How novelty effects and Dopamine Culture rule the tech industry

AI apps with shitty retention, consumers with zero patience, and what happens next

Dear readers,
First off, hello from Substack! I am writing from andrewchen.substack.com where you’ll find me from now on. After 15+ years of writing via wordpress on andrewchen.com (and its predecessors), I’m going to try Substack. I’m long overdue, of course, led the company’s Series A in 2019. I will be maintaining my archives at andrewchen.com and will occasionally refer back.

My (very bad, very lazy) excuse is that I wasn’t very active on blog anyway, being heads down on my book, and as a result, I didn’t want to undergo the big project of transitioning platforms. For now I’m going to keep andrewchen.com where it is for the SEO juice, and will just do my writing here.

Anyway, onwards. Let’s talk about “Dopamine Culture” and this provocative chart via Ted Gioia:

If you stare at this graphic for a moment, you’ll note a few things:

• Everything on the left is real/offline. Whereas everything on the right is virtual and online. This is, of course, because what is real/offline tends to take more intent and longer engagement, since you have to physically go somewhere. But it also means it’s very convenient to exist in the digital world.

• Moments are measured in seconds not days/years. In slow/traditional culture, where everything on the left takes years to produce, and often hours to consume. And of course, this also restricts the supply and thus there’s a smaller supply of books than, say, articles, and then even fewer than of tweets. So we have more, but they are fleeting.

• This is hitting everything hard. If software is eating the world, then of course the ability to instantly satisfy people is pushing into every part of their experience. The above lists sports, video, communication, and relationships — but we could say the same about shopping, travel, work, and every other aspect of life is also becoming more instant, more convenient, but also more ephemeral.

You could look at this chart and it’s a bit of a Rorschach Test on how you feel about the conversion of IRL/slow experiences into digital/interactive moments. The term “Dopamine Culture” is sort of negative in that it implies that we’re hacking our own biological systems, but let me actually argue the opposite.

This is instead, a result of everything…

• … becoming more accessible and universal. Everyone (poor or rich, educated or uneducated) has a smart phone, and we are giving people much more choice.

• … having endless variety. Many of the items in the chart above historically were controlled by central gatekeepers, be it news publishers, movie studios, telecom companies, or otherwise. Now it’s much more democratic, which lets people consume what they want — which means more niche and long tail interests. We can produce a lot more, because it’s cheaper to produce.

• … being consumed on new channels/devices/etc. Instead of needing to go to a movie theater or a museum to consume movies or art, we can do it on our handheld, mobile supercomputers. Whereas before we would have waited in line or sat on our toilets and stared at the ceiling in boredom, we can consume what we want — and sometimes those are long-form Substacks or audiobooks or they might be short form videos. OK, more often videos and feeds than books, I admit.

We could debate all of this further, and we’ll have a grand ol’ argument about whether this is all good or bad for us, but this essay isn’t meant to address all that. We’ll debate that in the cultural/political sphere for ages to come, but this essay is instead about the ramifications of this new accessibility — particularly in the world of building/launching/growing tech products.

A couple observations on how all of these trends manifest into tech:

• Shitty retention is ubiquitous

• Novelty effect is real, and affects growth metrics

• The dominant culture of product management iterative towards KPIs

• Tap into Dopamine Culture, or counterposition - both can work

Let’s break each of these down in more detail.

Shitty retention is ubiquitous
First, perhaps obvious, but no wonder retention in mobile products is terrible. I’ve often written about these benchmarks for retention:

• daily retention of consumer apps: D1/D7/D28 of 60/30/15

• the % of active users — that is, DAU/registered — should be 25%

• or, for consumer/SMB subscription: M12 of >30%

When you reflect on these metrics, they are actually outright terrible, but I assure you that these particular benchmarks are in fact the good numbers! (That is, most of what I see in pitches/data rooms/etc are much worse than this). In other words, tech products live in a world where you can lose ~90% of our DAUs after a month, and >75% of your user might be inactive. And that’s not only acceptable, but it’s actually good! I’ve written much more about these disturbing metrics here and here, and in general it’s a depressing topic. But obviously a small % of products do succeed each year.

