Episode 746 | 9 Startup Predictions for 2025
22:25 Share

Welcome to Startups for the Rest of Us. I'm Rob Walling. In this episode, I talk through my nine startup predictions for 2025. I like to take a bit of time each year to think about how I think the landscape might change in the coming year. And frankly, most of these are around 70%.

SaaS, but there are a few that aren't directly going to impact SaaS. And so I just said startup predictions in the title. And predictions are really tough. If you're actually looking ahead and trying to forecast things, no one knows what's going to happen.

So take these with a grain of salt. There's at least one of them that I kind of hope doesn't happen for the sake of all the startup founders that I am rooting for. But nonetheless, if I think there's a good chance it'll happen, I want to include it in this episode. Last episode, I went through my predictions for 2024. And you probably noticed, I think maybe I had a...

what, a 30 or 40% hit rate, 50% at best. So keep that in mind when you hear predictions from me or anyone else that at best, we're going to be as good as maybe a coin toss.

Before I dive into the predictions, if you missed backing the Kickstarter for my new book, co-written with Dr. Sherry Walling, it's called Exit Strategy, The Entrepreneur's Guide to Selling Your Business Without Regret. You can now buy it on pre-order in paperback form. So the Kickstarter was for hardcover, but if you go to exitstrategybook.com, you can either order an electronic copy or

or a paperback copy, and you'll get that in a few months once everything is ready. So if you missed it, exitstrategybook.com. And with that, let's dive in to my first prediction. ♪

I'm going to go through three of them as a group because of these three I made for 2024 and I'm just carrying them through. And I realized that's kind of cheating. So I'm kind of combining these all into one. My first is that Twitter will change hands in 2025. I think either the debt's going to be called due, someone's going to do a hostile takeover. The valuation is so low, I just think it is inevitable that Twitter will change hands here at some point.

And so I'm calling my shot like Babe Ruth and I'm saying 2025. My second prediction is that no code and low code will get unit tests and version control. So in 2024, I predicted it would be professionalized and I put that in quotes. And then I did a text expander on professionalizing and I defined what I meant by that. Realistically, the fact that no code and low code don't really have unit tests and version control at this point, as far as I know, the tools we use, the Airtables, the Bubbles, the Softers, the Zapiers,

stuff gets pretty brittle and it reminds me of how we coded in say the late 90s or the early 2000s and I do think it will evolve. So unit tests and version control coming for no code and low code and I do think there's opportunity there for startup founders. My third prediction carried over from last year is that Stripe will go public in 2025. Not much more to add to that one. So my fourth prediction is

for 2025. And the first new one is that search volume for Google organic SEO, and I'm talking about content-based keyword targeting, where you're cranking out an essay, an ultimate guide, a blog post, I think they will slide across the board by at least 15% in 2025, specifically as they give way to AI searches. Now, realistically, when I first wrote this down, I put 25% and I thought that

That might be too much, but I think there's a real chance it can be between 15% and 25% specifically due to AI searches. And the reason I've said specifically due to AI searches three times now is because Google has messed around with their search engine results

for what, the past decade? In terms of making ads at the top look like organic results and then just bumping the organic results further and further down the page to the point where if you're searching on a lot of laptops, you don't even see organic results above the fold. That's not what I'm talking about. I think Google will continue to pull shenanigans with that.

But the number of times now that I do searches in Google and the AI answers it quite accurately and cites sources is just shocking. And so this zero-click thing

that, you know, a lot of folks like Rand Fishkin over at SparkTour have been talking about, but a lot of other folks as well. It used to be that it was a little, you know, a little box with a summary, what's the weather like, and it just had the weather at the top instead of linking you to weather.com like it would have done 10, 15 years ago. But now AI is doing that.

And now AI answers pretty complex questions. In fact, in preparation for this episode, I have a prediction later about self-driving cars, self-driving taxis. And I had remembered that people have projected how many traffic deaths can potentially be avoided around the world if everybody used self-driving cars. But I couldn't remember what the percentage was. Is it down 90? Is it 95? Is it 98?

