a16z is right that marquee proof wins categories. Here’s the contract clause that decides whether you ever get to use it.
Years ago, we helped a payments startup get ready for its high-stakes launch. The startup had signed a pilot with a household-name national retailer, the kind of logo that makes a Series B deck sell itself. We built the launch around one of the industry’s biggest conferences: analyst briefings, a full media plan, weeks of preparation, onsite meetings with opinion leaders. The works that any typical high-stakes launch requires.
Then we secured interest from one of the most enviable business publications to do an exclusive on the news. The reporter made no bones about what they wanted and that was the exclusive to write the story before anyone else. Often this situation is nirvana for many a tech startup.
One condition held the whole thing up. The client had promised that the pilot customer would provide a quote and take the interview call from the reporter. We planned around that promise. When the moment came to schedule the exclusive, the customer reference didn’t materialize as the CEO and co-founders had promised. The retailer wasn’t ready to go on the record. The exclusive slipped away. We still managed to go to plan B and secure a lot of embargoed interviews with some solid coverage. But the client still expected to get coverage from the national business publication despite knowingly not producing the customer reference, quote or advance interview.
The customer pilot was indeed real. It simply wasn’t ready to turn on the light and shine.
For this particular launch, the mistake was certainly not the reporter’s. For that matter, it wasn’t really the customer’s either. The proof existed. What was missing was the right to use the proof when the startup wanted. That right is something you secure long before launch week or not at all. And sometimes unexpected situations unfold. This situation, however, is not uncommon at all. It’s a carefully choreographed act where all the stars need to align and timing is everything. What’s also important to note is that the pilot customer holds the leverage, with the startup at their mercy around the timing. The startup’s ‘launch date’ is immaterial to the realities of what’s unfolding with the pilot customer. Unfortunately, many times there are too many unforeseen variables that can interfere.
Lighthouse or Landgrab, and the assumption underneath
I was reminded of the whole episode recently after reading a really sharp piece from an Andreessen Horowitz blog. In late July, a16z’s Joe Schmidt and Julian Marx published “Lighthouse or Landgrab? How to Pick Your AI Sales Strategy,” and it’s one of the better go-to-market frameworks I’ve read this year.
Their argument is clean. AI companies selling into the enterprise face a choice between two plays. The Lighthouse: win a few marquee customers in a brand new category and let their credibility (proof) pull the rest of the market in. Think Harvey in legal or Hebbia in finance. The other play is called the Landgrab: skip the prestige race, win on math (and speed), and close on ROI before an incumbent bolts AI onto its own product to compete. Think Decagon or Stuut. And the way you choose between the two plays is with a pair of key questions. How much does the buyer risk when they sign? And does social proof travel in that market?
One line in the piece is memorable. Your buyer doesn’t purchase the future. They purchase proof or they purchase math. That’s it.
That risk question is the one founders tend to underrate. Risk runs high in regulated industries, when you’re replacing a system of record, or when the output faces outside scrutiny, say, around compliance. On the other hand, risk runs low when a wrong call only annoys someone, say a manager, internally. The higher the stakes, the more a buyer needs to watch someone else who is credible in the industry go first. That’s the appeal of the Lighthouse play. It’s a bit of a trap however.
I agree with almost all of it. One quiet assumption sits under the Lighthouse half, and it’s the one that sank my client’s launch.
Proof does not travel on its own because it can’t
The framework treats “does your proof travel?” as a fixed property of your market, as if some industries are simply built for word of mouth and others aren’t. In my experience, proof doesn’t travel on its own in any market, and that’s simply because it can’t. Proof travels because someone has to make it travel. It sits in the dark static because no one did the work, or because no one got the thumbs up to turn the light on in the first place.
And even inside the markets where proof is supposed to travel freely, a hidden variable decides how far that proof actually gets. I think of it as a Name-ability Spectrum. How nameable is your proof allowed to be?
The Name-ability Spectrum
Picture three industries sitting at three different points on that spectrum.
Legal is the lighthouse shining brightly in the open. Firms will go on the record as a reference, because using the most sophisticated tool is itself a status signal. Harvey’s marquee wins get talked about precisely because the buyers want to be seen making them. The light points outward, on purpose.
