Notes on scale + quality
or why we need to build human-scale internet networks
Greetings from Anna Maria Island where I’m spending a few days with my extended fam (in total 12 adults and 9 kids). Not exactly relaxing, but in between too many candy store visits, sticky hands, and celebrating my little one’s 3rd birthday, I managed to escape the human dramedy to enjoy my own company at the lovely North Shore Cafe and write my way out of (into?) a question I’ve been thinking about: what are the trade-offs between scale and quality when it comes to building Internet businesses? Like it or not, this is the kind of stuff I think about when I’m on vacation.
Notes on scale + quality
It’s hard to have a conversation with anyone in tech without the terms “Internet scale” or "network effects" entering the conversation.
Because the laws and constraints of the physical world don't apply to the digital world, the theory goes, anything online should be huge.
In practice, things are more complicated.
Let me use Sublime as an example, since it lives rent free in my head.
On the one hand, one additional person using Sublime costs us nothing.
The more people use Sublime, the more content there will be, and the more likely you find something relevant.
On the other hand – the most valuable part of Sublime is the people.
The fact that Sublime is like a collaborative commonplace book where you save snippets and thoughts and ideas and can discover related ones from other people’s libraries is fantastic – but that's because of the quality of the network and the signal to noise ratio being very high.
This leads to a dilemma: is Sublime a software or curation business? Are we, and trust me I hate this phrase as much as you do, a “SaaS” tool, a social network, or a boutique collective?
Software encourages infinite scale and rapid growth.
Curation focuses on shared context, quality, and deliberate community growth.
The common pattern for consumer internet entrepreneurs is to scale as fast as possible, aim to convert the whole world into a user, then sell your company and go into a deep depression.
writes:“I have a theory that chasing things that scale makes you need therapy, and the therapy is pursuing things that can’t scale. I once wrote that every entrepreneur’s dream is to succeed at building an impossibly hard business and then finally open a local coffee shop to be happy.”
My friend Ben Pieratt had something similar to say about his experience building Svpply:
Damn. I don't want this to happen to Sublime.
As anything scales too effectively – from restaurants and ad agencies to social networks and search engines – the market opens for more non-scalable alternatives. Once Starbucks opens on every block, we crave the artisanal coffee shop. There's the identity piece of it, where we want some degree of distinctiveness. But there's a practical side too: now that Google can crawl every page on the web and return it in an instant, we have a different problem: an increasingly unmanageable flood of bullshit.
Here’s Bill Bernbach’s letter of resignation from Grey in 1947, warning about the hidden costs of scale (I tried to pick an excerpt but the whole thing is just too good):
Dear ________:
Our agency is getting big. That’s something to be happy about. But it’s something to worry about, too, and I don’t mind telling you I’m damned worried. I’m worried that we’re going to fall into the trap of bigness, that we’re going to worship techniques instead of substance, that we’re going to follow history instead of making it, that we’re going to be drowned by superficialities instead of buoyed up by solid fundamentals. I’m worried lest hardening of the creative arteries begin to set in.
There are a lot of great technicians in advertising. And unfortunately they talk the best game. They know all the rules. They can tell you that people in an ad will get you greater readership. They can tell you that a sentence should be this sort [sic] or that long. They can tell you that body copy should be broken up for easier reading. They can give you fact after fact after fact. They are the scientists of advertising. But there’s one little rub. Advertising is fundamentally persuasion and persuasion happens to be not a science, but an art.
It’s that creative spark that I’m so jealous of for our agency and that I am so desperately fearful of losing. I don’t want academicians. I don’t want scientists. I don’t want people who do the right things. I want people who do inspiring things.
In the past year I must have interviewed about 80 people—writers and artists. Many of them were from the so-called giants of the agency field. It was appalling to see how few of these people were genuinely creative. Sure, they had advertising know-how. Yes, they were up on advertising technique.
But look beneath the technique and what did you find? A sameness, a mental weariness, a mediocrity of ideas. But they could defend every ad on the basis that it obeyed the rules of advertising. It was like worshiping a ritual instead of the God.
All this is not to say that technique is unimportant. Superior technical skill will make a good man better. But the danger is a preoccupation with technical skill or the mistaking of technical skill for creative ability.
The danger lies in the temptation to buy routinised [sic] men who have a formula for advertising. The danger lies in the natural tendency to go after tried-and-true talent that will not make us stand out in competition but rather make us look like all the others.
If we are to advance we must emerge as a distinctive personality. We must develop our own philosophy and not have the advertising philosophy of others imposed on us.
Let us blaze new trails. Let us prove to the world that good taste, good art and good writing can be good selling.