The important part and how it fits into Dopamine Culture is that whatever value your product provides, it has to deliver it fast. If you don’t provide value in the first session and the first day, you’ll have such severe drop-off that you won’t get a second chance.

If you had “Products” as a row at the top graphic, you’d get something like this:

• Slow/traditional culture: Save up for years, buy a product, use it even if it sucks

• Dopamine culture: Download the free app, tap a few times, if it doesn’t do what it says, swipe out, go to Instagram, go back to scrolling

Good luck.

How Novelty Effect impacts our product growth curves
Dopamine culture drives novelty effects as well. I recently wrote about the S-curve of AI apps versus mobile apps, where the tech industry seems to be obsessed about the former and just meh about mobile apps, even though the latter is still a platform with a lot of innovation and a customer base of bullions. And of course, it’s because we’re on a new S-curve with AI apps, where everything is new and shiny.

In other words, there’s a Novelty Effect. In the early era of a technology S-curve, novelty means that every new interaction hits dopamine hard. Click a button, and something magical happens. Post a demo, and all your followers ooh and aah. Talk about your experiences with a new product, and people think you’re a genius.

And the Novelty Effect has some very concrete actual effects on growth:

• More organic growth / word of mouth

• Higher signup conversion

• More one-and-done usage

• More comparison shopping

• Higher churn

As many AI founders are figuring out right now, products are often finding spiky growth metrics, popping with each new model release (or shiny new feature) but ultimately with bad long-term retention. A critic might say, easy come easy go.

But to be honest, it’s a good trade. At least you’re getting new users, lots of engagement/activity, and thus a lot of investment activity as well. Better to be in a high novelty S-curve and have a chance to be successful than a tired curve where you have to pay/grind for every new customer.

Product management culture
The culture of tech product management has also moved us all towards Dopamine Culture. As many of you know, most large companies use processes like OKRs to set goals, or they are often slaves to the quarterly earnings reports. As a result, planning and thinking happen in 3 month increments rather than multi-year plans.

Product Management at most tech firms have thus aligned to quarterly review schedules where each team promises they’ll raise one metric or another by (a rather unambitious) 2-3%, and then try to do exactly that. If they fail, they blame seasonality, and if they succeed, they try to get promoted :) Jokes aside, the easiest way to deliver on these small promises is to grind out incremental progress — I’ve referred to this as the Next Feature Fallacy — particulalry at helping users get their dopamine hits. This means using algos to help users scroll, or to help them invite their friends. Or to recommend the next bit of content that will keep them engaged. Even small % gains in these core areas drive so much growth that it’s worth it.

I argue this is mostly a good thing. We are giving people what they want. And coupled with an engagement-driven business model (online ads) and access to real-time metrics/dashboards, it means we are constantly driving towards making it even more efficient over time.

Tap into Dopamine Culture, or counterposition - both can work
The final point I’ll make about Dopamine culture, if you’ve read this far this will resonate I think, which is that counterpositioning also works. Yes it’s true that dating apps have increased the amount of casual dating, but actually divorce rates are recently declining. And yes there are a lot of photos and videos and tweets, but there are also wonderful services like Substack where you can read to your heart’s content, and long-form podcasts/audiobooks/videos are also on the rise. It turns out that while Dopamine Culture grows, many times all the short form content and experiences are just the commercial for the real thing.

Perhaps after watching one short form video you might watch hours and hours of video from a creator you like. Or after you read a funny tweet from someone, you decide to subscribe to their Substack or 1-click buy their book on Amazon. Yes it’s a dopamine hit to be able to get the book onto your Kindle quickly, but it’s also an entree into a multi-hour reading experience.

On the other hand, it’s also obvious that a new product has to tap into any new emergent culture in order to succeed. In a world where visual media is the dominant force — whether that’s photos, videos, or clips — any product that figures out how to naturally produce media in its usage will succeed. No wonder we’ve seen the Novelty Effect of diffusion models that produce media, being shared on social media, become so viral so quickly.

And finally — yes, ride the Novelty Effect. Build upon Dopamine Culture. But also keep in mind that to be successful in the long-run, your product must also be useful. It must retain. Because when the S-curve progresses to the next stage, anyone who is novel but useless will surely come to an early end.