come back and we'll tell you what is the optimal solution to drive the action you want. Yeah, and it can it actually continues to be able to do that.

SPEAKER_00 36:12

That's the point, it might be that now, but for whatever reason, yeah, weirdly, like why are we even changing like that? But yeah, it's super yeah. Uh so I that's super cool. So, you guys, that'll be an interesting AI space. Oh, I was gonna ask you about it. Used to be SEO, search engine optimization. And in fact, AEO. Yeah, AEO. So when I first got to Megaport, we actually had to rebuild the entire websites, all these pieces, get all the SEO work. It was a disaster. It was something that was built in 2015 and everything was broken, it was a mess. And I arrived there in 2023. So think of the distance in that time. So we went and rebuilt, and it was all about speed and fast and beautiful and all the links and all the um web optimization, all those pieces. And even interestingly, to this day, the largest amount of leads come from organic um web uh from our website, organic. Organic search. Exactly, and it's linked back to that, which is I think it's cloud connecting between clouds, such a weird thing that we become the thing that we show up on. Now, we obviously are a cloud connector, but um, it was interesting because Amazon spent so much time never talking about another cloud, and so we ended up as the the top result for everyone because no one else would ever acknowledge there was another cloud, and we were the only ones saying multi-cloud, it's just this bizarre thing. Anyway, it still feeds us a lot of stuff. Um, when uh AI came along, I was like, holy, we're gonna lose our lose your positioning positioning or whatever, and now AI is making these decisions. And so explain to me how that's by the way, that's blowing up everyone's view on how these SEO works, our human tire industries of advertising was built on this and marketing and what have you. It's literally almost thrown out, feels like it's thrown out the wind. No, you could probably convince me one way or another of whether that's true, but it feels like that. Um, and then we go, all right, well, um, I'm so worried now. How do I, I don't know, become the answer in ChatGPT or Claude or whatever it is that for me it'd be like um CPU, GPU, storage, compute, network, whatever. So if anyone wants that, how do I become the answer to that equation? As a CEO, like it's really, really difficult. And the other thing that's so hard, I don't know, it used to be easy to do Google advertising.

SPEAKER_03 38:22

Yeah, you paid.

SPEAKER_00 38:23

Oh my god, you paid. I like I had a little real estate business um that my parents ran, and I just did all this Google advertising, and I used to get it for like three dollars a click or whatever it was, and now the click-through cost is just seems like so. I thought this was just us. I was losing my mind, thinking this is just unbelievable. I find other people are paying, I'm like, oh, we're not that bad, even though it was felt it feels so painful. Um, and the other thing that's so annoying is we used to be able to track everything so beautifully with all this GDPR or whatever it is, all this rules really hard. All the cookie, blah, blah. We don't you can't actually tell if it's the same person, how long they stayed. So it's actually we've removed so much visibility that we used to have. So I actually think that the CMO's job in today's today is infinitely harder than it was literally five years ago.

SPEAKER_01 39:14

Absolutely.

SPEAKER_00 39:15

And actually, it's so I don't even know. It's so hard to make um make decisions on it. You used to be able to invest money in something and do the data to say exactly what the return was. Now it's it's like it's gone backwards to you can't even tell in many cases.

SPEAKER_03 39:32

So part of measure success.

SPEAKER_00 39:33

It's a measure success. And then then you have this debate. Well, maybe we just stop investing in it. And you're like, yeah, but then cut the budget and then what happened?

SPEAKER_03 39:40

Exactly. Was it because of that or was it something that happened? It's really hard.

SPEAKER_00 39:45

We're talking for bigger, bigger companies. This could be that a millions and millions and millions of dollars a month of decision that actually is we're a lot of people are blinded by. So anyway, I I feel that it's a really difficult place. I assume that it's so time. We see it all the time.

SPEAKER_03 40:00

We see it with all of our I mean, we sell the CMOs, right?

SPEAKER_00 40:02

Exactly, yeah.

SPEAKER_03 40:03

And they're feeling it. So one, what you're seeing on organic search.

SPEAKER_00 40:06

So it's not just us.

SPEAKER_03 40:07

No, not just you, it's everybody, right? Like everyone with organic search is seeing anywhere from 15 to 50 percent declines in their or organic search traffic.

SPEAKER_01 40:15

Yes. So organic search is declining.