Respectfully,
Bill Bernbach
The allure of scale
Internet logic goes: because software has zero marginal cost (selling one copy costs just as much as 100 million) and in many cases has network effects (the product gets better as more people join), there are massive profits to be gained from reaching huge numbers of people. There's also a less mercenary argument: why should the most valuable things be artificially constrained, when the beauty of the web is they can reach whomever they are relevant to?
That's different from a Hermes Birkin handbag, where each one is handcrafted and takes 25 hours to make. You can't simply walk into Hermes and ask for a Birkin bag. Hermes flips the buying process on its head and decides whether or not it wants to sell you one.
Luxury brands have their own special breed of network effects – the more people recognize and covet the Hermes brand, the greater its signaling value. But these network effects are slower moving (and in my opinion more powerful) – it took nearly two decades for the Birkin bag to gain widespread fame. It was the patient cultivation of desirability and exclusivity that allowed them to preserve their premium status and pricing power.
Artisanal software
In a speech delivered years ago, Ev Williams used agriculture as a metaphor to understand what’s possible on the Internet.
Agriculture was a tremendous invention – it got people fed and freed them to do many things. But agriculture – taken to the extreme in the pursuit of profit leads to a sick and obese population and industrialized farms with little regard for the environment or animals. And in response to that, we now have amazing artisanal and healthy food choices that prioritize quality ingredients over mass production.
If we apply this metaphor to the Internet – it’s amazing that Google, Amazon, Spotify, Instagram, and Twitter exist and bring us convenience. But taken to the extreme, they flatten our interactions and reduce them to a number of likes, they exploit creators (and our attention), they prioritize breadth over depth.
The web hasn’t had its artisanal moment. Right now, giant, ad-based networks created by six men that everyone begrudgingly uses with diminishing emotional returns, control the vast majority of the web.
Human-scale internet networks
I think there’s room for a different approach.
You can always trace a product’s flaws back to its motivations. If you think about a car or a PC – they’re not trying to get you to use the car or the PC as much as possible. The purchase is the purpose. But the predominant business model of the digital age — advertising — incentivizes consuming more, faster. And because ads need a lot of eyeballs to work, they also necessitate mega scale.
That these incentives are spoiling the Internet is perhaps our best hope to take alternatives seriously. A renewed interest in human curation, a slow move away from big social, a clearer understanding of platform incentives, the economic feasibility of subscription businesses, and builders driven by a new set of values – combined, it feels like a good time for Internet businesses that are both human-scale and profitable.
To be clear, subscriptions are not a panacea, which is why we’ve been experimenting with a choose what you pay model.
There are no solutions, only trade-offs. This is the journey we’re on with Sublime. We hope you find the opportunity as interesting as we do.
Related reading
Craft by Paul Stamatiou
Tiktok’s enshittification by Cory Doctorow
Karri Saarinen, co-founder of Linear, wrote a post about quality “As companies grow, they often have a hard time with quality, and usually just give up on it. [The] main reason is that quality is something which cannot be easily measured or defined. As the companies scale, the way they operate or make decisions [is] based more on measurements.”
In other news
Remember phone books?
Sublime is a proud sponsor - like actually, we’re proud of supporting this because it rules - of Internet Phone Book, a physical directory of poetic, personal, and human websites. Kristoffer and Elliott are two of the most interesting and genuine human beings in the cozy web universe - can’t wait to see what they put together.
Submissions are open now.
Sublimers out in the wild
Sublime member Patricia Mou has opened pre-orders for The Rabbit Holes Coffee Table Book, a compilation of her amazing
newsletter, and it looks beautiful (only 200 copies available!):
Sublime legend ’s last two essays - Slow the decay in our creative work and Smaller rooms lead to bigger doors perfectly embody Seth’s work over at of which we at Sublime are proud members.
Pairs nicely w Yancey Strickler’s latest, Small is more meaningful than big.
Itay Dreyfus asks whether screens are making us dumber.
Are you a Sublime member who’s doing cool stuff? Let us know - DM me here on Substack or email Alex at alex@sublime.app.
"The first Velvet Underground album only sold 10,000 copies, but everyone who bought it formed a band." - Brian Eno.
Also: thank you for the support in making the phone book possible <3
A phenomenal piece as always Sari
It’s interesting to ask what the levers are if you want to limit scale. We know the levers for growth. What are they for identifying a specific maxima?
The one that’s clearest to me is headcount, which is the ceiling we’re using at metalabel to maintain a certain spirit and way of being. But as I’ve told the team many times, success makes these limits harder. At kickstarter the strategy was never be bigger than 50 people, but one day we started becoming overwhelmed with interest and what to do but lean into it? It was only later that I began to see why not.
I’d say 37 Signals, Craigslist, and Arena are probably strongest in this way of thinking and living it out. It takes real discipline and a willingness to swim the other way.