What is the difference between top-down and bottom-up sales?

Top-down sales starts with a decision maker high in an organization: you get an executive's attention, prove you solve a strategic problem, clear procurement, and sign a contract. Bottom-up sales starts with the user: you ship a self-serve product individuals adopt on their own, let it spread inside a company, then convert that usage into a paid contract. Top-down suits expensive products that appeal to executives; bottom-up suits products that solve a pain for individuals or small teams and are easy to adopt.

Should my startup use top-down or bottom-up sales?

Go bottom-up if individual contributors or small teams feel the pain; go top-down if an executive does. YC's Pete Koomen notes that the top YC B2B SaaS companies are split roughly evenly between the two, so neither is inherently better. Many companies eventually run both: bottom-up to acquire individual users, then top-down to expand into larger contracts.

Why does top-down sales set a price floor for your product?

Because the only way to scale a top-down motion is to build an enterprise sales team, and that team is expensive. The cost of those salespeople has to be covered by each deal, which pushes the minimum viable price up. Koomen puts the floor at roughly 10,000 dollars for mid-market and 100,000 dollars for enterprise. Below that, the cost of selling outweighs the revenue and the unit economics break.

Does bottom-up sales still need salespeople?

Yes. The common myth is that a self-serve product sells itself, but Koomen calls the need for salespeople the dirty secret of bottom-up sales: Slack, the textbook example, employs many. The real advantage is not skipping sales, it is a more efficient sales cycle, because the prospect already uses the product before a salesperson ever calls. That efficiency lets you close faster and cheaper, then choose to pass the saving on as a lower price or keep it as margin.

What are examples of top-down and bottom-up sales companies?

Bottom-up examples include Slack and Dropbox, which individuals and small teams adopted on their own and spread inside companies before any contract was signed. Top-down examples include enterprise software like Salesforce, Workday, and SAP, which is sold to executives through a dedicated sales team and a formal procurement process. In the video, Pete Koomen uses Slack as the canonical bottom-up case.

Sales motions: top-down vs bottom-up

There are two ways to sell software to a company, two sales motions, and most founders pick the wrong one for too long. Getting this top-down vs bottom-up sales decision right early saves you from rebuilding your go-to-market later. In a Y Combinator video, Pete Koomen, a YC partner who co-founded and was CTO of Optimizely, lays out the two motions cleanly: top-down, where you sell to an executive, and bottom-up, where you win individual users first. Here is how each one works, how to tell which fits your startup, and what the current wave of AI sales tools changes about the math.

Top-down sales: start with the decision maker

A top-down motion starts with a decision maker high up in the organization. The playbook is consistent: identify a lead and get their attention, which is the hard part, validate that they have a problem you can solve and convince them you can solve it, clear the procurement hoops, sign the contract, onboard the customer, and you are in business.

This works for products that appeal to executives, usually because they move an important number in the right direction or accomplish a strategic goal. Those products are often expensive to implement and need a lot of support. The biggest advantage is that there is a playbook and it usually works. It takes grinding, but if you know who your customers are and understand their problems, you will eventually find someone who pays you. Executives are hard to reach, but they make great customers, and startups that start top-down often show better early retention than bottom-up startups.

The downsides are real. It is easy to get pulled into building one-off features for big customers, and those can quietly turn into consulting relationships. And the only way to scale a top-down motion is to build an enterprise sales team, which is expensive. That cost sets a floor price for your product: roughly 10,000 dollars for mid-market and 100,000 for enterprise. Charge less than that and the unit economics stop working.

To build the motion: define your target customer (which companies need the product, and who inside those companies has the problem), find leads that fit (LinkedIn is the workhorse here), and get their attention. Warm introductions through your network or your investors are best. If you cannot get those, cold email works, using tools like ZoomInfo, Hunter.io, or LinkedIn Sales Navigator to find contacts. Write every email by hand and make it as personal as you can.

One hoop founders underestimate is the security review. As a CISSP, I will flag that for almost any enterprise deal, procurement includes a security and vendor-risk assessment: the SOC 2 report, the questionnaire, the data-handling questions. It gates the contract as surely as price does. Have a short security one-pager ready and know your SOC 2 status before the first big deal, or you will watch a closed deal stall for a quarter in someone's review queue.