SPEAKER_03 40:17

It's declining really fast. And some of them are seeing some sort of makeup and traffic that's coming from answer engines, ChatGPT or Anthropic and the R. You can't see that.

SPEAKER_00 40:26

So you can see via You can see that it's someone that's clicked in ChatGPT and landed on your website.

SPEAKER_03 40:31

Yeah. So you can you can see sort of the referring agent, right? Um so you can tell like what traffic is coming from, you know, organic search, what's coming from an answer engine.

SPEAKER_01 40:41

Exactly.

SPEAKER_03 40:42

And you can actually see the breakdown by the types of bots. Is it's uh a bot that's like indexing you, is it a bot that's coming directly from a search, or is it a bot that's just like refreshing content? So you can actually see the breakdown of that as well. Uh we offer visibility into the code.

SPEAKER_00 40:54

And indexing it, that's more like a Google thing, or is that actually like very similar to a Google thing.

SPEAKER_03 40:59

So it's a it would be like a Chat GBT bot. So you might actually see a ChatGBT indexing bot that is just indexing.

SPEAKER_00 41:08

Exactly.

SPEAKER_03 41:08

Very similar. So there's there's multiple types of bots per answer engine. Uh some are like a sort of like a search orientation, some are just indexing, some are just refreshing your content.

SPEAKER_00 41:17

And Google's growing.

SPEAKER_03 41:19

Yeah, it's growing. I mean, remember, well, because Gemini, right? AI overviews is their answer engine. So you'll see a mix of Google search and Gemini AI overviews traffic, which is doing incredibly well, right? I think they just announced um at I.O. they're doing like you know, two and a half billion like users monthly.

SPEAKER_00 41:37

Yeah, it's insane.

SPEAKER_03 41:38

Yeah, it's wild.

SPEAKER_00 41:40

It's so funny because like a year ago, everyone was like, Oh, I actually checked, or maybe it was a year and a half ago, like Google's dead and ChatGPT is gonna take it all. And actually, what ended up happening to Google just accelerated. Yeah. It just took off. It's wild, yeah. Yeah. Yeah. Um, so how do you s what what do you think about in terms of an AI, how do you optimize for AI?

SPEAKER_03 41:55

Like what the Yeah, so there's this new practice called AEO that everyone's talking about. So it's it's answer answer engine optimization. So versus search engine optimization, which is SEO.

SPEAKER_01 42:05

Okay, answer engine, not AI optimization.

SPEAKER_03 42:07

Exactly. It's answer engine. And it's very similar to SEO.

SPEAKER_01 42:11

Answering engine, the engine that can answer everything.

SPEAKER_03 42:13

Engine that answers everything.

SPEAKER_01 42:14

What is the meaning of life, answer engine? Answer me.

SPEAKER_03 42:16

Tell me, answer it. Um but yeah, so with AEO, you do very similar things to SEO in that you're trying to optimize for the answer engine to understand you, to essentially rank you, to index you.

SPEAKER_00 42:27

Is it similar?

SPEAKER_03 42:28

Or it's like how much changes are you doing to it's similar conceptually, and that you're trying to achieve the same results of like you're trying to be recognized by these answer engines and ultimately be you know, referred or recommended.

SPEAKER_00 42:38

Um these answer engines also looking at all the links too. You remember the whole thing. Exactly.

SPEAKER_03 42:42

They're looking at your backlinks, but like they look at a lot more. So like the top things that we talk about is content authority. So it'll look at your website. And let's say you're inconsistent about how you articulate what your company does, what you know, what products you have, and the content that it also sees beyond your website doesn't match up. Um, it gives you a low content authority score at that point. If you're telling fundamentally different stories between like your social and your third-party published content and your website, so it's really important.

SPEAKER_00 43:08

It's probably an issue for companies as they take on more products. We're taking on more and more products. Yeah.

SPEAKER_03 43:13

And these answer engines don't understand what a company does because they're like, well, what products is it? And what is your purpose? What's your vision? What's your mission? So content authority is really important and clearly articulating what you do.

SPEAKER_00 43:23

Is that still something that can be delivered via blogs or like yeah? So that is sort of that search and optionalizing.

