Most buyers decide before they ever talk to you

The short version:

• Most B2B buyers finish the bulk of their decision before they ever contact a vendor. Roughly 70% of the journey is done first, about 80% reach out to sellers themselves, and most already have a preferred vendor by that first call (6sense; Gartner).
• At any moment only about 5% of a market is ready to buy, the 95-5 rule. Competing only for that 5% is the most crowded, most expensive game there is.
• Content is the one instrument that reaches buyers during the long, invisible research phase. Unlike paid, it compounds into an asset instead of resetting to zero when the budget stops.
• AI search raises the stakes rather than ending the game. With roughly two-thirds of Google searches now ending without a click, the goal shifts from ranking to being the answer that gets cited, which is what SEO, AEO, and GEO are really all about.

Somewhere in the last decade, the moment of decision quietly moved.

It used to happen in the room, on a sales call, in a pitch, across a table. The salesperson was there at the start, because there was no other way to learn about a product. That stopped being true a while ago. By the time a B2B buyer books a demo today, the decision is mostly already made.

The numbers are stark. Gartner's research on the B2B buying journey finds that buyers spend only about 17% of a purchase meeting with any potential vendor. The rest goes to independent research and internal conversations. By the point someone raises a hand, 6sense finds roughly 70% of the journey is already behind them. Most buyers reach out to sellers themselves, and about 81% already have a preferred vendor in mind before that first conversation. The demo isn't where the sale is won. It's where a decision that already happened gets confirmed.

That changes what marketing is for. If the buyer makes up their mind in a stretch you can't see and can't attend, the job is no longer to persuade them in the room. It's to be useful in the room you're not in.

The 95% you're ignoring

There's a second, quieter fact underneath the first. At any given moment, only a small fraction of a market is ready to buy, roughly 5%, according to the 95-5 rule from Professor John Dawes of the Ehrenberg-Bass Institute. The other 95% aren't in the market yet. They will be eventually, but not today.

Most companies spend almost everything on that 5%. They over-invest in demand capture, paid search on high-intent keywords, retargeting, demo requests, the trackable things that show up cleanly on a dashboard, and under-invest in demand creation. It's a reasonable instinct, because those are the leads you can measure. But it means competing hardest for the smallest and most expensive slice of the market, at the precise moment intent is most obvious to every competitor at once.

The 95% is where preference is formed, long before anyone is ready to buy, in search results, peer conversations, communities, and podcasts. Whoever is useful to that group during the long quiet period tends to be the name they recognize and trust when they finally do enter the market. You aren't buying their attention at the moment of readiness. You're earning it well before.

Why content is the instrument that fits

Once you accept both facts, the decision happens off-stage, and most of your future buyers aren't ready yet, content stops looking like a tactic and starts looking like the only tool shaped for the problem. 

Content works in this hidden stretch because it doesn't ask the buyer to identify themselves first. It's there when they search at midnight, when a peer forwards a link, when they're three vendors deep in a comparison. Most of that happens anonymously and on the buyer's own schedule. By the time they reach out, 6sense finds around 85% have already largely defined their requirements. A salesperson can't be in all of those moments. A good page can. 

And unlike paid, content compounds. An ad stops the instant the budget does. You're renting attention. A page that answers a real question is an asset. It keeps getting found and keeps doing its work months and years after it's written, gaining authority as it goes. One is an expense that resets to zero. The other is a balance that grows. For any company where time and money are scarce, which is every early-stage startup, that difference eventually becomes a moat. 

The catch is that "content" here doesn't mean volume. It means being the best available answer to a question your buyer is already asking. That's a far higher bar than publishing often, and it's the part most teams get wrong.

Why this matters more now than it did a year ago

For a long time, being the best answer meant ranking near the top of a search page and hoping for the click. That arrangement is breaking.

Roughly two-thirds of Google searches in the US now end without a click to any website. SparkToro's 2026 data puts it near 68%, up from around 60% two years earlier. AI Overviews have cut click-through on even the top organic result, by about a third when they first appeared, and closer to 58% in Ahrefs' analysis. And a growing share of buyers now ask an assistant, ChatGPT, Gemini, Perplexity, Claude, instead of scanning ten blue links at all. The assistant reads the sources and returns a synthesized answer, frequently without sending a single visitor to any of them. 

At first this reads like a threat to content-led strategies. It's closer to the opposite. When the interface becomes an answer rather than a list, the entire game becomes being the source that answer is built from. The acronyms that have grown up around this are still contested and overlap heavily, but they're worth defining plainly, because the target audience for this shift, founders, keeps running into them:

• SEO (search engine optimization): being found and indexed by search engines. Still the foundation, because organic search still drives the majority of web traffic.
• AEO (answer engine optimization): being used inside the answer, the featured snippet, the AI Overview, the voice response a user reads or hears without clicking through.
• GEO (generative engine optimization): being cited and represented correctly by generative AI like ChatGPT, Gemini, and Claude, based on how the model understands your brand across the whole web, not one page.

