Awin is the all-in-one platform for smarter partner marketing. With award-winning technology, expert support and a global network of 1 million partners, Awin helps brands discover, manage and scale high-performing partnerships that drive measurable growth. Learn more.

[00:00:00] Marcus Johnson: Today's episode is sponsored by Awin, the world's leading brands grow through partnership marketing with Awin's enterprise technology, AI-powered insights, and expert US regional support. With 30,000 advertisers, one million partners worldwide, and more than 20 billion in partner, $20 billion that is, in partner-driven revenue generated last year, Awin helps brands turn partnerships into growth.

[00:00:19] Marcus Johnson: Learn [00:00:20] more at awin.com/us. Join EMARKETER’s The Future of Digital Summit on September 15th in New York City for a full day of research, expert analysis, and executive conversations on AI visibility, commerce, media, and the future of marketing. Click the registration link in the show notes for some tickets[00:00:40]

[00:00:42] Nate Elliott: Hey gang, it's Friday, August 28th, and this is Behind the Numbers, the podcast from EMARKETER hosted by Marcus, who's now gonna take over.

[00:00:50] Marcus Johnson: Used to be hosted by me. Uh, Nate, Grace, and listeners, welcome to Behind the Numbers, an EMARKETER podcast made possible by Awin. I'm Marcus, and joining me for today's conversation we have two [00:01:00] AI experts, uh, one part-time host, uh, we have living in, I think he's in New York right now.

[00:01:05] Marcus Johnson: He's there sometimes. I am. Principal analyst Nate Elliott.

[00:01:08] Nate Elliott: Buongiorno, Marcus.

[00:01:09] Marcus Johnson: Hello. We're also joined, uh, from California, analyst Grace Harmon.

[00:01:13] Grace Harmon: Hi, Marcus. Nice to be here.

[00:01:14] Marcus Johnson: Hello. All right, today's fact.[00:01:20]

[00:01:20] Marcus Johnson: So this one, this one's quite sweet, I think. Looking at distant stars means looking into the past because light takes time to travel through space. So like when you look at the sun or the light you get from the sun, it takes like eight minutes to reach Earth and so technically when you're looking out into the galaxies, the [00:01:40] universes, you're looking into the past.

[00:01:41] Marcus Johnson: It's so sweet, isn't it? Although kind of concerning 'cause if the sun has blown up, it might have been blown up already. We wouldn't know about it, which was one of my other facts that I had from earlier this week when I said that the sun was gonna implode in five billion years.

[00:01:58] Nate Elliott: Have any of the Avengers [00:02:00] ever tried to blow up the sun or save the sun from being blown up?

[00:02:03] Marcus Johnson: Maybe that's the new Doom movie.

[00:02:04] Nate Elliott: I just had to help you get the Avengers in the podcast again.

[00:02:06] Marcus Johnson: Thank you. Today's real topic, the AI platforms race.

[00:02:14] Marcus Johnson: So Nate, we were talking about, we've just finished some research on AI platforms. It seemed to go, I mean, relatively [00:02:20] unnoticed, you could argue, from people who aren't tracking this as closely as you are. But a lot of these platforms are starting to hit the one billion user mark. Take that as you will. Uh, i- is, is that as a marker of significance or not?

[00:02:31] Marcus Johnson: But that's something, um, that you could say i- is a milestone in terms of, um, AI adoption becoming, um, pretty significant for a lot of these major [00:02:40] platforms. That said, i- i- one billion isn't the goal. Folks are gonna need a lot more users than that to really put their stamp on and make an impact i- in this space.

[00:02:48] Marcus Johnson: What- To you makes the next few years so important, as you write in the piece, and industry-defining for, for, for AI marketing?

[00:02:55] Nate Elliott: Well, for marketing in particular, it's important because the companies that win [00:03:00] in the next year or two are gonna be the ones who decide what advertising in AI looks like. If ChatGPT wins, then their ad formats, their targeting tools, and everything else they do around advertising is likely to become one of the standards, if not the only standard, for how we do this for years to come.

[00:03:17] Nate Elliott: If Google wins, then the [00:03:20] standards for AI advertising will probably look a whole lot like the standards for traditional search advertising. If Claude wins, then perhaps the standard expectation will be that there aren't ads in AI at all. But whoever wins this race, this horse race over the next year or two, is gonna have a really big say on what all of this looks like.

