GoDaddy CEO Aman Bhutani on the enduring power of the website

Today’s episode of Decoder is a little different: I’m talking with GoDaddy CEO Aman Bhutani, and we did this one live onstage last week at an event hosted by AlixPartners in Palo Alto, California, so you’ll hear the audience in the background.

GoDaddy is one of those companies that feels tied to an earlier era. In this case, the company had a hugely controversial reputation in the web 1.0 days, when it built a huge business selling domain names and ran a bunch of pretty sexist ads — but all of that is long in the past. The company still runs a huge domain name business, but Aman has been CEO since 2019, and he’s been building out what he calls adjacencies — basically, helping the very small businesses that run websites on GoDaddy grow their operations. You’ll hear him talk about a barbershop owner that runs a couple sites as one of his biggest customers — one he personally handled customer service for.

But the business of the web has really changed in the past few years: the walled garden social network era really took over in the past decade, and now huge changes to Google Search and the addition of generative AI have put a massive strain on the very foundations of the open web.

So I started out by asking Aman the question I’ve asked so many other guests on Decoder in the past year: What is the point of a website in 2024? Who needs one, and how do they benefit from paying GoDaddy to host it for them? 

You’ll hear Aman say that a website can convey legitimacy — that his customers feel pride in having their own domains, something they own and control for themselves, to build a direct relationship with their customers, instead of relying entirely on other platforms to do it for them.

You’ll also hear us talk about AI a lot because, at this point, it’s everywhere. GoDaddy launched a service earlier this year called Airo, which allows site owners to generate everything from logos to full site designs with just a few prompts.

That’s a big idea — and it’s one GoDaddy’s big competitors in this space are doing, too. I talked to the heads of both Squarespace and Wix last year, and they’re all chasing AI-powered web design as well. But I wanted to know if any of this was making money for Aman — and if he thought a flood of AI content would change the web for better or worse.

And then, because we were live, the audience got to chime in, too. They had some smart questions at the end that I think you’ll enjoy.  

Okay, Aman Bhutani, CEO of GoDaddy, live onstage. Here we go.

This transcript has been lightly edited for length and clarity.

Aman Bhutani, you’re the CEO of GoDaddy. Welcome to Decoder.

You have been the CEO of GoDaddy since 2019. I think a lot of things have happened since 2019.

Do you think you still have the same job you signed up for in 2019?

I have not just the same job, I have way more to do than I did in 2019. I think COVID changed things and the company changed, and now we have so many more opportunities — so much more to do, even more.

One theme that we’ve been thinking about a lot at The Verge is that the internet feels like it is on the precipice of major change. AI has changed. Our distribution networks and search might be changing in huge ways. The way we bring traffic to websites might change. The creator economy might have flipped the very idea of needing websites upside down entirely. How do you see GoDaddy’s business now? Are you still in the business of websites, or is it something bigger than that?

At GoDaddy, we’re known to the world as the largest domain registrar. And when I came to GoDaddy, it was still mostly a domain registrar and a hosting company. Of course, hosting changed in a very big way too with the cloud and other providers. With me coming in, the idea was that it’s great to be the largest domain supplier, and it’s great to be the first company people think of when they have an idea. Because when you have an idea, what do you want to do? You want to lock that name down. But there’s so much more we can do with that. So we should go into adjacencies. We should find the things that our brand has a right to sell to our customers. 

Our customer is the micro business. This is the very, very small business. If you get to companies with 10 employees, you’ve already traversed more than 90 percent of GoDaddy’s 21 million paying customers. So it’s mostly very, very small businesses. And what are the needs that these folks have, and do they want those from us? What we found is that our customers had identity needs — not identity as in their personal identity, but in how they show up to the world. They had needs around their presence, whether it was in store or online, and that’s where websites come in. And they had more and more commerce needs. 

So today, GoDaddy has grown a segment we call Applications and Commerce that is entirely around these new businesses that we’ve built. That’s taken us from, let’s say, chapter 1 of the company to writing chapter two of the company. And yes, the internet’s going to change, and I think AI will change a lot of things, but websites are as important to our customers today as they were 20 years ago.

All right, hold on to that. One of the things I think about a lot is why you would start a website. I’m in media, so if I were to start The Verge today, I don’t know that my first move would be to start a website with a CMS and a staff of people who write. I might just start a YouTube channel or a TikTok channel.

One of my favorite kinds of TikTok is small business TikTok. I’m deep into the pressure-washing community. Pressure-washing TikTok is one of the most calming things you can encounter. I’ve gotten my house pressure washed a number of times, needlessly now, because it’s so good, and I own a pressure washer. That’s a kind of direct marketing that’s very cheap, it’s very effective, it’s very human. And it seems like that community isn’t racing to go set up hosting and have a website.

