I was driving down 101 towards Portola when five words completely derailed my train of thought.
The massive Vanta billboard said “Compliance that doesn’t SOC 2 much.“ It was witty. It was hyper-targeted. And most importantly, it has lived rent-free in my head since that day.
In a world obsessed with digital attribution, click-through rates, and algorithmic targeting, Silicon Valley’s top tech companies are still spending millions (okay, maybe hundreds of thousands) of dollars on static, physical sheets of vinyl.
But why exactly? Because not all are like Vanta’s board. Some are provocative, some are downright stupid, and most are forgettable.
The answer lies at the intersection of brutal unit economics, old-school psychological frameworks, and a masterclass in modern brand positioning.
But you do not need a multi-million dollar venture capital runway to exploit these mechanics.
As an early-stage solo founder, you can steal the exact psychological blueprint behind 101’s most iconic or sometimes idiotic, signs to dominate your niche digitally.
Does The Math Math?
The first thought that popped into my head was, how much would it cost to own one of these things?
A quick search and here is what I found.
SF billboard rates, 2024:
To put that in context, here is what Stytch’s April 2024 campaign actually looked like. They ran three highway 101 billboards, eight SF city billboards, eighty bus ads, and a hundred and thirty bus shelter ads. Stytch’s CEO Reed McGinley-Stempel wrote up the campaign in detail on Lenny’s Newsletter, but did not share the total spend.
So let’s do the math.
Stytch April 2024 campaign, estimated media spend:
Six hundred thousand dollars. For four weeks.
That is a year of runway for a solo founder paying themselves fifty thousand a year. It is a pre-seed round at most funds. It is six hundred customer interviews where you fly the customer in and buy them dinner.
But here’s the question that actually matters. Did it work?
Per Reed’s writeup: branded search lifted 25% during the campaign. Branded search clicks lifted 70% globally. The lift sustained at 10–15% after the boards came down. More than 10% of sales opportunities created since the campaign cited the billboards as a factor.
Now let us try to explain this in terms of the ROI.
Estimated ROI of Stytch’s $600K campaign:
Break-even on direct attribution. Plus the 70% global branded search lift, the team morale, and the brand presence that fed the rest of the funnel for years.
So the math works for Stytch. But for most of us, it doesn’t.
The good, the provocative, and the stupid
Not every billboard works. Some work for the wrong reasons. Some don’t work at all. Here’s the rough taxonomy I’ve landed on.
The good ones say one specific thing in the language of one specific person.
Vanta is the textbook example. “Compliance that doesn’t SOC 2 much.” or “Don’t SOC block your engineers.”. Someone should give their marketing department a raise. But the point is that Vanta, in 5-6 words targets their exact audience and makes them laugh. They exclude everyone who is not a buyer, and the people who are buyers feel seen. Per Sarah Scharf, Vanta’s Senior Director of Demand Generation, the campaign helped Vanta land their largest enterprise customer at the time, and the billboard went viral on Reddit (which, if you know Reddit, is a small miracle).
Bland AI’s billboard, paired with a clever video, racked up 25 million views and hundreds of qualified leads, per CEO Michael Burke.
These are real numbers. This is a real channel.[1]
The provocative ones also work, but the tax is steep.
Artisan put up Stop Hiring Humans across San Francisco in late 2024. Per the company, the campaign generated $2M in new ARR, and they raised a $25M Series A from Glade Brook Capital not long after. CEO Jaspar Carmichael-Jack called the campaign “deliberate ragebait” on the record. The same campaign generated thousands of death threats by Carmichael-Jack’s own count, a satirical counter-campaign called Replacement.AI, real protest billboards from a Las Vegas virtual receptionist company, and a city full of people who openly hate them.
It worked but it came with a price.[2]
The genuinely stupid ones don’t work at all.
Vercel’s first big SF billboard was, I am not making this up, a single line of terminal code: ~/ npm i ai. The SF Standard called it the most unintelligible billboard of the bunch. Vercel’s CMO Morgane Palomares had to explain in an interview that the campaign was “about the front-end cloud.”
If your CMO has to translate the billboard, the billboard didn’t work.
Then there is AI runs better on US. Four words. Tells you nothing. Could be hosting, could be a chip company, could be a recruiting platform.[3]
The good billboards say one specific thing in the language of one specific person. The bad ones shout their LinkedIn bio.