SPEAKER_03 43:30

It's very similar in that sense, in that, like, you know, if you're adding like fresh content to your site, it's written away that it's like getting high authority scores, it's matching content that you have elsewhere, that ranks really well. Um, you know, answer engines are ultimately their their agents, right? They're bots that are looking at your site. And so the other thing that we tell you to optimize for is making sure that your site is legible by these agents. And so site schema is really important. Structure, how you think about even like your ta metadata tags and things like that are really important.

SPEAKER_00 43:59

Plain, you got that. Which I guess is a similar thing to SEO. So it's a similar fundamental. Um, we have found interestingly, we have an incredible docs page for all of our products.

Cloud Phone Startup Then NFL Digital

SPEAKER_03 44:10

Yeah.

SPEAKER_00 44:10

And it breaks down how to deploy every one of our products.

SPEAKER_03 44:12

That's excellent.

SPEAKER_00 44:13

It's so important. That seems to be what gets picked up. People say how to connect between this side to that, how to deliver bare metal, how to go to a cloud from this. And it actually somehow has obviously figured out that our docs page, which is, and by the way, like our docs people, I love our docs page. It's like megaport.docs, I think it is, and it's just like so simple. But every product is in there, how to deploy with images, with steps and all those pieces.

SPEAKER_03 44:38

Every single product structured content, really clear information, organized in a way that's legible, right? Like, you know, agents love markdown, yeah, or just structured content. And when you organize it that way, it ranks really well. It's funny. That's like cues do really, really well. Yes, right. So it's just having really rich content that clearly articulates what you do helps your AEO score a lot.

SPEAKER_00 44:59

That's funny because yeah, my most of the fluffy marketing stuff anyway, lands poorly. Yeah, yeah, yeah.

SPEAKER_03 45:04

You know, imagine someone just reading it and not knowing like what websites they're on. Could they figure out where you are, who you are?

SPEAKER_00 45:09

No, no.

SPEAKER_03 45:10

How's Nathan gonna figure it out?

SPEAKER_00 45:11

No, 100%. Or if it's a picture. Exactly. Maybe, yeah. Uh okay, cool. So you're you've got a ton of developers inside your business. Yeah. I assume lots lots of developers. It's more of a you're a very big software platform.

SPEAKER_03 45:23

Or a software platform, yeah.

SPEAKER_00 45:25

Um what's happening with coding? So are they just all going crazy? Are you getting all these? Are you turning up um with like uh these giant bills coming out that you didn't expect, or everyone wanting tokens and I want the token maxing? Token maxing. I'm gonna work, yeah. Vampires turning into like I got 11 agents building code of the same. Yeah. I'm never stopping. Yeah, good.

SPEAKER_03 45:47

We do have, I mean, I think it's uh, I think every company sort of is going through a very similar life cycle of, you know, give everyone the tools, turn on the tools, like token max, like build things. And um, you know, I think it's actually more of a like divergent, convergent, divergent, convergent sort of experience where um you want people to like be really familiar with the technology. You want them to build, you want them to experience these aha moments of wow, with AI, I can go so much faster. I can do this. But at the same time, AI is not perfect. It needs context. And so, you know, the risk is you sort of just start with this divergent like experience. And what happens very quickly is suddenly developers are being DDoSed by, you know, it's so easy to create code. Who's reviewing the code? Yeah, that's right. And so you need to figure out like how do you actually transform how you work, not just how you take one piece of it and blow it up. Yeah. Um, and so we're thinking about how do we ensure that AI is creating high quality code that fits with our business that is like actually has a full life cycle of like not only creating the code, reviewing the code, testing the code, like actually running it out in production, having our whole CI CD cycle like work with it.

SPEAKER_02 46:53

Yeah.

SPEAKER_03 46:53

And so the things that we're investing in are not only enabling our teammates and and like putting the tools in their hands and letting them play, but transforming our processes to work well with it and building what you know, a lot of people are talking about coding harnesses, but like building a harness around our software. So um, you know, like right now, if you use, I don't know, like Opus, you're gonna just be like, write me an application that does this. And it's a generic application that follows some high-level best practices from like online, but like it may pick a language, it may pick a structure, like there's not really much there. But when you want to actually build on an existing product, there's already best practices. There's like, how do we think about APIs? How do we actually think about the different methods that we create? How do we actually structure it? What is the language?

SPEAKER_00 47:38

Yeah, yeah. And so what you need is a harness of the machine and the harness, yeah.