What’s the thing that tips people over from, “oh, I should actually go away from these platforms the audience is built in” to “I need to run something of my own”?

I think the platforms are fantastic in a way. Thanks to these platforms, these micro businesses — your pressure washing community — can not only meet and talk on their Usenet group, they can share videos, they can transact, they can reach people who have interest. They reached you, which would’ve been impossible in the past, so I think it’s really good.

I don’t see that removing the need for them to have their own website or their own space where they represent themselves in the best way they can. While they may not be racing, as you say, I think what we do find —  and I talk to our customers on a very regular basis —  is that there are many reasons they build their website. They might build their website because they want to have a place where, once they reach someone on a platform, they want to be able to direct them to something that they own and create, and where they have a full representation of themselves. It may be to accept a bill. It might be a payment, or it just might be everything in my art gallery. One of the customers I talked to has an art gallery and she does it for a very specific group of people. She put it online and it brought her a ton of traffic. It brought people to her studio that she’d never seen before.

I could tell you so many of those stories, and what you’ll find is that those websites give these micro businesses legitimacy. In fact, there’s a stat from one of our recent surveys where even in the millennial or Gen Z generations, four out of five people are saying, “before I buy something from a small business, I go check their website first. I just want to know that they’re legit. I check the reviews because I want to know that they’re legit.” Those things really allow a person who is a sole opener and just started their business to reach people in a manner that is not just good and positive but where they have a chance to compete with the big player.

But it goes further. I don’t know if people in this group have run a micro business, or somebody in their family just sells jewelry online. GoDaddy’s average customer sells like $100,000 a year on one of our systems. That’s really not much, right? That’s very small. But when you talk to them, it’s not just that the website gives their business legitimacy. It makes them feel more confident. That actually is the number one thing. If you are a sole opener, the number one thing that’s going to take you from there, to hiring your first employee, to being successful is that you have the confidence to get through the tough times. For many people, looking at that website and building that website makes them feel more confident about what they’re doing. And that’s something.

I was looking at your last earnings, and it seems like the total number of domains you’re servicing is reasonably flat, but then the revenue inside of those customers is growing. So you’re providing more services to a neutral amount of people.

The approach we took over the last year is to go into a lot of businesses and grow the customer account significantly. But GoDaddy has also done some acquisitions that had a customer attached to it in different parts of the world. Five years ago when I joined, my vision was to grow into adjacency. So the revenue is going up because we’re big in productivity solutions, in websites, and we have commerce offerings. Those are the businesses that have grown very fast. What we call the core business, which is the older business, grew by low single digits, while the new businesses grew 20 percent. That’s where you are seeing the growth.

While we see that growth, we also have a strategic imperative to divest and end-of-life some older platforms. That’s been a headwind for us, not just on customer count but on revenue as well. We actually shared with The Street that over the last couple of years, we had one to two points of headwind on revenue just with the divestitures and what we chose to shut down. But I fundamentally believe that us focusing the company… and we are not a very large company. We have over $4 billion in revenue, and we’re about 6,000 people. We need to make sure our people are working on the most important things. If that means that we need to divest certain pieces or just end-of-life and reach out to a set of customers and say, “Look, we just don’t do this anymore,” then that’s what we need to do. And you see that pressure in our numbers.

The idea that you’re going to provide more services to a set of customers —

Well, we’ll add more customers once we get past this headwind, but — 

The reason I bring it up that way is I’ve talked to a number of companies who are nominally your competitors. Squarespace has been on Decoder, Wix has been on Decoder, and companies that even aren’t nominally your competitors like Intuit. All of them are kind of in the same space, which is, “Okay, you have a site. Now we’re going to use your site to take over your back office.” I think the canonical example is you have a yoga studio or you’re a personal trainer, and you need to do scheduling. So the front end of your website needs a scheduling functionality, which means now you need a back-end scheduling functionality. That means you need a billing system, and probably an account management system or a sales management system. We’re just going to build all that for you. So you’ve gone from setting up a Squarespace website to having the entire back office run by Squarespace.

It feels like you’re headed in that direction. You have a number of these services you could offer. Where does that stop for you, or do you see more growth there?

I think we’re squarely focused on what our customers continually tell us is their number one priority and what they want from us: how do they work better with their customers? So our tools actually don’t go into the back office at all. That may be the strategy for some other companies. If you look back at GoDaddy five years ago, it had this mode called, “Dream It, Create It, Grow It, Manage It.” One of the first things I did was say, “No, we’re just going to do Dream, Create and Grow. We’re not going to do Manage.”

The company had actually bought another company or two, and there was a set of products that we end-of-lifed, and got out of those businesses. Because what our customers want from us, where we have the right to win, is in the early phase. For this customer, it’s how they engage with their customers. It’s not in the back office. And yes, over time that may be a partnership opportunity for us because, like you’re saying, we have a certain relationship and we can expand it. We can look at that. But strategically, we’re very much focused on what we believe is their number one opportunity: how they better serve their customers and what they want from us, which is the tools that allow them to do that.

