Distribution Is the Product

When anyone can build a working product in weeks, the product stops being the hard part. The scarce asset is a reliable way to reach the people who need it, and that route deserves the same design care as the software. Plan distribution from the first day instead of starting it after launch.

Building Is Cheap Now

AI-assisted development, managed infrastructure, and open-source components have made software far cheaper to build. By early 2025, about a quarter of Y Combinator's Winter 2025 startups had codebases that were 95% AI-generated. That helps you, and it helps every competitor just as much. If your product can be built quickly, assume a lookalike can be too.

What competitors cannot copy quickly is your relationship with customers: an audience that trusts you, or a sales channel you understand better than anyone else. Both take months or years to build.

This changes how you spend your early months. A founder who ships a decent product and talks to customers daily usually beats one who ships a great product and waits to be found. In Traction, Gabriel Weinberg and Justin Mares propose a "50% rule": spend half your time on product and half on getting customers, running both in parallel from the start.

Founder-Led Sales

Your first customers should come from you personally. Nobody else knows the problem as well, and nobody else can change the product the same week a buyer raises an objection. Paul Graham's 2013 essay "Do Things That Don't Scale" makes the case: "You can't wait for users to come to you. You have to go out and get them." His examples include Stripe's founders setting up new users on the spot ("Right then, give me your laptop") and Airbnb's founders going door to door to recruit hosts in New York.

Every sales call is also research. You learn the words customers use, the objections that recur, and who actually signs off. Write down the exact phrases buyers use; they become the copy on your landing page and in your outreach.

A simple rhythm:

• Keep a list of everyone who fits your target customer.
• Reach out personally, with a specific reason for writing to each person.
• Use calls to learn first and sell second.
• Track every conversation and its outcome in one place.
• Review each week which objections repeat, and change the product or the pitch.

Hire salespeople only after you have closed deals yourself and can write down a process someone else can repeat.

Find the Channel That Works

There are many ways to reach customers: outbound sales, search, paid ads, content, communities, partnerships, marketplaces, events, referrals. Spread effort across all of them and you get weak results everywhere. Traction catalogs nineteen traction channels and a five-step process it calls Bullseye: brainstorm, rank, prioritize, test, and focus on what works. The authors quote Peter Thiel on why: "It is very likely that one channel is optimal. Most businesses actually get zero distribution channels to work." Test a few channels cheaply, then go deep on the one that works.

Choose channels where your customers already spend their attention and that match how they like to buy. A developer tool and a product for dental clinics will not share a channel. Measure the cost and time per real customer in each test, then commit to the winner long enough to get good at it.

Going deep means learning the channel's craft: which headlines rank in search, which messages get replies in outbound, what a partner needs before recommending you. Skill in one channel compounds over time, while dabbling in ten teaches you little about any of them.

Content and Communities

Content works when it helps your customer even if they never buy. Write about the problem you solve, with detail only someone close to it would know. Show your work, including what failed. Building in public turns your progress into distribution, as long as each post teaches readers something instead of only announcing news.

Content is slow. It rarely produces results in the first few months, and then it can keep working for years. A 2025 Ahrefs study found only 1.74% of newly published pages reached Google's top ten within a year, and the average page ranking first was five years old. Pick a format you can sustain, such as one solid article or video a week, over a burst you will abandon.

Communities and partnerships borrow trust that already exists. Join the places where your customers gather, and contribute before you promote. Look for companies that serve the same customer with a different product, and find integrations, referrals, or co-marketing that help both sides. One good partnership can outperform months of cold outreach, provided the partner gets clear value too.

Brand, Design, and Motion

Brand is how quickly a stranger decides whether to trust you. A clear name, a sharp one-line description, a consistent visual identity, and a site that explains the product without jargon all lower the cost of every other channel. Good design is a distribution multiplier: it makes ads cheaper, outreach more credible, and referrals easier to pass along.

How you sell should follow from what you sell. The two main motions:

	Product-led growth	Sales-led growth
Best when	Users can get value alone, quickly	The buyer is a committee or the setup is complex
Price point	Lower, often self-serve	Higher, often annual contracts
First investment	Onboarding, activation, in-product sharing	Founder sales, demos, pilots
Main risk	Many signups, few paying	Long cycles, slow learning

Atlassian shows how far the product-led motion can go. When it filed to go public in 2015, Tomasz Tunguz calculated that it had spent between 12% and 21% of revenue on customer acquisition over the previous three years, against 50–100% for the median SaaS company, relying on the product to attract users and move them into higher-paying tiers.

Many companies end up running both: users adopt the product on their own, and a sales team turns that usage into larger contracts. Start with the motion that fits your first customer, and add the other once the first works.

Being Found by AI Assistants

A growing share of people ask AI assistants for recommendations, comparisons, and how-to help. In a 2026 Pew survey, 49% of U.S. adults said they use AI chatbots, and 42% use them to search for information. AI answers also change what people do on search pages: in Pew's 2025 browsing study, Google users clicked a traditional result in 8% of visits when an AI summary appeared, versus 15% without one. Agents are starting to act on users' behalf by calling tools and services directly. The channel is young. How assistants choose what to mention is poorly understood and shifts as the systems change, so anyone promising guaranteed placement is guessing.

What you can do is make your product easy for machines to understand and use, which tends to help human readers as well:

• Clear, public documentation that states what the product does, who it is for, pricing, and limits, in plain text rather than only images or video.
• Structured data on your site, using the schema.org vocabulary in JSON-LD, the format Google recommends, so key facts are machine-readable. It helps machines classify your pages but guarantees no particular placement.
• Honest comparison and use-case pages that answer the questions buyers actually ask.
• A well-documented API and, where it makes sense, support for open protocols such as the Model Context Protocol, an open standard that assistants including Claude and ChatGPT use to connect to outside tools and data.
• Consistent facts everywhere, so your site, listings, and docs never contradict each other.

Treat this as a supplement to your main channels. The same basics apply: say plainly what you do, and make the next step easy.