SPEAKER_03 47:41

You build a harness. A harness is actually like your context layer. This is how we operate principles. Here's our documentation, here's what things do, yeah, here's how these systems connect. But it's a it's essentially a context layer that when you're actually creating code, that code is creating within that harness, it's understanding the context of your business so that whatever you're doing. More or less.

SPEAKER_00 48:03

You you sort of say, here's our code.

SPEAKER_03 48:05

Sort of like the harness ends up being sort of it's a mix of like markdown skills and capabilities and best practices.

SPEAKER_00 48:14

This seems like like a massive scalar of the company if you get the thing right. Absolutely. So you need someone super clever to you need a lot of someone. You need a harness.

SPEAKER_03 48:23

Exactly. You need people who are investing in your harness, who are building it and adding the context of your code base to it so that whatever gets generated is leveraging the harness and creating stuff that fits your product, that works well. And then you still need to like do all the other things in terms of like making sure things are getting QA'd properly properly. You're writing the right test coverage. That's um, but that's a big unlock. That's what really is going to transform.

SPEAKER_00 48:46

So you'll be you'll be nervous on this one. We've got, I think, nine developers inside the business, and each one of them is like going wild on flawed in some way. Some of them are extreme versions of that, some of them are uh, you know, probably slower to take that up. But doesn't they're still producing way faster than they were. The problem is all the QA lands like tons of so we've had to, and I don't know if everyone gets this right, but we've had to pivot the folks to actually review each other's code as opposed to having to have QAs, because basically you had a whole bunch of QAs that were just getting fed like a 10,000 fire just getting crushed. And they're like, actually, this doesn't work. Actually, it doesn't scale because they can't actually get anything through anyways. And then they go, actually, we need to flip this on its head. We get everyone to each review each other's code or something. And that look, hopefully, um claw gets a little bit better in terms of the QA piece or something, or someone solves that equation on the next release. But that seems to be for me, it feels like that's the bottleneck or or the well.

SPEAKER_03 49:41

So for us, we're like using, you know, using AI to write test suites. Yeah. And we're using AI to actually like run code against these test suites and like actually check against like different synthetic experiences, so that instead of having someone QA it, like AI should QA it. Yeah. If AI is running it, it should QA it. And so it's about creating this, like these constraints and these guardrails. Yeah. Like because ultimately, if you have a closed loop here where you can actually write, create, test, like iterate, like you can actually get really great output. Um and so we are looking for areas of our code base where you know it's less complex, less integration, where we've actually have agents. We have AI that's basically running on the side. So as a ticket comes in that's like, hey, build this feature, it can actually pick off the ticket, write the spec, code it up, uh, test it, push it to production. And these are like sort of the low risk areas, but like, you know, we're creating these like closed loop AI experiences where AI can actually run the whole thing end-to-end. Um, and what what

From AppDynamics Customer To Operator

SPEAKER_03 50:40

allows it to happen is that it's a good idea.

SPEAKER_00 50:43

I want this, there it is.

SPEAKER_03 50:46

Like build it, and here it is.

SPEAKER_00 50:47

Great news.

SPEAKER_03 50:48

Yeah. So like that is the feature we want to build towards because anywhere where like, you know, everything is like a, you know, whether it's your software development lifecycle or your end-to-end like build experience, you have all these different steps along the way. And if only one step is AI powered and everything's not.

SPEAKER_00 51:03

Yeah, that's everything else is.

SPEAKER_03 51:04

You're just gonna like, you're gonna find a different bottleneck.

SPEAKER_00 51:06

Well, you just hit another, yeah, the roadblock or the bottleneck is there, yeah. Yeah, yeah. Yeah.

SPEAKER_03 51:10

So you have to sort of you have to think about how you leverage AI to unlock every part of that life cycle.

SPEAKER_00 51:15

Have you found it easy to get the engineering team to change? Uh, you know, I we find different varying views. Some folks actually love coding.

unknown 51:25

Yeah.

SPEAKER_00 51:25

And one of my one of our head guys actually was like, I actually, it's taking away the one thing that I love doing.

SPEAKER_02 51:31

Yeah.

SPEAKER_00 51:31

I'm like, I I said, you'll find a way to actually be able to uh instruct it to code for you lots faster, but I actually love the art of coding, you know, and I'm like a little bit sad in that instance, in that, but it was like there's extreme versions of where people are at. Um and getting humans hate change.