There are two themes I want to pick up on there. One is around AI and one is about growth generally. But you’ve brought up the size of the company and some decisions you’ve made, so I want to start with the Decoder questions, which are a gimme. This is just LinkedIn thinkfluencer bait. So get ready. You said 6,000 people, $4 billion in revenue. How are you structured?

At GoDaddy, we have a picture that looks like a house, and it’s part of what we call our operating model. Just like a kid would draw a house, you have the roof and it has two slanting ends, and that represents our brand. That’s a strategic change from five years ago where I said, “Look, all things are going to come in to the GoDaddy brand, which serves this customer. We can grow the brand to serve more customers, but we’re not going to go out and be a house of brands.” The company had done an acquisition to sort of pursue that, but I said, “None of that. We’re going to be very focused.” So that’s the roof.

Under the roof, imagine a few rectangles that we call rooms. Each of those rooms is what we call “customer-backed.” Each room serves a certain customer population, and those are our business units. So at GoDaddy, you don’t have a business unit that delivers a product. You have a business unit that serves a particular customer base. For example, you have an independent customer. This is a customer who wants to do things themselves. There’s a business unit that’s entire job is to do the product, go to market, P&L management, and technology for that customer. So you have five rooms under that roof.

Below that, we have a set of horizontals, which are the platforms across the company. The first one is our care platform. This is what most companies call the call center. We call it care. We don’t believe we’re a call center, and we don’t believe we have call center agents. We have guides; We have care and guides. Their number one metric to hit is NPS, not a cost function.

Under care, we have marketing. This is not the marketing decisions for the customer population, it’s the mechanics of marketing. You don’t want five rooms fighting for their own Twitter accounts or SEM accounts. You want that executed in one place, so that part is centralized. You have the CTO group, which is e-commerce, data science, the data platform, the pieces that are shared across the company. And then the corporate platform, which is legal, HR, and finance. 

That picture is something I drew probably five years ago when I joined. Every leader in the company knows that picture, and the board sees that picture. It’s called a house, and that is foundational to our operating model. It defines the structure, and it defines how different teams engage with each other. If you’re a room or a vertical, you have a P&L and you behave a certain way. If you’re a horizontal, you have a way that you engage with the five verticals to simplify and be able to move fast while you centralize certain things.

My joke on Decoder is, “If you tell me your org chart, I can name 85 percent of your problems.”

That’s probably true. It’s also true for code. When I used to take over teams — it’s one of the things my bosses love to do because that’s how my career grew — I would take over teams and they’d always ask me, “What do you want to do when you’re taking over this tech team?” I said, “Send me two things. Send me the last annual HR report: every person, what raise we gave them, all the money for everyone. what do they get. And send me a link to the codebase.” Because if your code is structured a certain way, I can tell you which two leaders don’t work well with each other, everything about it. It’s Conway’s law, which says that the structure and organization is the function of the relationship of the organization that designed it. So it’s absolutely true.

And not just your org chart but your operating mode. You want the org chart. You want accountability structure, you want culture, you want your opmax. You want a few things. If the operating model doesn’t tell you 80 percent-plus, you basically don’t have an operating model. People are just making shit up then. They just work however they work.

All right. So instead of me guessing your trade-offs, you tell me. What trade-offs do you think you bought with the room’s structure?

The biggest thing about our structure is that by building a customer-backed model that goes into the same brand, we have dedicated people that focus on solving a certain customer’s problem. As the company grows and addresses the needs of different types of customers, the same group can’t really divert their attention to different customers at the same leadership table. The people in the company — and because we’re the world’s largest domain player this is a bit of a niche thing — serve the largest group of domain investors in the world. The people who serve independents should not be thinking about domain investors. That should not happen. That’s what the operating model does. It puts the customer population first, and then the business unit. Then the business unit worries about the go-to-market, the product, the P&L and so on. That’s the positive part.

The trade-off is that anytime you have a many-to-one relationship, you’re going to have some friction. You’re going to have some prioritization conflict. You’re going to have people saying, “I’m not getting my part enough because I have to depend on somebody to get it.” The horizontals that we have works in an accelerator brake model, especially as you get lower down into the tech platform or the marketing. The business unit wants to spend as much as possible as quickly as possible for their customer, but now they have to talk to a CMO who says, “Hey, wait a minute, this other business unit will get a much better return on this investment. So why are you getting this? Somebody else should get it.” The same marketing team that was the brake on the business unit is now an accelerator because marketers want to spend as much as possible. Then you’ve got the tech organization under it that says, “Hey, here’s the media mix model. It runs on machine learning. Here’s what it’s saying, and here are the returns. You want to spend in this channel, but actually the data says you should spend on this other one.” 