[00:03:38] Nate Elliott: And if we've learned anything [00:03:40] from how this went in search and social and mobile and lots of other ad channels and formats, the companies that win in the first three to five years tend to stick around for a very long time and be the outsize player in the market for years to come.

[00:03:54] Marcus Johnson: Grace, any thoughts on this, this timeline?

[00:03:55] Marcus Johnson: It's so interesting how the next few years will be so important given that we weren't really talking about large [00:04:00] language models until a couple of years ago.

[00:04:02] Grace Harmon: I think that Nate said it, you know, it's gonna be things like formatting, like placement, like cost measurement that are gonna be decided by who the usage leaders are.

[00:04:09] Grace Harmon: I think you make a really good point about if it ends up being, um, an established current advertising giant, like if it were to be Meta, which is unlikely, or if it was Google, it's gonna be interesting to see [00:04:20] how the threads of traditional advertising and digital advertising as we know it now might show up in AI advertising.

[00:04:26] Nate Elliott: Although have you noticed that the ChatGPT ads look a lot like Facebook ads? They do. I kept thinking that they would look like Google ads the more I look at them because they have the image in them. They're just, they're basically just Facebook right-hand rail ads.

[00:04:38] Marcus Johnson: Nate, what's the, um, [00:04:40] potential prize here in terms of, um, the AI advertising pot of gold at the end of the rainbow in, you know, you said a couple of years, so call it 2030?

[00:04:48] Nate Elliott: I mean, it, it depends on what part of the AI market you're looking at. But our forecast says that in the US alone, AI advertising or advertising in AI interactions [00:05:00] will generate about $68 billion in advertising revenue, uh, in 2030. Now, we make some distinctions within that 68 or so billion dollars. And so when we look at the ads just within standalone chatbots, when we look at the AI advertising in ChatGPT and Gemini and Claude and [00:05:20] Perplexity and standalone AI chatbots like that, it's less than $10 billion.

[00:05:23] Nate Elliott: I think it's eight and change. Because what we're seeing right now is that- Google on the advertising side has figured out how to advertise alongside AI by putting AI overviews on the tops of their traditional search engines results pages. And so right now there are [00:05:40] millions of advertisers, uh, who are advertising alongside AI responses in the form of AI overviews who've never asked to be advertisers alongside AI, and they're advertising next to AI that consumers didn't ask for, 'cause people don't request AI overviews.

[00:05:56] Nate Elliott: And yet today, in 2026, this [00:06:00] is the main way we're seeing ad revenue flow towards AI search results, and we think that's likely to continue to be true. We think that'll still be the majority of the market in 2030. So there's this big prize in the US, you know, $68 billion four years down the road, and the question becomes more complicated when you get [00:06:20] into is that gonna be in chatbots?

[00:06:21] Nate Elliott: Is it gonna be in AI summaries at the top of search results pages? Is it gonna be in those weird chatbots that live inside search results pages, like Google's AI mode? Um, the answer is it's gonna be in all three places, but we actually think chatbots may end up being the smallest of those three categories.

[00:06:37] Marcus Johnson: So in your research of AI [00:06:40] platforms, you're looking at not just the platforms, but you're, you're breaking that down by basically everything from kind of top of funnel to bottom, awareness, trial, usage, um, and then paying for stuff. And one of the points you, you note early on is ChatGPT from OpenAI i-i- is the best at convincing people who have heard of [00:07:00] it to try using it at least once, uh, 60% of people you say.

[00:07:05] Marcus Johnson: Why is ChatGPT been able to build a s- a stronger adoption than the highly recognizable competitors like a Meta AI or a Copilot?

[00:07:14] Nate Elliott: The bottom line is it just ow- owns the top of the funnel. Uh, first mover advantage is not [00:07:20] necessarily durable over the long term. If you don't believe that, then go ask the people who Own AltaVista or MapQuest or Friendster or any of the other first movers who over the years have fallen by the wayside.

[00:07:34] Nate Elliott: So it, it doesn't guarantee that ChatGPT is going to have the biggest awareness and the biggest [00:07:40] trial in the long term, but certainly right now it has guaranteed that it has those caveats or, or, or those, uh, plaudits, I should say. You know, in our survey, about 95% of US online users said they'd heard of ChatGPT, and as you say, almost 60% of the people who've heard of ChatGPT in the US say they've tried it at least once, and [00:08:00] that's what you get for being synonymous with the category.