Bottom-up sales: start with the user

A bottom-up motion starts with the user. You build a self-serve product that people can adopt on their own without talking to you, find a cheap distribution channel that scales (a way to get a lot of attention without spending a lot of money), use it to get many users, then find the companies where your product is already being used and convert them to contracts in exchange for more features or bulk pricing.

This works for products that solve a pain point for individuals or small teams, are easy to adopt, and spread virally inside an organization. Slack is the textbook case: small teams adopted it without ever talking to a salesperson, it spread on its own, and the sales team could wait to reach out until adoption hit a critical mass, which made the sales cycle very efficient.

Now the reality check, because the bottom-up model looks like free money and almost never is. Viral growth is genuinely hard, and the companies that pull it off usually found some non-obvious, untapped channel to acquire their early users. The dirty secret, in Koomen's words, is that bottom-up sales still require salespeople. Slack employs plenty. The difference is that when your prospect already uses the product, the sales cycle is efficient, and you can spend that efficiency on lower prices or higher margins, whichever helps you win.

If you go bottom-up, a few things are non-negotiable. You still cannot build a great product without talking to customers, and most bottom-up startups still start with cold calling to reach early users. After that, be obsessed with friction. Show your landing page to people and ask them to explain it back to you, then rewrite it until it is crystal clear. Watch real users go through onboarding and cut anything confusing or annoying. Instrument the funnel so you can see where people drop off, and use A/B testing to fix those points. And use freemium deliberately: make the features individuals love free, and charge for the features teams need to collaborate.

So which one? Match your go-to-market motion to whose problem you solve

Neither motion is better. If you look at the top YC B2B SaaS companies, it is a roughly even split. The deciding question is simple: whose problem are you solving? If your pitch lands hardest with individual contributors or small teams, you will probably go bottom-up, the motion behind most product-led growth. If it lands hardest with executives, you will go top-down. And the two are not mutually exclusive over time. You often use a bottom-up motion to acquire individual users so you can later sell to the executives they work for.

What AI changes about the math in 2026

Koomen's framing is evergreen, but it predates the current wave of AI sales tools, and those tools change the economics of each motion more than they change the motions themselves.

On the top-down side, the hard part Koomen names is getting an executive's attention, and the thing that sets your price floor is the cost of an enterprise sales team. AI moves both. An AI SDR now does the prospect research and first-touch outreach that used to require headcount, which lets a smaller team run a credible top-down motion and pushes that price floor down. The caution is the same one Koomen gives about cold email: personalization still wins, and a blast of generic AI-written emails is the fastest way to get ignored. The YC startups building AI for sales are worth studying for how they keep the outreach specific.

On the bottom-up side, the friction obsession Koomen prescribes is exactly the work AI is good at: summarizing session recordings to find where users drop off, drafting and testing landing-page variants, and turning funnel data into the next fix. AI does not pick your motion. It lowers the cost of running whichever one your customer's problem demands.

What to do this week

Answer the one question first: is the person who feels your product's pain an individual contributor or an executive? That answer is your default motion.
If top-down, write a one-line target-customer definition (which companies, which role inside), pull 20 leads on LinkedIn, and hand-write 20 personalized emails. No templates.
If bottom-up, show your landing page to five people and have them explain it back to you. Rewrite until they get it in one read, then watch five people go through onboarding.
Prepare the enterprise gate early. Draft a one-page security overview (how you handle data, your SOC 2 status or plan) so a procurement security review cannot stall your first big deal.
Plan the second motion. Most companies end up running both, bottom-up to acquire and top-down to expand. Decide which you will add, and when.

The sales motion you choose shapes your pricing, your hiring, and your product roadmap, so it is worth getting right on purpose instead of by accident. If you want the system around it, from go-to-market to operating with AI, that is what we teach in the AI Operating System for Startups. For the business-model version of this question, see how to build a service-as-software company, and for the same YC partner on building internal AI infrastructure, see internal AI infrastructure for startups.

Sources

Pete Koomen on bottoms-up vs top-down sales, the Y Combinator video this article distills.
Pete Koomen: YC profile. He co-founded and was CTO of Optimizely before becoming a YC partner.
Slack is named as the canonical bottom-up example; the mid-market and enterprise price-floor figures (around 10,000 and 100,000 dollars) are Koomen's from the video.