So you have this accelerator brake model throughout the structure. That causes what a lot of people call pain in the organization; I call it pressure. I say, “Look, if it’s pain, bring it to me and I will help you resolve it. If it’s pressure, just realize that it’s designed to do that. If you feel pressure talking to the other tier, it was built that way because you want the accelerator brake model.” Cars run better with an accelerator and a brake.

We were joking before that I’m a journalist. I just spend money. I don’t make a dime. It’s one of the great perks of being on the editorial side of the house. But what you find often is those models don’t have risk priced into them. If I take a riskier bet, I might get an outsized return, but it’s hard to quantify the risk inside of that dynamic.. It’s hard to say, “Okay, there’s actual real growth here, but we have to stop doing this.” We can use AI as an example. We’re going to forward invest a bunch into AI because we think that’s really useful and everyone’s talking about it. But actually, if you look at the whole of the AI industry right now, no one knows how this is going to make money at the scale that the investment is currently driving. How do you think about putting risk in them all?

The way I think about this exact comment about risk is that it’s a function of who is in the room when the decision is made. This is one of the reasons I don’t use the operating model that’s fully functional. I think if you have a leader — and it works great for many companies, it just happens that what works for us well —  but in the functional structure, you end up at the CEO table with people who are leading a lot of corporate functions. 

So let’s say that you have somebody that’s chief revenue officer or somebody that’s CTO, chief legal officer, chief people officer, or chief risk officer. I think the numbers matter. If there are four corporate leaders in a room and two people who are accountable for the performance zone for the business and making money, you’re going to end up with a lower risk decision-making set up. Those corporate leaders are designed to be in the productivity zone, and the productivity zone’s job is to reduce risk.

At GoDaddy, we have the GoDaddy leadership team. It is specifically weighed where the leaders of the rooms and the verticals — the business unit leaders —  are part of the group. So in any room where we’re making an assessment at my level on whether we should take a risk or not, there are more risk-takers than not. Your CFO’s always going to tell you to be more conservative. That’s his job or her job. Your CLO’s always going to do that. That is their job. What I want to do is create an environment where I have the head of the independents business there, the head of the partners business there, the head of the commerce business there. I’m there and the COO’s there. Now suddenly there are five, six, or seven of us who want to lean in into the risks, who want to explore it.

Of course, we can use our judgment, take a step back, and say, “Maybe this is a risk we don’t want to take.” But we’re not going to get stuck in a room where we’re the minority because then the majority will win. And you don’t want that. Business is about taking risks. It’s about taking risk in a manner that’s reasonable, that’s thought out, that’s backed by data, and that’s well understood. Or at least, you’re working hard to understand it as well as you can. So you just don’t want to create the setup where that conversation doesn’t happen.

So this is the other big Decoder question: How do you make decisions? What’s your framework?

At the company, we’ve had two large transformations over the last five years. The first was about evolving our software platform. That’s really about GoDaddy being a domain and hosting company, and your margins when you’re a domain player are your gross margins. But when you’re a software player, you have much higher gross margins. So, let’s build software, which was a transition for the company. 

Your question is really about our second transformation, where we make decisions based on evidence. I’m a big believer in belief. I want to believe in things that are not there yet, but I’m also equally big on the scientific method. So the approach for us is, “If you have an idea, I don’t want to spend a lot of time talking about it. If it can be tested, just test it. Once you’ve tested it, you can just put it in GD experiments Slack channel and anyone in the company can see it.”

You have the same thing that’s now taught to first graders or kindergartners. You have an observation of the world, you have a hypothesis you want to test, you have a test you’ve designed, and you’ve got the result. When you run it, you put it in Slack. If the data makes the decision, we’ll go against it maybe 0.1 percent of the time,. But normally the data is leading us, and the decision is delegated because the metrics are clear and the prioritization is clear. How you track that metric is clear. In that 0.1  percent of times, let’s say it’s a one-way door or you need a HIPPO decision (HIPPO being “highest paid person’s opinion” wins). It’s kind of the antithesis in my mind of data-driven decisions. But if it has to be a HIPPO decision, well  let’s reduce that number to as small as possible, and then figure out whoever’s the best person to make it. If I’m the best person to make it, let it come to me. If the chief architect should make it, then he should make it. But you try to reduce those decisions to the fewest possible. Hopefully that helps.

The two Decoder questions help me figure out how to ask all the rest of the questions. So now I want to ask about AI. If you all haven’t seen the GoDaddy AI platform, it’s called Airo. I believe it can even fill out your LLC paperwork for you. You can tell it, “I want to buy a domain name,” all the way to, “I’m interfacing with the government to tell them I have a business.” It’s a lot.

We’re doing quite a few tests in that area.