[00:08:03] Nate Elliott: By launching first, ChatGPT became synonymous with this category, and it, it may or may not stay that way over the long term. Our forecast model says that Google, between Gemini and AI mode, will have more monthly AI users in the US by the first quarter of next [00:08:20] year. But for now at least, uh, you know, ChatGPT is the Kleenex of this category, and, uh, and for that reason, it's got really high brand awareness, and if people are just gonna try one AI chatbot, that's the one they're most likely to try.

[00:08:34] Grace Harmon: Yeah, I mean, I think you painted the picture of it's, it's... Those are the words I was gonna use. I think it is first mover advantage, exactly. It's just, [00:08:40] like you said, that that doesn't guarantee stability in that category forever. I mean, I think it's like a three-player race now, uh, for consumers specifically, because ChatGPT has the scale.

[00:08:49] Grace Harmon: Like you said, it's synonymous with GenAI and chatbots. Gemini has Google's distribution, and then Claude is doing so well with engagement and monetization.

[00:08:58] Nate Elliott: I mean, it's remarkable that it's a [00:09:00] three-horse race right now. Mm-hmm. Grace, you and I were talking about this offline, but it was a two-horse race six months ago, right?

[00:09:06] Nate Elliott: Mm-hmm. And then Claude, which always had a really good reputation but never had a lot of users, kinda came, not out of nowhere, but as one of the leading players, as a genuine competitor in this market. It kinda came out of [00:09:20] nowhere in the last six or eight months, aided by things like the Super Bowl ad and the Department of Defense, uh, conflict and a bunch of other things that brought it into public attention and separated it from a lot of other tools that are sort of languishing with a lot fewer users at this point.

[00:09:36] Marcus Johnson: Wonder how much of it is people not knowing how good- [00:09:40] To use it. Uh, and what I mean by that is, in your research, Nate, you know that ChatGPT has got the highest awareness, 95%, Gemini 90%. But some of these others have very high awareness, too. Meta AI, almost 90%, uh, Copilot 80%, and then Grok and Claude further behind.

[00:09:57] Marcus Johnson: But still, people are pretty, pretty aware of [00:10:00] them. What's interesting, though, is getting from awareness to using it. Gemini people might think to themselves, "Well, how do I use Gemini?" Similar to what we've talked about with Google having AI mode and AI overviews and Gemini, and people not knowing when am I using AI and when am I not?

[00:10:14] Marcus Johnson: Copilot, what does that mean? Meta AI, do I have to use one of their services to use it, but is it, is it just that I'm using it, [00:10:20] the product, and therefore I'm using it or do I have to do something specific? So I wonder how much of this is just ChatGPT people know how to start using that product.

[00:10:27] Nate Elliott: Yeah, I mean, certainly the, the distribution that Google has and their ability to promote Gemini and the, the other AI tools that Google offers to folks who are using all the other tools from Google that people like to use, like [00:10:40] Android phones and Gmail and the Chrome browser, and lots of other tools that have billions and billions of users, never mind the traditional search engine that certainly helps bring people into the Google AI ecosystem.

[00:10:52] Nate Elliott: The Meta AI example's an interesting one for me because they've spent the last, what, year and a half, almost two years Slapping a meta AI label [00:11:00] on the regular old-fashioned, not terribly good search experience on Facebook and doing likewise on Instagram. Really, the, the main place that they've done a genuine AI integration into search is on WhatsApp, which is, in the US at least, by far the, the least popular- Mm-hmm

[00:11:18] Nate Elliott: of their big tools. [00:11:20] Uh, and so if anything, this tells me how few people use Facebook Search or Instagram Search- Mm-hmm. Mm-hmm ... if trial of those tools is so low, because when you just search for a friend or an account on Facebook or Instagram, you're told that you're using meta AI, and still the conversion rate from awareness to [00:11:40] trial on meta AI is surprisingly low.

[00:11:42] Marcus Johnson: One of the things that I, uh, came to mind when we were thinking about this episode was this, these recent headlines about OpenAI and safety concerns. Mm-hmm. Uh, so, uh, I was... I mean, my question I had was how much can headlines like OpenAI pausing some model work over safety concerns impact adoption, or even just model [00:12:00] perception.