And so I typed that into Google and the AI at the top had excellent results. I think it said between 90 and 94% of the projections and these are the different sources and blah, blah, blah. And it just summarized everything. I didn't have to click anything. So that's just one example, but I find that being more and more common. And what used to happen is I would Google it, I'd see the AI result and I would question the AI result. You know, is this legit? Because AI can hallucinate as we know.

but they are citing sources and you can see the sentence from the source. So things are going to get dicey. It's going to be interesting to watch as search engine click-throughs drop significantly

And if you look back, Ruben Gomez and I had a conversation about this three or four months ago. You can search for his name at startupswitharestofus.com to listen to that episode where we talk through organic SEO. And his company, SignWell, relies a lot on organic SEO. But you'll hear his perspective on it. It is very much facing the present reality, but also there are ways around this.

content-based keyword targeting is only one way to do SEO in Google. What are the other ways that maybe won't be impacted by AI? And this, among all the other predictions, is the one that I don't want it to happen because guess how many startups I'm invested in, I advise, or I'm rooting for that rely on content-based keyword targeting? There's a lot of them, but we do have to face the current reality when we are building our companies.

My fifth prediction is that ads in AI chat interfaces like ChatGPT or Clod or whatever other chat interface you use will become commonplace because this is the next frontier. As eyeballs move from clicking on the 10 blue links on the Google search engine result page,

it's natural that these free interfaces have to make some type of money. They have to start monetizing at some point. And so they're going to be toying around with different types of ads

And I don't know if they'll look like AdWords or not. Will they be display ads? Will they be text ads? Will they be a video pop-up? Will they be just a sentence embedded in the AI's thing of like, hey, if you like this, maybe if you searched for this, maybe you want to buy a self-driving taxi. That doesn't actually make a lot of sense, but you get the idea. It can be interwoven into the message.

There's a lot of creative approaches to this. I think some companies will overdo it. It'll be super annoying because marketers ruin everything and monetization ruins everything. But my prediction is that we will, I know we're already seeing a little bit, I think there's one company right now that's testing ads in their AI interface as I'm recording this, but it will become, dare I say, almost ubiquitous one year from now.

My sixth prediction, another, there's two more related to AI. And look, I'm sorry. I'm not like a big AI proponent or opponent. I think it is just shifting the landscape of everything. And the fact that Google's 10 blue links are being disrupted shows you what a monumental shift across the entire world

tech community and the startup community and frankly the world. It reminds you of how big of a shift it is. So my sixth prediction is that Google will see its biggest ever drop in revenue due to the transition from

10 blue links being the prominent way people search into AI ads. And I think Google's going to have a bit of time that they're going to need to figure it out. Because realistically, it's some astronomical number, and I should probably go to ChatGPT or Claude and ask this, but I think it's something like 95% of Google. I say Google, it's Alphabet, right, is the parent company. 90 to 95% of Alphabet's revenue comes basically from online advertising. And this AI

phase shift, this huge transition is going to disrupt that. And they'll figure it out eventually, I think. But I don't think it will happen without some type of drop. And Google really hasn't seen...

any major drops in revenue from what I can recall over the past, you know, 25 or whatever years. So that is my prediction number six. And I want to pipe in before my seventh prediction. As you listen to these, think about where the opportunity is going to be for bootstrap startups. And as we see shifts and new tools arise, like let's say they do have ads in AI chat interfaces.

Well, they're going to have to have an ad management tool for that that they will build, they being whoever, you know, ChatGPT, Claude, whoever. Usually those are not going to be very good at the start. So if they have an API, is there a way to build a third-party tool that can manage the ads better? This is what tiny seed company Linklo has done for LinkedIn ads. It's linklo.io.

If you are using LinkedIn's built-in ad interface, you know how painful it can be. And Linklo helps make that a lot better. So is there an equivalent for these other new ad interfaces that are going to be created? It's just one idea, one example I think of as I look through these predictions.

My seventh prediction is that the term AI will be used in fewer H1s than it is today, especially across startups, because AI will just be assumed. It will just be the norm. It will be the default. You will assume AI is in every product today.

that you log into in every SaaS app that you use as it becomes ubiquitous. I think back to maybe 2005 on the internet when as SaaS companies, they weren't called SaaS back then, they were called software, web software, web-based software, or ASP, application service providers. As those became more prominent, they literally would have

statements on their homepage. You don't need to operate a server to use our software. It's not client server. Access from any internet browser.

And also entering your credit card is safe and secure. Like all this stuff that was such a paradigm shift to get people to try to get their head around it. Well, it's like, well, I don't need a server. How do I do it? And there would literally be like FAQs of, well, you just log in here in the web browser and you use the software. And you had to explain that to people. And these days it's like, well, it's AI for this and it's your AI assistant and it's your AI that. I mean, I guess if it's an AI assistant, that'll stick around. But I do think there's going to be a new term that comes around because...