You set up the website for you, you can fill out the copy for you. This is a big investment. 

Today as we speak, we’re competing with Satya Nadella, who’s onstage at Microsoft Ignite talking about scaling laws. But all the big companies are telling me this is a platform shift. I don’t really quite know what it means, but I have some ideas. Do you see the scale of that investment, the decision to go after using generative AI to make things for your customers, at the same scale or return on investment as the big players you’re seeing?

The piece we’re really focused on in our business is that last mile: how generative AI adds value for our customer. One of the things we always say at the company is, “Yes, we are going to take some of the friction out for the customer to build something, but we’re always going to make sure that the customer is in control of their brand.” It’s ultimately their decision. It shouldn’t just auto-magically happen. It should happen magically, and then they should decide whether they like it or want to edit it. What we find is that our customers don’t have copywriters. They don’t have those people. They don’t have lawyers to advise them. They don’t have marketing people to advise them. They don’t have a webmaster. So if generative AI can start them off on an About Us page that makes sense or give them an image for a logo, they’re like, “Oh, this is pretty good and I’ll customize it from here.”

That’s a fantastic start because a lot of people are not going to start their side hustle or their micro business even if they have a template and everything. When they actually have to choose the images or type in the text, they may not be as proficient in English, or they’re just not able to bring it together. They’re a really good plumber, just not as good at writing about how good of a plumber they are. Well, if generative AI can do that, I think it’s a fantastic thing. We don’t look weirdly at people who use calculators to do math, do we? And I think generative AI does that work for text for our customer.

I want to be 100 percent clear: my dad looks super weirdly at me when I use the calculator.

But the question is, will you at your children?

I guess that’s the general AI case. I’ve heard that from everybody. I’m saying that, specifically for GoDaddy, you had to invest in the Airo platform.

And when I talk to the big companies, with their big investments in foundation models, and H100s. It’s hazy how you get from that to what they think will happen. Is it hazy for you? We’re going to invest in Airo and that’s going to drive growth?

No. The investment level in Airo is much smaller than building a $500 million model. It’s nowhere near that. Because Airo is very, very close to the customer, we can run 100 tests within one year to see whether it’s actually creating value or not. Does it actually move customer willingness to pay or not? Because if you gave the customer something and they’re not willing to pay more for it, you didn’t really meet a true need for the customer. Yeah, you gave something away, and maybe some people use it, but the real breakthroughs are when you create something, get the customer to use it, and they’re actually willing to pay more for it. And we have that data, right? We are able to see that data in a year. 

In terms of the very large investments in AI, typically what I say to people is, “I remember when the internet came along, and there were some very big companies putting cables all over the world.” We’re still using those cables even though those companies went away. We’re using that cable today, and they put in so much cable. They didn’t imagine there was going to be video on the internet back — or maybe they did imagine and invested really early. I think that’s the cycle of technology. In many cases, from electricity to mobile phones to satellites to now AI and the internet, large investments have to be made to make it available to everyone, and sometimes the usage comes over time.

Whose models are you using for your AI products?

The way we use it on our platform is that we have a single API. Whether I want to write a review for my direct [reports] or GoDaddy’s system is writing something for its customer, everything goes through one common set of APIs. It’s easier to do moderation, and behind that API, we have all the models connected. So we have connections to OpenAI, to Gemini and the Microsoft Azure version of OpenAI, and that allows us to do some testing at the API tier to see where we’re getting a better response, and, frankly, often a cheaper response that’s just as good. Because we can pass signals to the API tier and the API tier can decide, “Okay, I’m going to use this model instead of this one. Was the response just as good or not?”

Do those feel commodified to you? I’ve heard a lot of people say that the core model functionality is going to get commodified and prices will come down.

They know something I don’t know because I think it’s too early to say. I think it’s changing so fast. The new models, even in the [OpenAI] GPT-4o model, are just different. It’s different from what we expected, and I think the more you have intelligence and understanding, this is going to change a lot over the next year or two. It’s too early to term it as a commodification.

You’re one of the few CEOs I talked to who’s willing to put real price tags next to AI work. Do you see consumer price sensitivity, like, “I’ll get you a better answer and that’ll cost you more money,” or is it just AI?

There is consumer sensitivity, especially for our consumer. I’ll tell you a quick story. The first customer I really talked to in depth was an HR person at a roofing company. Her husband had two barbershops and she was a webmaster. She was upset because her $200-a-year subscription was going to increase because the Dreamweaver instance we supported for her — that was at end-of-life, 15 years, when I talked to her — was not working anymore. The reason she emailed me and called me was because that little delta in the year was a big deal for her. Other people may say, “What’s $50 a year?” But for this customer, it was really a lot of money. So I would say the micro business customer is very, very sensitive.