SPEAKER_03 51:49

Yeah, and so change is tough.

SPEAKER_00 51:51

And engineers hate change a lot because it's they love a black and white component. So how yeah, how do you how the hell do you manage that bit? Like, um, I actually think this is a big issue for everyone right now. So I'm just like, I don't know, you don't have to do the answer and probably be one, but yeah, what do you what's the philosophy that you're at least taking with it?

SPEAKER_03 52:06

Honestly, it's it's standard change management. Any any leadership role, you have to do some sort of change management, whether it's uh bringing people along the ride for AI or changing things just with the business, change management is hard and it's tough. And you know, I think it's for engineers who are feeling this a lot on the air.

SPEAKER_00 52:25

Big change for engineers. Like for 20 years it's been the same. Exactly. And all of a sudden it's not now, it's different. And now these people are like pretty strict in their thinking on things, you know. Yeah, yeah.

SPEAKER_03 52:36

And I think the key is is um when you see these technical transformations, these like step function changes that we see, you know, like probably the last time you saw this was like frankly the internet.

SPEAKER_01 52:46

I don't know.

SPEAKER_03 52:47

Yeah, the internet or even like before that, the industrial revolution. Right? Like you know, see these like massive transformations and but it just happens there are people who go with the channel.

SPEAKER_00 52:58

That's the that's the bit that the humans it's scary because it's so fast. Yeah, like at least industrial revolution, it's like trains. Yeah. It takes you ten years to build a train track, whatever it is. Now that's fast in that period, but this is happening in a month.

SPEAKER_03 53:09

Yeah.

SPEAKER_00 53:10

Two weeks.

SPEAKER_03 53:10

These models transform every two weeks.

SPEAKER_00 53:12

Every two weeks, too.

SPEAKER_03 53:13

And so it's a it's a scary change, and it's one where the decision of are you gonna go with it or are you gonna be left behind? Are you gonna be a part of like every person has to make that call for themselves.

SPEAKER_00 53:22

And then there's what about one of the things that all the SaaS company, you're not public, so you don't have to deal with this pain, thankfully. But it's like there's been this SaaS pocalypse. We've seen everyone thinks that everyone's software is just gonna be replaced by AI. It's like the market woke up one day and everyone's just literally gonna, oh no, code can replace everything you do, which I think is a total over exaggeration and they're gonna rip pivot pretty quickly back. But maybe not quickly, maybe it takes some whatever. But that that has occurred. How does that look for your business in terms of yeah, are you just gonna be disappeared by is AI gonna take over your business?

SPEAKER_03 53:54

Like, yeah. I mean, I think it's a I I think everyone's afraid of that. The same way that they're afraid of like, you know, every SaaS company is sort of down in the market right now.

SPEAKER_00 54:03

They're all down in the market's like crazy things.

SPEAKER_03 54:05

I don't believe it. No, but I think there is just the there's a narrative about it. Yeah. And I think the narrative is wrong because markets are so extreme, they don't understand.

SPEAKER_00 54:13

It's so ridiculous.

SPEAKER_03 54:13

Well, I mean, like the the idea that a company is just the code that it produces, yeah, is the if it was that simple, everything would be replaced already. Exactly. Yeah, you have to think about like what's actually behind that. Like the easiest argument is people are like, oh yeah, all the software is gonna be replaced. Like work day is dead, right? Like these HRI, and every every employee hates whatever HRIS system they're using. Yeah, sure.

unknown 54:34

Yeah.

SPEAKER_03 54:34

And I'm like, let me tell you how much I'm not gonna spend time building my own HRI.

SPEAKER_00 54:40

Yeah, I see.

SPEAKER_03 54:41

Because it's like even if I could build it, let's say I could build it in a couple of weeks, I then have to maintain it. I have to think about client. Yeah, I have to look up all the compliance. In Germany, you're like, oh, when things break, it's like even when it comes down to software, right? When you build your own software, there are pieces of it where you're like, it doesn't make sense for us to build in-house. We should go to another firm and use their you know authentication system, their security system, right?

SPEAKER_00 55:08

Like there's so many examples of better software today, before AI, but we still don't move.

SPEAKER_03 55:13

Exactly.

SPEAKER_00 55:14

Yeah, it might be cheaper, easier, better, blah, blah, blah, blah. But we still will that now know.