AI means too many things right now. And I almost wish the term like virtual assistant, you know, that came to mean someone who was remote, right? It's like, oh, back in the four-hour workweek days of, I don't remember when that was, 08 or 09. Virtual assistant came to mean, oh, I have someone in, you know, in the Philippines or in APAC or in just an inexpensive region. And really virtual, I don't know, I wish that term hadn't already been co-opted for that because AI means too many things.

And I think there will be terms that splinter off of this to better describe what they actually are. But all that said, my prediction is not that. I'm digressing. I am predicting that the term AI will be used in fewer H1s one year from now because it will just be assumed that AI is built into everything.

Now in the near term, I think the number of applicants to TinySeed and to Y Combinator that include the term AI will stay the same. Hopefully it doesn't go up. I mean, I think the number of companies or the percentage of companies accepted into Y Combinator that say they use some form of AI, it was like in the most recent batch was 85 or 90%. Like it was some astronomical number. TinySeed wasn't that much, but it was definitely, if I were to just ballpark it, maybe 5%.

40 or 50% in our most recent batch. I'm not sure it's that high, but you know, you get the idea, which is very much up from zero. No, we actually had, I think we had maybe 5% two years ago, three years ago, like before chat GPT was kind of one in each batch was using AI to do something.

And then, you know, it has slowly slash quickly ticked up to where we got a lot of applicants. The problem with a lot of the AI tools that applied to TinySeed is they have this super sharp growth curve and crazy high churn. And so you can't outrun that kind of churn, 15%, 20%, 25% churn.

And I just think a lot of them are catching a temporary wave. And there's opportunity there in the short term, but to build a long-term sustainable business to make it that long, most of them will not. And it's the same thing you see on X Twitter or Blue Sky, folks launching and like, ooh, AI, look at this growth. And then...

Yeah, a few months later, the hubbubs died down and you don't hear from them anymore. And then you see it for sale on a choir for a fraction of what you think it would be worth. And so there are very, very few. I'm not saying there's zero. There are very few that are, you know, quote unquote, AI rappers, chat GPT rappers that will be around for years. My eighth prediction is one I referenced earlier. It's about self-driving taxis specifically.

And the prediction is that self-driving taxis are legit. And anywhere that allows them to operate will quickly see them become the norm. I'm speaking from firsthand experience, having been to Scottsdale, Arizona. You know, you fly into Phoenix, Scottsdale's right there, butted up against it.

And I saw ads for Waymo self-driving taxis in the airport. But of course, I have Lyft and Uber. And so I used one of those apps to get to the hotel. But then a friend of mine who lives in Phoenix said, while you're here, you have to try the Waymos. He said, it's all I take now. So download the app and try it. And I did. And Sherry and I rode around in a car with no driver. And the steering wheel moves fast.

And it is, it takes up like one trip. It took me one trip to get used to. It is a super interesting experience to be in there. And you have full control of the climate, so you can make it hot or colder. You have control of the music. You could feasibly pair your phone, but it was kind of a complicated thing for music. But they had playlists. Of course, it was a bunch of Christmas track from, because, you know, it was late November. But after doing that, I was struck by just how pleasant the experience was.

And the fact that self-driving cars are so, so, so much safer than human-driven cars is just incredible. And again, referencing earlier in this episode, I went to Google and asked how many traffic deaths, not just accidents, but deaths do we estimate will be reduced by these self-driving cars? And the numbers that quoted were between 90 and 94%.

So in the United States, for example, there's somewhere around 50,000 traffic-related deaths every year. And a lot of those are from people who... It's driver error. It's humans being tired, being drunk, being high,

making mistakes, unable to see in certain conditions where an autonomous car can see through the rain or whatever and humans might have difficulty or darkness. And so even if we take the low number of 90%, instead of 50,000 people dying each year, you're going to have 5,000. It's an estimate, but even if they're off by a factor of two, it's just shocking. It's shocking. And I know there's a whole conversation, just like AI, there's a whole conversation around jobs. And I get it.

And I am totally open to having that conversation. Not going to have it in this episode. It is much like self-checkout and much like robot manufacturing. You know, instead of humans putting a rivet in something, robots doing it in a Toyota plant. There's whole societal impacts that we have to deal with. But put all that aside for this podcast talking about startups.