The shift underneath all three is simple: visibility used to mean ranking, and now it means being the thing that gets cited. The reward for being the clearest, most trustworthy answer to a question just went up, because the machine in the middle now picks a single answer to repeat.

So the discipline hasn't so much changed as intensified. Answer real questions. Say the useful thing plainly and near the top. Structure it so a person can skim it and a model can lift it. Build enough depth around a subject that you read as the authority on it, not the author of one lucky post. None of that is new advice. It's simply no longer optional.

What makes it work

The strategy is easy to describe and hard to run, and the hard parts are where the returns hide. A few things matter more than they appear to. 

The first is that it has to be a system, not a campaign. Campaigns start and stop, and the compounding stops with them. The returns only accrue if the output is continuous, something useful shipped consistently, without an end date. The least glamorous truth in all of this is that most of the advantage comes from simply not stopping. 

The second is that intent beats reach at the bottom of the funnel. The highest-returning content is rarely the broad, top-of-funnel piece that draws the most traffic. It's the narrow, decision-stage material, comparisons, documentation, straight answers to "which of these should I choose," that meets a buyer who has already concluded they need something. Most teams over-produce awareness content and under-produce this. One buyer at the edge of a decision is worth more than ten at the start of one. 

The third is that in high-consideration categories, trust is the product. Nobody hands over something that matters, their books, their infrastructure, their data, because of a clever ad. They do it because they've read the answers, watched the same name show up consistently, and slowly concluded it knows what it's talking about. That kind of trust can't be bought in the 5% moment. It's assembled across the whole 95% stretch, in public, over time. Showing up in person, at events, communities, the offline half of the same funnel, is part of that motion, not a separate line item. 

The fourth is that being found has to connect to being chosen, or it's just traffic. The signal that someone read three comparison pages and a pricing doc is worth more than their email address, because it tells you who is in the 5% right now and why. The teams that make content pay route those signals into their sales conversations, so the person who finally raises a hand is met with relevance rather than a script.

How this looks in practice

We run a version of this at Inkle, where we build accounting and tax automation for US startups. A few things have held up in our own experience. 

Our marketing team is small, a handful of people, and it's run as a continuous operation rather than a series of campaigns. The output maps to the questions our buyers actually ask, at the stage they ask them: explainers and help docs for founders working out what they need, comparison pages and customer stories for those choosing between options, and the occasional handbook or webinar for topics worth going deep on. Alongside the writing, a couple of lightweight signals, which pages someone reads, what they download, help the sales team spend its time on the people close to a decision, with enough context to be useful rather than generic. And because trust in finance is built as much in person as online, we show up at events across San Francisco, New York, and Bengaluru, which feed the same loop: the conversations become content, and the content brings people to the conversations. 

The result isn't dramatic. It's a steady tilt in where customers come from. Today the large majority reach us on their own, and most, by the time we first speak, are already fairly settled on the choice. The first conversation is rarely a pitch. It's usually the confirmation of a decision made somewhere we never saw.

The takeaway

That's the heart of it. The buyer was always going to decide in the dark. You don't get to be in the room for that part. You only get to have already been useful by the time they walk out of it. For an early-stage startup with more time than budget, that turns out to be the most durable advantage there is, because unlike a campaign, it keeps working after you stop paying for it.

Frequently asked questions

What is intent-led (content-led) go-to-market? 

It's a GTM approach built around being useful to buyers during their own independent research, so you become the preferred option by the time they're ready, rather than interrupting them with cold outreach. It's aimed at both the anonymous research phase, where most decisions actually form, and the roughly 95% of a market that isn't in-market yet.

Does content marketing still work now that AI answers everything?

Yes, but the objective moves. With most searches ending without a click and AI Overviews cutting click-through even on the top result, ranking first matters less than being the source the AI answer is built from. Useful, well-structured, trustworthy content is exactly what search engines and AI assistants pull from, so the work shifts from chasing traffic to earning citations.

What's the difference between SEO, AEO, and GEO? 

SEO is being found and indexed by search engines. AEO (answer engine optimization) is being used inside a direct answer, featured snippets, AI Overviews, voice results. GEO (generative engine optimization) is being cited and represented correctly by generative AI like ChatGPT and Gemini. The terms overlap and are still debated. The shared goal is being the clearest, most trustworthy answer to a real question.

Should an early-stage startup do inbound or outbound first? 

Both have a place, but content-led inbound compounds while paid and outbound reset the moment you stop. A practical order for early-stage founders: start with high-intent, decision-stage content, comparison pages, docs, honest answers, that captures buyers already searching, then widen into broader educational content. Whatever you publish, connect it to sales so traffic turns into conversations.

How long does content-led GTM take to work? 

It compounds slowly, then meaningfully. It's a system, not a campaign. Most of the advantage comes from consistency and not stopping: publishing useful answers continuously so a library of pages accrues authority over months and years. Expect a gradual shift in where customers come from rather than a single spike.