[00:12:00] Marcus Johnson: So this headline from Julia Shapiro of The Hill, the ChatGPT... She explains the ChatGPT maker said in a blog post that it temporarily slowed the pace of scaling after its models independently gained access to the internet during testing and hacked into the tech company Hugging Face. It instituted a two-week pause in reinforcement learning, uh, training, in which AI models [00:12:20] learn through trial and error without human involvement.

[00:12:22] Marcus Johnson: In addition to the Hugging Face incident, OpenAI separately pointed to the recent finding, uh, that its next generation model, Astra, may have critical cyber capabilities under its preparedness framework. OpenAI CEO Sam Altman said, "Model progress is now extremely rapid, and we always said we would take action if we felt the model [00:12:40] capabilities were outstripping the pace of safety and alignment."

[00:12:42] Marcus Johnson: And there's been a few other examples. Anthropic revealed late last month that, uh, its models escaped their testing environment and accessed the systems of three different organizations during testing by a third-party partner. The Claude maker saying this was due to a misunderstanding. Meta had a similar incident as well.

[00:12:55] Marcus Johnson: What are your thoughts on these big headlines about safety concerns surrounding AI [00:13:00] impacting adoption of any of these models?

[00:13:01] Grace Harmon: I think for public perception, it really cuts both ways, 'cause I think that it adds to all of these anxieties that these frontier models are moving a lot faster than their safeguards, and in that aspect, I think it could dissuade some people from adopting or exploring AI just out of anxieties and fear.

[00:13:15] Grace Harmon: I also think that for other people, this voluntary pause is going to [00:13:20] add to OpenAI's credibility that, you know, this, this narrative Whether or not it's true that it's willing to prioritize safety over speed.

[00:13:27] Nate Elliott: It's funny to me that everyone's piling on and saying, "Oh, wait, we've had a breach as well." I don't know what you think about this, Grace, but it feels like these platforms, the companies that make these platforms recognize that they're not going [00:13:40] to win over people who are skeptical about AI.

[00:13:42] Nate Elliott: And so they seem to be playing the game of how can we look more advanced and more cutting edge to the folks who do use AI and who like these technologies. And you kind of saw it with early this year when companies were saying, "We're not gonna release our frontier- Mm-hmm ... models because they're too [00:14:00] dangerous."

[00:14:00] Nate Elliott: And, you know, now they're all saying, "Oh, they've escaped their testing environments and hacked into other companies." And I've no doubt that these things have happened, but I also think that the companies maybe would've downplayed this if ChatGPT, uh, hacking Hugging Face hadn't gotten some amazed [00:14:20] and wondrous reception.

[00:14:22] Nate Elliott: Uh, I think the other platforms may have downplayed, uh, their incidents. I, I think, you know, it sounds like Anthropic, for instance, uh, their models escaped their testing environment before ChatGPT made the announcement, and they chose not to announce it until they saw that people sort of thought [00:14:40] ChatGPT was really good because it was able to do this.

[00:14:42] Nate Elliott: And as Grace says, it's this double-edged sword, but it's, it's fascinating to watch them try to play the game of saying, "We're really concerned about safety, but also our tools are strong enough to break out of their safety environments as well."

[00:14:54] Grace Harmon: I think Nate brought up a really good point around Anthropic earlier this year saying that they have a model that is so powerful that they [00:15:00] can't release it.

[00:15:00] Grace Harmon: And what Sam Altman had said is that it is really good marketing to say, "We've built a bomb. We are about to drop it on you, and we will sell you a bomb shelter for $100 million to run across all your stuff, but only if we pick you as a customer." So there's that accusation of it being fear-based marketing.

[00:15:15] Nate Elliott: Yeah.

[00:15:16] Grace Harmon: And again, so double-edged sword, you know. Are consumers gonna have the perception of these [00:15:20] companies are taking all of these extra steps to protect them, or is the fact that they believe that their technology can be that dangerous scary enough that they're not gonna wanna touch it?

[00:15:29] Marcus Johnson: Mm-hmm.

[00:15:29] Nate Elliott: I think that quote was a bomb

[00:15:32] Marcus Johnson: It certainly was.

[00:15:33] Marcus Johnson: Okay, so we talked about, that's the kind of awareness part, you know, how many people are aware of it, and can you get them to at least start using it, [00:15:40] uh, or at least to try it. But Nate, we also talk about trial to usage, and you point out that Gemini is doing the best job at the moment of getting from trial to monthly usage.