SPEAKER_03 55:17

Exactly. It's because software is not just the code behind it, it's the thinking, it's expertise, it's the context, it's like all the like and it's living. Like the idea that software is just like a thing you build and you leave it is silly. Like as soon as you build it, it's in a state of like decline, right? It has to be maintained and managed. And so the reason why I don't think software companies are going away is because it's not the code that's actually the value.

Webflow Beyond Building Sites

SPEAKER_03 55:40

It's the team, it's the context, it's a space. It's like finding a narrow niche in which you can build code, but maintain that code to deliver value because no one else wants to build it.

SPEAKER_00 55:49

And also behind these software platforms is a huge amount of infrastructure. Absolutely.

SPEAKER_03 55:56

They have to figure out how to operate at scale because like the only reason why these companies make money is they've found economies of scale somewhere.

SPEAKER_00 56:01

Yeah, yeah. Yeah.

SPEAKER_03 56:02

And I'm you don't want to have to do that for every single thing you're gonna run.

SPEAKER_00 56:05

Market's overreactive on that side. Um, it's probably it's good nice to be private when that happens because you just you can just avoid Yeah, you just avoid all the valuation conversations.

SPEAKER_03 56:13

Yeah, exactly.

SPEAKER_00 56:14

Um, yes. But then I guess what does uh what do you guys want to do in the future? Is it everyone's like every private CEO wants to be public, every public CEO wants to be private? This what's the um yeah, what what is what do you want to do for the broader business? What's the plans there?

SPEAKER_03 56:29

Yeah, you know, for me, like private versus public doesn't really make a difference. Like it's just like a it's a moment in time for a company to go public or you know, just say private. For me, it's a it's just being mission driven, right? Like the only reason why I'm the CEO is because all I care about is our mission and our vision and what we want to do. Um, I'm not if you want me to be a CEO to take a company public, you got the wrong gal. Like I'm just here to take the vision and figure out how do I keep us moving in that direction and try to like pursue this, like this infinite dream as effectively as possible. Um, and so for like the future of Webflow to me is where we started as this technology that helped people visually build websites. I now want to visually manage them. I want to like create the new endpoints, like whether you want to access us via an MCP, a CLI, or directly with an API. I want you to be able to build, manage, and optimize those experiences and have the agentic fleet that can augment your team and work with you on how you operate. And it's a it's a big bold vision of where we think marketing is going, but that's what I want to build. And, you know, like ultimately, if we're successful, what we'll have done is we'll have supported, you know, tens of thousands or hundreds of thousands of customers and running their organizations, drive growth for their businesses, and like built something that it's like truly delivering value on a regular basis.

SPEAKER_00 57:44

So I mean, that's awesome. Like get behind that mission. Um, you mentioned something throughout that that you would should explain what all these MCPs and things like what is like what does that even mean? That's sort of uh we found found this out the other day. I learned this, and I was like, wow, this is really cool. But this is it's like the interface of an how an AI would it's like an AKI for the AI, is that exactly more or less.

SPEAKER_03 58:06

It's like um so if you think so. One of the things that's top of mind for me is this concept of AX. It's you know, where UX is user experience, AX is agent experience. And if you think about um, you know, we've always thought about software as something that we have to build for an end user, someone like you or I, right? So it's a typically a visual experience. But if you imagine a world where it's not just like agents building code, but agents running workflows, the the touch point is actually not you or me logging in and doing something, it's actually even an agent logging in and doing something. And so you need to build your product such that an agent can access it, interact with it, and actually run those workflows effectively. And so from an AX perspective, right now, a lot of people are using MCPs, so model context protocols. These these MCPs are a way where you can essentially uh allow an agent to come and interact with your product and run it via this, you know, right as like a prompting experience.

SPEAKER_00 59:01

But like push it things to it, actually make decisions, execute. In our case, they can build up an entire network and buy it or execute. Exactly.

SPEAKER_03 59:08

Yeah, and it's just it's a it's a protocol that allows agents to essentially interact with your product.

