Last week I posted a line that annoyed a lot of engineers. Anthropic has the best model, and OpenAI wins anyway. Fable 5 is, to me, the best model on earth right now, and the alternatives feel like a joke next to it. Anthropic made one focused bet on code and it was the right one. OpenAI had spread itself thin across agents, images and writing, lost the lead for a few months, then noticed, went all in on code and agents, and pulled back in front. The post did 352 reactions and 157 comments, and almost all of them argued about which lab is really ahead. That is the fun argument. It is not the one that pays your rent.
The one that pays your rent is what that story does to everyone watching it. The best product in a market does not automatically win the market. The company that gets in front of the right buyers again and again, with a focused motion, wins it. I run Growth Cab, a go to market advisory, and founders hire us to fix how they sell software above fifty thousand dollars a year. So I spend most of my week staring at the exact gap the OpenAI story is about, and nowhere is that gap wider than in lead generation for saas companies.
in
“Anthropic has the best model. OpenAI wins anyway.”
Why Lead Generation for SaaS Companies Stalls
Almost every SaaS founder I meet has quietly built their plan on the same assumption Anthropic seems to be testing in public. Build the best thing and the market will find it. So they pour a year into the product, ship something genuinely good, turn on a signup form, and wait. For a few weeks a trickle of inbound arrives from their own network. Then it flattens. The product kept improving and the pipeline did not, because a better product does not create demand on its own. It only converts demand that already exists.
That is the whole trap. Lead generation for saas companies is not a marketing chore you bolt on once the product is ready. It is the engine that decides whether anyone is ever in the room to judge how good the product is. OpenAI understood this even while it was behind on raw capability. It kept its name in front of everyone through launches, integrations and sheer presence, so that the moment its models caught up there was already an audience paying attention. A quieter, better lab does not get that second look for free, and neither does your software.
What Lead Generation for SaaS Companies Actually Requires
Strip away the tooling and a working engine has four parts, and none of them are exotic. A narrow definition of who you sell to. A channel you actually own rather than rent. A motion you run every week whether or not you feel like it. And a number you check that tells you the truth. Most founders have none of the four and a subscription to six tools that quietly assume all four already exist.
Start with the narrow definition. When we take on a SaaS client, the first thing we usually cut is the target list. A founder selling to everyone with a pulse gets ignored by everyone with a budget. Anthropic did the opposite of a broad play and it worked. It picked code and became the obvious choice for one crowd before trying to be the choice for every crowd. Your lead generation gets sharper the moment you can name the two hundred accounts that matter and forget the rest.
Then pick one channel and own it. For most of our clients that is founder led outbound stitched to a real presence on LinkedIn, because the founder is the one asset a competitor cannot copy and a template cannot fake. We put AI inside the research and the list building so the human spends their time on the fifteen messages that need judgement instead of the two hundred that need typing. One client went from a dead inbound form to nineteen qualified conversations in a quarter by doing nothing more exotic than that, run consistently and measured every week.
That last part, measured every week, is where most engines die. If the only number you look at is signups you are flying blind, because a signup tells you nothing about whether the right person is in the funnel. We count conversations started with someone who fits the profile and could actually buy. It is a smaller and uglier number than traffic, and it is the only one that moves revenue.
Where Lead Generation for SaaS Companies Stops Working
I owe you the limits, because the people selling lead gen services never mention them. The first is product truth. If your software does not solve a real problem for the people you are targeting, better lead generation just gets you rejected faster and by more qualified people. Distribution amplifies whatever sits underneath it. Point a great engine at a weak product and you have built an efficient way to hear no.
The second limit is patience. A founder led motion compounds, which is a polite way of saying it is slow at the start. The first month feels like shouting into a well. The conversations that turn into deals in month four came from messages sent in month one, and if you kill the motion at week six because the dashboard is still quiet, you never see the payoff. OpenAI did not rebuild its lead in a fortnight either.
The third is mistaking activity for progress. It is easy to feel productive sending three hundred generic messages a week. That is not lead generation, it is noise with your name on it, and it burns the exact audience you are trying to warm. A handful of sharp touches aimed at the right two hundred accounts beats a spray every time, and it protects the one reputation you cannot buy back.
How to Start This Week
Do not buy another tool. Open a document and write down the profile of your best current customer in one honest paragraph, then list twenty accounts that match it. Pick the one channel where those people actually pay attention, most likely LinkedIn or email, and commit to reaching ten of them by hand every working day with something specific rather than a template. Track one number, conversations started with a real fit, and look at it once a week.
That is unglamorous on purpose. The lesson from watching the best model lose and the loudest company win is not that quality has stopped mattering. It is that quality only gets rewarded once distribution puts it in front of someone who can say yes. Anthropic may still win the model race. It will only win the business if it fixes the part that has nothing to do with the model. Your SaaS is no different, and the fix starts with the twenty accounts you write down today.
Every Thursday, AI Frontier gives you one signal, my read on it, and one practical play from the AI and GTM systems we run inside Growth Cab, all in under five minutes, including the prompts behind our outreach. Drop your email below and confirm your subscription to get the next edition. And if you think I am wrong about where lead generation for saas companies breaks, tell me on LinkedIn. I answer everything.