[00:15:51] Marcus Johnson: Why?

[00:15:52] Nate Elliott: That's a really good question, and I don't know that we have a really good answer to it right now.

[00:15:56] Marcus Johnson: Hmm, that's fair.

[00:15:56] Nate Elliott: I, I think Google's distribution and the fact that it [00:16:00] has a chance to remind people about Gemini on a really frequent basis certainly helps. I think that not just the distribution Google has through all the other apps people use, but the integration that Gemini can offer with those apps.

[00:16:15] Nate Elliott: I mean, you click one or two buttons and personal intelligence connects Gemini to [00:16:20] your Gmail, your Drive, your Photos, to everything that you have on Google. That makes the tool more useful, and even if it isn't as good at meeting some of the benchmarks, even if Gemini was the 10th best AI tool in the marketplace, the 10th best AI tool that knows what's happening in your life, on your calendar, in your email, in your Drive, in [00:16:40] your Photos, in your Maps, all of that- Is a lot more useful than the best tool in the market that doesn't know any of that information.

[00:16:46] Nate Elliott: And so I, I think there's a real argument here that a lot of the reason Google is doing a good job of converting trial to usage, and also converting monthly usage into weekly and daily usage, is that they have a chance to remind people. But [00:17:00] also when people use it, it knows more about them, and that makes it more useful to consumers.

[00:17:05] Marcus Johnson: Grace, any theories here about why Gemini is doing such a good job at getting people from trial to, to usage?

[00:17:10] Grace Harmon: I think like Nate said, just the number of touchpoints that it has, like this ecosystem advantage that it has so many different chances to turn, um, awareness into experimentation, into habit. And then [00:17:20] it's also pretty aggressively offering paid access, which I think helps a lot.

[00:17:23] Marcus Johnson: We keep going down and you've gone from, um, awareness to trial, trial to usage, usage to give me some money for that usage. Nate, why are, as you point out in the research, Claude and ChatGPT better at getting users to pay and converting them to subscribers?

[00:17:38] Nate Elliott: They're number one and number two [00:17:40] in the market at conversion from daily usage into subscription, but I don't wanna put them in the same breath because Claude is- Mm

[00:17:46] Nate Elliott: way out in front on this. About three quarters of people who use Claude every day pay for that privilege, and only about half of the people who use ChatGPT every day are subscribers- Mm ... to ChatGPT. So, um, they are number one and number two. [00:18:00] It's a really big difference between them. And I mean, Claude has a lot going for it, but I wanna throw this to Grace because I know Grace, you are a Claude user, and I wanna hear your take on

[00:18:10] Nate Elliott: I, I also use Claude, but, but I think you might use Claude exclusively or almost exclusively, and I wanna hear your take on why that is and why you think they're so good at converting to pay.

[00:18:19] Grace Harmon: [00:18:20] I only use Claude, and I only use Claude for personal reasons. But I ... And, and that's just having to do with, um, how I perceive both companies' efforts within the AI space and which one I've chosen to use.

[00:18:31] Grace Harmon: Because I think that, that going back to, you know, OpenAI slowing down its development, I think that different people are gonna have a different view of how both companies have handled all of the events, [00:18:40] especially the last six to nine months in different aspects. We've talked about the DoD and things like that.

[00:18:44] Grace Harmon: I think a lot more of my writing as of recent has been about Claude in the enterprise space. And for that, you know, as y- Nate and I talked about, it's still the number one tool used by marketers. And I think part of the reason that that conversion into paid users is going so well is that, um, Social [00:19:00] Media Examiner found that only 50% of global marketers use free A- AI tools.

[00:19:04] Grace Harmon: So once these enterprise customers are finding what works best for them and what they're getting the most use out of, they're willing to pay for that. So I think that has a lot to do with why Claude is doing so good, um, with enterprise customers and with marketers.

[00:19:16] Marcus Johnson: If we look at the amount of awareness of all these [00:19:20] services and then go straight down to the subscription part, it's quite the drop-off.

[00:19:24] Marcus Johnson: Pretty much everyone knows about these things, but very few people pay for them. Folks, I'm wondering in the future, next couple of years, let's say, or four or five years, say that 2030 marker, are we expecting the subscription part of these, uh, companies' [00:19:40] business to, to grow into 10, 20%, uh, of their earnings?