SPEAKER_00 59:13

So you can help build, like I was just thinking as you said that we've remember with this whole vision of product-led growth, this whole it was the like the holy grail. Yeah, that was a holy grail. Everybody wanted it. No salespeople, everyone just finds it engross. What we're like, we've seen like almost Claude-led growth. Yeah. Um, just where Claude is deciding that this is a great software platform to leverage or this is more efficient VM or whatever it may be. Is that where you may it can also be a part? Like, is there a way in the world where we used to have to market for product-led growth using a website, maybe there's ratings and reviews or whatever it may be, and then a then a human might decide without talking to someone and buy something? Are you doing that with your uh webs, websites and effect to actually try to convince or actually pitch the product better to an AI platform? And if you can pitch your product better via that platform, you would no different to the product-led growth, you win the the deal here, but you are gonna win the AI.

SPEAKER_03 1:00:11

Absolutely. That's the hope, right? So one from a mix of how do we show up in the answer engines? Yes, right, with AEO. But two, the access points, the agent UX. And so if you actually have Yeah, can it actually interact with you? Can it run on your platform? Can it take action? And so are you servicing an MCP or are you servicing like can someone access you via like a CLI, right? Command line interface, or just directly if a developer wants to hit your API, I think the new experience and you know, credit to Salesforce, they're the first ones to really put it out there, but like the concept of headless software so that the first head touch point, no, it's not necessarily like a visual experience, right? Like, yes, you will have one, but the if you imagine the primary way, the primary way someone's gonna interact with your software is via like an agent. Yeah, you're gonna want to create this headless experience that is essentially API oriented that allows like an FCP to come in or you could use a CLI to come in.

SPEAKER_00 1:01:05

But somehow is pitching your product over someone else and it makes a decision to use that versus other products within the right.

SPEAKER_03 1:01:12

In a way, yeah, in doing that. So you're creating the access points, right, by making it headless. And then things like AEO and like integrating with these, you know, with you know, Claude or with like Codex or something is what makes it easy for someone to build in Claude and be like, oh, actually, like if I'm building a website, yeah, build it in Webflow, use this MCP server and build it, show me the thing, like integrating with the point that people are creating. Uh, so whether they're creating in Cloud Code or they're creating in Codex or they're creating in Gemini, like having those integrations with these frontier labs and in terms of like these workers or spaces. You can create an MCP in which um you'll want to create like essentially integrations with each of these um each of these sort of model providers. Yeah, but like it's still one MCP and then it's just sort of like making it work for each.

SPEAKER_00 1:02:00

Yeah.

SPEAKER_03 1:02:01

Yeah.

SPEAKER_00 1:02:02

I only learned about MCP the other day, which is when you mentioned that's really interesting. Um cool, okay. So you got a lot to do.

SPEAKER_03 1:02:08

Lots to do. And the work is moving really fast.

SPEAKER_00 1:02:10

Oh, that's well, yeah. There's so much chat so much change. I think um the one thing that I caught up with a number of our customers in the Bay Area, they're all just blowing their mind at how fast things are changing. And the answer has been how quickly can we um adapt to it? So that I like it's funny, I think the winners are gonna be the ones that move the fastest at the moment. That's it, because it's pivoting so quickly. Um and that if you can pivot your entire company, and this is why I think it's all of a sudden the CEO becomes a key part of this because it's only one person that can 100% move an entire company in a moment, and that has to come from the top. Whereas before things would move and decisions eventually get up to you, but it's you could take time, but you need to make decisions almost on a daily basis and actually place bets um on these pieces and watch them fail and f figure it out and scale it. So I think um you're super well positioned for it, you're the right leader for it. So we I love seeing what you're doing. Um, yeah. Keep building, keep going crazy. Anything you wanted to share with anyone?

SPEAKER_03 1:03:06

No, that's it. You've shared it all.

SPEAKER_00 1:03:08

I've I've unpicked your whole life. I know. Way to just like dig in there. I know, but it's amazing. Um, yeah, so cool. So check out Webflow. Um, follow LT. How do you follow you?

SPEAKER_03 1:03:20

Uh LinkedIn or Twitter. Okay, or X.

SPEAKER_00 1:03:23

Tweets X, yeah, yeah, true. Tweeting the I still call it Twitter.

SPEAKER_03 1:03:26

I know.

SPEAKER_00 1:03:27

Do you know when you search on your iPhone you type Twitter, it still brings up X. Oh, yeah. It's so helpful for me. Yeah. All right, good to see you.

SPEAKER_03 1:03:34

All right, good to see you.

SPEAKER_00 1:03:37

Thanks for joining us on Uplink. See you next time.