We have a test right now where Airo scans your website and tells you what you should change on that website. It’ll do it for you too, but you can do it manually yourself — this is on managed WordPress. It also tells you how to improve your SEO or improve your labeling, navigation, content, or make your pictures lighter so it loads faster. It will go through your site and give you the top recommendation and order them. We have a price tag on that right now that’s being A/B tested, and so far, the signal is encouraging. That’s just one of many paywalls we’re testing right now where it’s real AI functionality that has a real price tag.

And look, not everyone’s going to pay for it, so we’re going to have to find some balance between giving enough for free that everyone values it and holding something really important back that the percentage willing to pay for it pays for it. Then, we can value-based price that and make it work for everyone.

I warned you that this would be a therapy session for me about the web before we started. You brought up Google. This is where I think I have existential fears about the internet that we live in. What I specifically mean is the idea that you can do automated SEO, which implies, one, that there’s a right answer to SEO problems–

It’s not automated. It’s a suggestion.

Sure. It implies that Google is static, which it is not, and that Google Search will continue to be a reliable source of traffic over the long term. Will Google be a going concern in the Trump 2 administration? It’s a little bit up in the air. I think it was yesterday that the Department of Justice was like, “We’re going to make you sell Chrome.”

[Laughs] I probably couldn’t afford it.

The guy from Rumble is like, “I’ll buy Chrome.” It’s an open market.

Let me bring you to that example since you talked about the tool we just launched. You may not be a technologist. By the way, lawyers are like coders. Contracts are sort of–

Yeah, you are. So there’s a logic mindset to it, and you’re absolutely right that those things in Google and search are changing. But you are a very sophisticated person when it comes to understanding that those things are changing. I’m talking about a customer who builds a website, sells pressure washers, but does not put the word “pressure wash” in the title or the H1 tag. Those are really important if you want to rank. That’s no secret. That’s not exposing some secret Google algorithm.

But let me just ask you this. This is the fight. There’s some SEO that people understand–

For these micro businesses, these folks don’t know the basics. 

But it’s the idea that the internet will be full of AI-generated content that will go into LLM training data, and that our search behaviors will change because OpenAI, SearchGPT, or Gemini will just tell us the answer. There’s incredible pressure on the very notion that those platforms will continue to send traffic to websites, as opposed to, “All of our app and marketing is on video platforms, and the last step is you come to my website to convert.”

So I worked in another business where the large platform companies, including Google, are very large. The hypothesis has existed for over a decade that the large platform companies will continue to take more and more of the traffic — which they have, it’s true, or the data suggests so. That will lead to fewer websites, fewer domain players, and so on, and it will be a loss for the open and free internet. That will be a big loss because the internet was built around this idea that anywhere in the world —  of course, if you’re not on [behind China’s] Great Firewall or something — you can reach any piece of information in the world in a very flat manner. You just have to type a URL and HTTP takes care of it. It’s amazing. It’s unbelievable what happens on the internet, right?

My fundamental belief, and maybe I want to believe it rather than prove it, but my fundamental belief is that AI can generate data and platforms can keep trying to take more and more of the data. They’re definitely creating supply and demand. It’s not like they’re just eating everything; they’re creating the supply too. They’re creating content that leads to people using a lot of data or bandwidth.

But my fundamental belief is that the more AI grows in that, the more important your words are going to be on The Verge because you are involved with it and no AI is doing what you’re doing. The same thing is true for our customer. Let’s look at micro businesses. If a mom decides to sell jewelry online, believe me, I can’t do anything too exciting with it. I can maybe give the mom a few ideas on how to structure it, but that is a real business. It may put food on the table or it might just give her joy and satisfaction in life, but that’s a real thing. I think it’s very hard to replace that, and I think it’s only going to make it more popular.

We have a PR line at GoDaddy. I’m going to mess this up because I didn’t read it carefully enough, but it’s something like, “Welcome to the smallest Black Friday ever.” Because this Black Friday, our services show that more customers in the U.S. want to buy from small businesses. And get this, they’re willing to pay more to buy from a small business than buy it from a big business. You don’t hear that very often, but this year on Black Friday, that’s what consumers are saying. I think the more you see this data and you see this AI, the human element is going to be more important. I want to believe that and I do believe that.

When you look across all of your customers, is Google still, by far, the largest referrer of traffic?

Google, between Google Search and YouTube, is a very large referrer of traffic. I think social is continuing to grow for sure, but I think Google still has the dominant position.

Bluesky is growing. I think right before we started talking, it hit 20 million users. It’s not a lot compared to Threads at 275 million and Twitter at 300 million, and then Facebook and TikTok. One of their differentiators is they don’t suppress content with links in it. They’re very proud of this. Jay Graber was on Decoder, and she said, “We are part of the web. We want this platform to feel like the web.” Other platforms like Threads, TikTok, or Instagram don’t allow links out, right? They’re closed ecosystems. 