Self-driving taxis are a thing. And if they came here to where I live in Minneapolis, 100%, I would be willing to take them. And in fact, late nights when sometimes I'm concerned that my driver is super tired or the driver might be drunk, I would want a self-driving taxi. And here's the other thing. I have two teenagers.

And at times, you know, there's Uber for teens, where as the parent, you kind of monitor what's happening. And so they can call a car. It tells you, you know, your child called a car to this location, and they try to give you the highest rated driver, like only 4.9 and up. And, you know, they take a bunch of safety precautions. I think they have background checks or whatever. I'm guessing they get more. It's the premium drivers, you know. It's a premium service. And then you can see the car the whole time, and then you get notified when they get dropped off.

But wouldn't that be even better if it was self-driving? Like the risk of someone doing something that you don't want to happen goes away because it is a self-driving car that pulls up and they get in and it's just, it's just incredible. Now, do I wish that

Mass transit was a thing? Yes. Do I wish in more U.S. cities it was just buses and trains like in a lot of European cities that I visited? Of course. Again, it's a conversation we can have about how society, you know, how America developed around cars and on and on and on. But self-driving cars are legit. Self-driving taxis specifically are amazing. And then what I'm thinking is, you know, I have my brother lives in the Bay Area of California and he has this massive commute because traffic's so bad. And as a result, he bought a

a Tesla purely for the self-driving because he spends so much time in the car each day, like 90 minute commute each way. Think about it like that. And it allows him to be more present with other things rather than sitting there pushing a gas pedal. But when do true self-driving cars change the game, right? Not just taxis, but that my brother could drive and not even need to sit behind the wheel anymore.

because that's what Sherry and I were doing in Phoenix. We were in the back seat as the car drove itself. And non-taxis, meaning just commuter cars that people are driving day to day,

don't have that yet. And so when that clicks, how many of those are going to be sold and how many cars on the road will become self-driving? It's going to be a lot. And I don't know if 2025 is going to be the year of it, but I think regulation is a big part of what holds this back. And of course, there are also driving conditions, like all the testing is being done. It's usually on the West Coast or the Southwest. It's not done in the snow.

So it may be a very long time until we see them in rougher climates. My ninth prediction is that platform risk will intensify in 2025. There are more platforms to build on. There are larger, more monopolistic platforms that exist. The Shopify's, the Facebook's, the Reddit's. And I think bootstrap founders will face increased challenges with platform dependence.

And if you listen back to episode 735 of this very podcast, just not even 15 episodes ago, the eight levels of SaaS platform risk, where I talk about the three key factors within platform risk, replacement, customer concentration, and lead flow, you will hear that it's not just one or zero. It's not just on or off, but I have eight levels of

where I talk about almost no platform risk, relying on a commoditized platform, all the way up to aggressive platforms, few replacements, and you have high customer concentration. So what does this mean for entrepreneurs? Well, be aware of platform risk. Be aware of the types of it, the dangers of it. Know what you are getting into. It's not to say you should never build on a platform. I think step one and step two businesses can be incredible opportunities for entrepreneurs.

As long as you're aware of what you're getting into, I would not want to build a 10-year business on any of these platforms because they do not give a shit about you and they will sooner cut off your API access accidentally, say, oops, killed your business. They will build the feature directly into their platform. Oops, killed your business. They don't care. They're like Honey Badger. They do not care. So know what you're getting into. If I were starting out online,

Trying to get my first app or first couple apps, trying to quit the day job. I am all for platforms because I think platforms lend, you know, it's all the reasons the stair-step method exists. But if I were looking to build my next app to seven or eight figures, I would think long and hard about whether or not I wanted to be on a platform that could take over the whole business, could kill it. Now, realistically, the hard part is this changes, right? Because WordPress...

I feel like has always been a pretty friendly platform until this year when Matt Mullenweg slash WordPress, you know, made the decision to be very aggressive with WP Engine and take, apparently take over plugins. I mean, there's a whole, there's a whole story there that I'm not going to cover here, but realize that platform risk exists in a lot of places, but it doesn't exist everywhere in the same level. And you will want to educate yourself, know what you're getting into as we head into 2025.

So thanks so much for joining me today to hear my nine startup predictions for 2025. Some of them are positive. Some of them are downright gloomy. But I did want to share these with you as I've been thinking about them. Thanks for joining me this week and every week. This is Rob Walling signing off from episode 746.