The Internet Is Fun Again

tl;dr = we are in a new golden age of experimentation because AI is fun + weird + different and people are more willing to try new products than at any point in the past few years = some weird consumer apps will become wildly mainstream.

There has been a fair bit of skepticism in recent years around the viability of new consumer internet startups to produce dramatic outlier returns. It was only 2018 when Peter Thiel decried consumer internet as functionally dead – “perhaps the big ideas have been tried,” he admitted that year at Fortune’s Dealbook Conference. Last month’s reporting in The Information of VC’s Consumer to Enterprise Pivot was merely a culmination of many years of continued pessimism around consumer in an era where Enterprise SaaS companies have become easier to underwrite, scale, and – more recently – have often commanded the multiples reminiscent of viral consumer internet plays.

Naturally, any number of investors and venture philosophers have rushed to take the other side of that sentiment. Most recently, Kirsten Green authored Consumer Startups Are a Better Bet than Enterprise Startups making a historical, empirical case that consumer (and consumer adjacent) startups outperform. Others rushed to Twitter to emphasize that generative AI has recast the opportunity set for musicians, videographers, writers, editors and more.

All of these takes – irrespective of their accuracy – miss a critical nuance: if consumer internet truly has another renaissance here in 2024, it’s because the internet is undeniably fun again.

The unprecedented acceleration of generative AI, generative imaging, AI search, and AI copilots has made our digital engagement feel magical again. And the internet being magical again has broken our long standing reticence to change our habits and try new things. Simply put: after a decade of cynicism towards new experimentation, our minds are being reframed to embrace experimentation of new genAI products.?

The Data

Early empirical evidence seems to validate this assessment, with mobile app installs per active mobile user in the US re-accelerating (ex Covid) for the first time since 2017 – numbers that are even more encouraging given that genAI mania did not evolve until the back half of 2023:

15 years after the introduction of the iPhone, with mobile users habituated into their existing app workflows and preferences, it becomes incrementally harder each year to convince a user to experiment with a new product because they are generally comfortable with how the existing products help them manage their life. As a founder once guided me: “beware the life hacks category.” This is why new app downloads per US user have been declining since 2017.

The Covid surge is an interesting outlier but also easily explainable: the behavioral change during Covid was out of necessity and utility, not experimentation. And that is why, like nearly every other Covid disruption, app installs ultimately reverted to their historical trendline by late 2021 and 2022

Additional data points supporting a shift towards experimentation include a September 2023 Salesforce.com survey that found “38% of generative AI users us[ing] generative AI for fun/messing around, [with only] 34% [using] for learning about topics that interest them,” suggesting users find more entertainment/novelty value today than actual utility. Similarly, seven of the top ten AI related searches in 2023 were for experimental keywords such as “free AI,” “AI art” and “what is AI?”

Historical Cycles

Truly understanding the viability of an emerging renaissance in consumer internet requires deconstructing the most two most recent of breakout consumer companies: the 2008-2013 e-commerce boom and the 2011-2015 historic transition to mobile.

The E-Commerce Boom of the late aughts produced many of the the iconic names in today’s e-commerce landscape: Wayfair, Chewy, Zulily, Stitch Fix, Jet, Warby Parker and more. While it’s tempting to attribute this group’s success to behavioral tailwinds of increasing comfort with online shopping, I would argue their success is moreso due to the unprecedented distribution and virality advantages generated by Facebook’s growth and connectivity. The ubiquity of an increasingly personalized Facebook ads engine, cost effective targeting (especially before large players adopted the platform), and tagging of friends for relevant products enabled an extraordinary mechanism for e-commerce sellers to reach new audiences – and, made it easier than ever for those eyeballs to subsequently click to explore a new site.

The Transition to Mobile was less about advances in distribution and more about novel, magical product experiences. This was the first time the consumer truly had a life operating system in the palm of their hand, synchronized to their precise location – enabling on demand transportation, food delivery, home security, remote car starting and more. Because these experiences were, at the time, so remarkably re-imaginative and futuristic, they prompted a large quantum of new product experimentation, especially as consumers sought to structure their phone home screens to match their life workflow needs.

The common denominator in both of these cycles was the willingness to experiment – whether due to ease of accessibility/discovery or due to technological innovation. And after a decade of mostly incremental social innovations, there appears to be a gold rush of consumers finding both outsized entertainment value and outsized utility in new products. Over the past several months, we have seen outsized momentum in new areas of the internet: a new purpose built social networks for sharing generative images, AI generated songs that have generated many billions of streams on Tiktok, generative AI avatar apps that have earned 100s of millions of revenue with zero marketing, dozens of generative profiles/accounts from faux celebrities to Marge Simpson (!!) that have gained millions of followers across Tiktok and Instagram – and even new digital marketplaces that have already sprung up to enable buying and selling of these AI generated profiles.??

It is this desire to experiment with new products that is the key variable underlying a new renaissance in consumer internet. Consumer investments are difficult because of the unpredictable war for customers acquisition, attention and mindshare. Success in consumer is fundamentally not meritocratic; the quintessential example of this being the famed case study of VHS versus Betamax – where an indisputably inferior product captured the market. Financial markets are inefficient because humans are inefficient and nowhere is this more explicit than in consumer investments.

But experimentation solves this. Experimentation generates low CACs, allows for more iteration cycles, and most importantly, allows for a margin of error: users do not expect perfection because they do not know what perfection looks like. The novelty and delight of new experiences is sufficient for engagement and positive sentiment.

It is impossible to diagnose every evolution of consumer behavior. And perhaps I am wrong – and the bias towards experimentation and novelty will wane (as it did roughly a year into Covid). Nevertheless, the data we have today and the insatiable demand and engagement for many of the new consumer applications we’re meeting, appears to demonstrate that we are entering into a renaissance of consumer investing. LLMs and Generative AI have opened the floodgates of experimentation dramatically. ChatGPT is, famously, the fastest platform in history to achieve 100M users , and, similarly, hundreds of niche AI driven apps have also grown at unprecedented rates – from face generators, to image generators, to video editors/generators, to chatbots and more.??

The cynicism towards consumer investments from institutional investors may not change anytime soon. But for the experimentoors themselves, it already has.

?