How do you think about running a company where we want people to grow by making websites, whether they’re micro-entrepreneurs or larger, but they’re up against marketing channels that want to keep them inside of their ecosystems?

Look, it’s a fact of life. You described it right; it’s exactly how it is. What we tell customers all the time is that, “Yes, you should use the platforms to reach new people, but ultimately, you should bring them to your website, your experience, or your store because that allows you to do that. The internet offers this amazing capability where now, if I start a brand, let’s say it’s the same example: homemade jewelry. If there are 100,000 people in the world out of 7-plus billion who are interested in my jewelry, they can all reach me through my website and transact with me, and technology allows us to do all of that. One person can run that whole thing.I think that’s just continuing to happen. 

Yes, many of the platforms don’t allow links out, but there’s a need for the consumer. If I see an article from you on a platform, yes it’s interesting, but if it engages me, then I want to see the full picture of what you’ve got to offer, and I’m going to get that on your website. I will come whether they link it or not. They can choose not to link it. I’ll find it if I’m really interested.

I’ve got a few more and then we’re going to take some questions in the room here. Your investment in AI, is that paying off? Is that a positive ROI yet?

No, we’re still in the investment cycle on Airo. We have really good early data for Airo, so we are going to continue to invest. In fact, I’ve publicly said in earnings calls that while we are testing significant monetization opportunities. But my primary metrics are discovery and engagement. I want people to discover Airo all over the world. I want them to have engagement with the different pieces of Airo — Airo has a lot of different components in it. Those are the two primary metrics that I’m focused on. Yes, we will monetize as needed. We did a three-year financial plan, we did an investor day in February 2024, so it hasn’t been a year yet. And in that financial plan, you’ll hear me say, “There is no Airo in this.” There’s $0 for Airo in the three-year financial plan. There will be some return, but I didn’t want to even have our CFO guide to it.

When do you think that will change? Is it after three years?

I think if we’re making good progress, it’ll happen earlier. But things were changing so fast when we did that presentation. Just between February and November, so much has changed in Airo: the take rate, the uptake, the engagement. Believe it or not, we’ve had 3 million people discover Airo, and we’ve had over half of them engage with some part of Airo. As that engagement deepens, we know that leads to monetization, so I think it’ll happen a bit sooner than we expected. But we’re going to keep driving that discovery and engagement number.

Is that based on a fixed-cost estimate for how much the models will cost, or are you expecting some variability?

It makes some assumptions on what the models will cost, but unlike some other people, that’s a bit of a smaller expense for us if you compare to others. The reason is that we have a number of machine learning models. They’re not generative AI models, but a number of machine learning models internally do a lot of the heavy lifting on what Airo does. We have actually tried very hard to minimize how much we would have to pay somebody else.

I always remind the team that when I came to GoDaddy, our ARPU was $167, and we crossed $210 last quarter. With that $210, we’ve got to pay [cost of goods sold], pay our salaries, market this product, and invest. And then we still have to have a margin that starts with a three handle. With gross margin in the mid-sixties and a three-handle operating margin, you don’t have a lot. You’ve got to make every bit work.

You run some managed WordPress hosting I believe?

Do you know what’s going on with WordPress right now?

I am familiar with this. Yes, I follow it closely, actually. 

How do you think that’s going to resolve? This is a big lawsuit.

I don’t know how the lawsuit’s going to resolve, but I do understand why it happens. I think you’ve got two big players in the WordPress space taking different positions, and it’s got to play out over time.

Which position are you going to take? Do you contribute back to open source?

We do. Depending on how you count it, we’re either the second-largest or the third-largest contributor. You can go to Five for the Future; it’s published data. You can look at it.

So Matt [Mullenweg, WordPress CEO]’s not going to yell at you?

Could change. If he’s listening.

He listens. All right. I think we can take some questions from the room if you have any?

Speaker 1: Over the years, GoDaddy has been innovating, really pushing the boundaries of what you offer, and reinventing your offerings over time. Do you have a sense that the pace of disruption and the required innovation you need to bring to market to stay competitive is becoming faster and faster to the point that it gets difficult for a single company to stay current and at the edge?

If the question is if the rate of innovation will continue to accelerate, I think it is. There’s this idea of “gradually then suddenly,” and tech follows that cycle again and again. The “suddenly” parts are getting steeper and steeper. OpenAI had more users faster than the company before it.

I don’t know about any one company. I think technology is moving so fast that it has a disruption cycle built into it. At GoDaddy, we’ve worked very hard to make sure that we can enter new businesses and continue to grow the company. That’s worked really well for us over the last few years. I often say to the team that, “You might have time, but I have no time at all. I’ve got to make everything happen now. But at the same time, for the right idea, I’ve got all the time in the world.” Sometimes people say, “Why do you say you have no time and then say you have all the time in the world?” So it depends on the context, doesn’t it?