Pharma is the lighthouse behind frosted glass. Almost no one announces early, and plenty never announce at all, ever. When proof does move, it goes through gated channels first: key opinion leaders, scientific sessions, published validation, people carrying the story as they move between companies. Then the moment one major player is known to be “in,” the herd moves fast, because they all want the choice someone else already de-risked for them.
Oil and gas, defense, banking, insurance, utilities. Here the bulb is barely lit. These buyers are so risk-averse and so disclosure-averse that even after they adopt, they often won’t let you name them at all. The proof exists yet still can’t move, because the reference itself is boarded up. Consider this group a lighthouse with a 5-watt bulb.
Here’s the rule that falls out of it. The less nameable your proof, the more the communications craft is worth. When you can’t get a named, on-the-record customer, the whole job shifts to the people who can make an unnameable win still land in the market.
A staircase, not a fork
This is also why I push back when a founder treats Lighthouse and Landgrab as a fork in the road. In practice it’s often a staircase. Win and document two or three real marquee deployments, use them to earn category trust, then convert the proven workflow into a broader landgrab. a16z’s own example in the post, Affirm moving from a Casper pilot to nearly every mattress company and onward, is exactly that sequence.
But you can’t climb a stair if you’re not allowed to stand on it. Which brings me to two moves most founders learn the hard way, the way my client did.
If you choose the lighthouse, price the first deal as a loss leader
Once a company decides that proof is the route into the market, the pricing follows. Treat your first marquee deal as a loss leader, on purpose. The return isn’t the contract value. It’s the reference asset that the rest of the market will react to. Go in expecting to give ground, on price, on a richer pilot, on extra support, because what you’re looking to get is a proof point you could not otherwise have as you move forward to scale.
If you are eyeing a whole category the way Affirm eyed mattress retailers, you walk into that anchor deal, look to strike a loss-leader arrangement, land the proof, and let it help you scale. The margin you’ll concede is arguably the cheapest customer acquisition you’ll secure, because you’re not acquiring one customer. You are acquiring the credibility that opens the category.
Put customer reference clauses in every contract, and hold the line
The loss leader is situational. This next move isn’t. It belongs in every contract you’ll sign, marquee or not. For years we advised our clients to introduce customer reference and co-marketing clauses straight into the agreement: the right to a case study, a testimonial, an approved quote, a customer success announcement, and select references with key opinion leaders who can wield influence. And then to stand firm on or negotiate on that language when the customer’s legal team pushes back, because they will.
The reason to lock it at signing is leverage, and timing is everything. The only moment you can win naming rights is before you sign, while the customer still wants your terms. Once you’ve delivered, you’re asking a favor, and favors collapse at exactly the wrong moment. Remember our old client’s launch?
Spell the rights out. The right to name the customer and use their logo, a jointly approved case study within a mutually agreeable window so it can’t be stalled indefinitely, an approved quote and testimonial, a set number of reference calls each quarter, some select interviews with key opinion leaders, and an announcement around the win, deployment or partnership. Build the customer’s approval into every element, so their legal and communications teams can comfortably say yes. You win this negotiation through making it easy for them, not through muscling anyone.
One detail quietly undoes all of it if you miss it. Your standard confidentiality terms will swallow every one of those rights unless you carve them out up front. Reference and marketing rights have to live as an explicit exception to the NDA, not a hopeful line buried elsewhere in the file. Work the exact language with your counsel.
And because not every customer sits at the open end of the Name-ability Spectrum, write the clause to ladder down. If you can’t get a named reference, negotiate the next rung: an anonymized but specific one, a “top-three US retailer” you are cleared to point to. Then analyst reference calls. Then a metrics-only result. Then a quote attributed to a title instead of a name. You negotiate as high up the spectrum as the category allows, and you never walk away with nothing you can use.
A lighthouse only counts if someone can see it
Had those rights been in my client’s contract, and had we held firm on the language, the reporter would have gotten the advance interview, the exclusive would have run, and the client’s offering deployed at a national retailer would have been national business news.
That’s the part left in the dark, and it decides whether a lighthouse strategy ever lights anything at all. Winning the marquee customer is the headline every founder chases. Earning the right to say their name publicly, and then doing the work to make that marquee name (the proof) travel, is the whole game.
A lighthouse only counts if someone can see it. Build the switch into the contract, and just make sure you’re the one holding it.