[00:19:44] Marcus Johnson: Or is it always gonna be a smaller piece and that reliance on advertising, um, gonna be a lot more significant?

[00:19:50] Nate Elliott: Subscription's going to be a surprisingly large part of the AI industry. I think a lot of people either were surprised or claimed to be surprised when OpenAI [00:20:00] introduced ads, and I think that's probably because Sam Altman had been so highly and publicly critical of advertising in AI for a number of years.

[00:20:08] Nate Elliott: But it was always an inevitability that- Mm-hmm ... there would be ads on ChatGPT. It was always an inevitability that there'd be ads in Gemini, and they're not there yet, but on the, I think, Q1 earnings call this year, [00:20:20] Google's executives acknowledged that they were starting to look at it. And so, you know, we fully expect there will be ads on Gemini by 2027 if, if not before the end of this year.

[00:20:29] Nate Elliott: It's really hard to find examples of large scale consumer tools, whether those are utilities, uh, content utilities, information utilities, that don't involve [00:20:40] advertising as a revenue stream. There are, you know, shockingly few examples of companies who rely exclusively on consumer revenues to fund what they do, and so advertising was always going to be part of this marketplace.

[00:20:53] Nate Elliott: I think it's easy to look at ChatGPT and Gemini and say this is going to be mostly advertising, [00:21:00] because most of what people do online is mostly advertising. Advertising is the vast majority of Google's revenue, certainly on the consumer side. It is basically all of Meta's revenue. Uh, it is the vast majority of revenue for almost every consumer tool online.

[00:21:16] Nate Elliott: And so it makes sense that it would be a very large source of [00:21:20] revenue for AI chatbots. But having said that, you know, this industry is getting really interesting. Um, we're seeing reports recently that Anthropic is looking for, uh, a higher IPO valuation than SpaceX got. We're seeing reporting- Mm ... that Anthropic has higher quarterly revenues right [00:21:40] now than OpenAI does OpenAI is much bigger.

[00:21:44] Nate Elliott: It's got five times more users in the US than, than Claude does, give or take. Uh, and it has both a, an enterprise, uh, as well as a consumer business, and the consumer business has both advertising and subscriptions. It's got more revenue streams [00:22:00] and something like five times more users than Claude, and yet we're seeing that Anthropic's making more money.

[00:22:06] Nate Elliott: So subscription revenue's gonna be really important here. Mm-hmm, mm-hmm. Um, and some of that, you know, is gonna come from enterprises, but some of that clearly is coming from people. And whether it's people paying to use it for their own lives, people paying to use it [00:22:20] out of their own pockets so they can do their jobs better for companies, for enterprises, whether it's small business people, individual entrepreneurs paying- Mm

[00:22:28] Nate Elliott: to use it to help them with their LLCs or their consulting businesses, um, clearly people are willing to pay for Claude at this point. And so I don't know that I have an answer [00:22:40] to how much of the revenue for these companies is gonna be advertising versus subscriptions, but subscription, I mean, as of right now, subscription is a lot more important to these companies than I would have said 6 or 12 months ago.

[00:22:55] Marcus Johnson: Unfortunately, folks, that's where we have to leave it. Um, but, uh, Nate's full report is [00:23:00] called, uh, Which AI Platforms Are Winning in 2026? How Top Gen AI Tools Stack Up on Awareness, Trial, Usage, and Subscription. Uh, link in the show notes, or you can head to, uh, EMARKETER.com if you're a PRO+ subscriber.

[00:23:11] Marcus Johnson: That's all we've got time for, for today's episode. Thank you so much to my guests. Thank you first to Nate.

[00:23:15] Nate Elliott: Arrivederci.

[00:23:16] Marcus Johnson: And of course, to Grace.

[00:23:17] Grace Harmon: Thanks, guys. Nice talking with you.

[00:23:19] Marcus Johnson: My pleasure, and [00:23:20] to the production crew, who is just Lance, because apparently he's the only one who does any work. Uh, and to everyone for listening to Behind The Numbers: an EMARKETER Podcast, made possible by Awin.

[00:23:26] Marcus Johnson: Subscribe, follow, all those things. We'll be back on Wednesday. No, we won't. It's the weekend. We'll see you [00:23:40] Monday.