Speaker 2: Now that you’re moving into all of this adjacent white space, how do you compete against the likes of Shopify? Because they’re coming at it from a different angle. What is it that makes you unique and different right now, and what’s the plan ahead?

Entering an adjacency, especially a large one that’s growing fast, is always going to be difficult because it’s crowded. It’s not quite white space. There’s a lot of players in it already. Our positioning is very clear on the commerce piece. We actively market our commerce offering only to our existing customer base. So that gives us a CAC advantage. It is literally built, designed, and invested in with that view. We don’t go out and try to tell the world that we have this offering because that would be prohibitive in terms of marketing.

What that allows us to do is to build a competitive commerce offering, which of course is a multi-year journey. It allows us to build it while maintaining a certain burn rate on it so we can get to scale. That’s one of our competitive advantages. We have almost 21 million paying customers today. In that base are a lot of customers that want commerce. In that base are a lot of customers that happily take GoDaddy. We have good conversion rates on it.

But you’re right. You’ve got to find your angle, right? Otherwise it’s just too hard. 

Speaker 3: When you started your conversation around generative AI, you said you mostly used generative AI to build something for your customers, and then that kind of conversation shifted to margins and the financial plan. So have you evaluated areas within the organization where your cost to serve each customer can come down using generative AI? Has that been a thought process?

We have something we call the GoDaddy innovation model, which is like a staircase. The idea is that everything we do starts with magic. Some human knows how to do it, but none of us understand how it happens, so that’s magic. When it scales, you have to structure it, which means it’s going to be process-powered rather than human-powered. If you can do better with that, you’re going to be automated, which means it’s machine-powered. Then the highest step in the staircase for us is that it’s data-powered, and we call it magical again, but it’s just AI.

So a number of processes at the company, even our corporate functions, have targets to drive processes up the staircase and get to that level. The simple idea is that as you go up the staircase, you gather more differentiation, more scale and more personalized experiences, even for a process internally. So yeah, it’s a big thing. Whether it’s in our care guides — we have better and better tools to turn our care guides into super guides — all the way to how we do our marketing, how we do our spend, or how we price, more and more of them are driven by data.

Speaker 4: How do you think about customer success and customer attention given the segment of the market you’re focused on? I’d love to hear your perspective on that.

I think people view this as a really tough customer retention or customer success vertical. We have an 85 percent published retention rate. We really think about customer churn, not product churn. In our customer base, you actually see people who tried a business, they got a domain name, but it didn’t work out after nine months or 12 months, and they need a new idea. But they still have to put food on the table or they have the drive to do something. They’re going to do something else. Are they doing it with GoDaddy? We want to keep that engagement with them.

I think a lot of businesses are not structured to provide a high level of care. It’s sort of treated as a cost center. In our case, the care organization is a true competitive differentiator. We call what they do “guidance.” Yes, our care guides have some targets, but they don’t have a script. They don’t get told, “Ask the customer X, Y, or Z.” The goal is to first understand what the customer wants and solve their problem. Then, we have sales motions because we can end up spending a lot of time serving a person.

We have a transactional NPS (net promoted score) in care and it was very, very high. When I came from the travel industry to GoDaddy, I was shocked. That transactional NPS was phenomenal for what I, at that point, called the call center. Today, we’ve publicly said that our transactional NPS in care is 65-plus. So we are one of those rare companies where if you have a problem, we’re going to work very hard to fix it, and we’re going to fix it a very high percentage of the time. And when we fix it, a lot of customers become customers for life. Then, you get the lower churn rate out of that.

What’d you do for the Dreamweaver barbershop lady?

I talked to her three times on the phone. She said, “This is really nice to talk to the CEO of a public company.” I said, “You’re one of my biggest customers. The  ARPU is $167, I’m asking you to pay $275.” I said, “You’re a big customer. You get to talk to the CEO.” She loved it. She talked to some of the tech folks. 

I asked her how she updates her website. I don’t know if you know how old Dreamweaver is. It’s been end-of-life for a very long time. Some designer-developer built it for her. She would copy the new image to a certain folder, rename it to the name that was in the code, refresh the site and the new picture would get picked up. That’s how she updated images. She didn’t know how to change the code. Iit took a couple of weeks, but we convinced her that she’d be much better off on a new platform. And she finally said, “Okay, I’m going to pay the $275.”

Oh, you didn’t give her a discount?

The team might have. I didn’t. I don’t think I gave her a discount.

That’s great. What platform did you put her on?

I think she actually went to one of the newer hosting platforms.

Hey, you know, sometimes there’s a discount and sometimes there’s no discount.

[Laughs] That’s how you get revenue and growth, everybody. No discounts. Aman, thank you so much. This was amazing.