Psychology Behind Billboards
Amongst the many Series B startup billboards, I also saw ones from Apple, OpenAI, Figma, Canva. And I thought why on earth would Apple need to show people the selfies they can take on iPhone 17’s camera? Or why would OpenAI feel the need to show an image of people on a ranch?
Everyone has heard of Apple. Everyone has heard of OpenAI. Why are these brands still spending on billboards? What are they paying for?
Then, I remembered a marketing concept I had read in How Brands Grow by Professor Byron Sharp at the Ehrenberg-Bass Institute for Marketing Science.
He says that:
brand awareness and mental availability are not the same thing.
Awareness is whether you’ve heard of a brand. Mental availability is whether the brand pops into your head in the exact moment you need it.
For example, almost everyone has heard of Yahoo. The awareness number is enormous. But when you need to look something up, you don’t say “I’ll Yahoo it.” You say “I’ll Google it.” Yahoo has the awareness, but it lost the moment.
Apple advertises to keep winning the moment you think I need a new phone. OpenAI advertises to keep winning the moment you think I have a question. Awareness is a one-time achievement. Mental availability is a renewable resource that decays. You have to keep refreshing it.
That is what the billboard buys.[7]
The Billboard Test
I will never have forty thousand dollars a month for a billboard. Probably neither will you. But the mechanics of mental availability are not gated by a billboard budget. They are gated by clarity, repetition, and discipline. And all of it is free.
So here’s the framework I now run on every piece of content I publish, every homepage line I write, every line I use to introduce Table For One at a dinner. I call it the Billboard Test.
Three questions. If your line passes all three, you are building mental availability. If it fails any one, you are making noise.
1. Is it specific?
Could a stranger glance at your line for two seconds and tell what you do and who it is for? Compliance that doesn’t SOC 2 much passes. Powerful AI fails. AI runs better on US fails. The specificity is the magnet.
The cost of vagueness is being forgotten too easily.
2. Is it in their language?
The Vanta line is only funny if you know what SOC 2 is. The people Vanta wants to reach know what SOC 2 is. The people who don’t are not the buyer. That is the feature, not the bug.
If your line uses the words your customer already uses when they describe their problem, you pass. If it uses words your investor uses, you fail.
3. Is it repeated?
Once is a stunt. A campaign is a presence. Stytch runs a major OOH campaign every two years, per Reed. They are not buying awareness. They are buying refresh. Mental availability decays. You have to keep showing up.
For a solo founder, that means: pick the line, the angle, the wedge — and say it every week. In your newsletter, your tweets, your homepage, your pitch. The same idea, repeated until it sticks. Most founders quit on a good line three months too early.
That’s the test. Specific, in their language, repeated. It is what every good 101 billboard does. It is what no AI homepage with a starfield background does.
You can run the test on any piece of writing in thirty seconds. Try it on your homepage tonight.
Sources cited in this piece
Rachitsky, Lenny and Reed McGinley-Stempel. “When and how to run a billboard campaign.” Lenny’s Newsletter, June 2024. (Stytch campaign details, lift figures, Vanta and Bland insights.)
SFGate “SF Tech CEO’s billboards are ‘dystopian.’ That’s how he wants it”, December 2024. (Backlash, death threats, Replacement.AI counter-campaign.)
SF Standard. “Stumped by San Francisco tech billboards? You’re not alone.” September 2023. (Vercel
npm i aibillboard; Vanta SOC 2 quote.)AdQuick. SF out-of-home market data. (Highway 101 and city billboard rate ranges.)
AdSemble. “San Francisco CA Digital Billboard.” (SFO digital billboard rates and impression data.)
Artisan Series A announcement, 2024. (Glade Brook Capital, $25M raise, $2M ARR claim.)
Sharp, Byron. How Brands Grow: What Marketers Don’t Know. Oxford University Press, 2010. (Mental availability concept.)
Romaniuk, Jenni and Byron Sharp. How Brands Grow Part 2. Oxford University Press, 2015. (Category Entry Points and mental availability extension.)
Covered a similar topic recently and the stat that blew my mind: SF Chron finding that 50% of SF billboards are AI products.
And to your point, many of them are VERY early stage. Super interesting.
Lovely read! 😇 The fact that a solo founder can incorporate the concept into day-to-day interactions is just awesome.