A startup landing page converts when a visitor can tell in a few seconds who it is for, what pain it removes, and what they get by clicking. Put those three things in the first screen: the audience in the headline, the pain in plain words, and a button that names the payoff, such as “Book a 20-minute pipeline review” in place of “Learn more.” Below that, show proof you can count and the product itself, then the price. Give the page one call to action. At Seed, many buyers arrive from a referral or an email you sent, so the page’s job is to keep the deal you already started. Nielsen Norman Group’s 2018 eyetracking study found that people spend about 57% of their viewing time above the fold.
This page covers the homepage and campaign landing pages of a B2B startup selling to businesses. For an AI product deciding whether “AI” belongs in the headline, read how to market an AI product without leading with AI. For the funnel the demo requests enter, read customer development as your early sales funnel.
Decide what the page has to do
6sense’s 2024 buyer study found that B2B buyers are about 70% of the way through their process before they contact a seller, and 81% already have a preferred vendor at that first contact. Your page is part of that silent research. For an early company, many of those visitors come from a founder email, an intro, an event, or a post. They already know your name. They want to confirm the fit and find the next step.
Unbounce analyzed SaaS campaign landing pages from July 2023 to July 2024. Median conversion rates varied by traffic source:
| Traffic source | SaaS median conversion (Unbounce 2024) | What the page should do |
|---|---|---|
| 16.9% | Confirm what the email promised and take the click | |
| Paid search | 4.1% | Match the search words in the headline, then prove fit |
| Paid social | 2.9% | Explain the problem fast to a visitor who was not looking |
| Display ads | 0.3% | Treat it as awareness. Do not judge the page on it |
Unbounce counts any conversion event on its hosted pages, from a demo request to a download, so read these as a spread by channel. A demo request on your homepage is a different event.
The FounderNexus session view: for a company whose first hundred customers came through founder intros and referrals, the website’s job was to avoid losing the deal. Sequence channels warm first, then cold and paid. Visitor identification tools that flag companies on your site help once traffic exists and do nothing before. Judgment from a FounderNexus session, not a survey.
Give the first screen three jobs
Nielsen Norman Group tracked more than 130,000 eye fixations from 120 participants in 2018. About 57% of viewing time went to the first screen and 74% to the first two. Within the first screen, more than 65% of viewing time went to the top half.
| Element | Write this | Avoid |
|---|---|---|
| Headline | The buyer and the outcome: “Month-end close in three days for multi-entity startups” | A category label or a vision line: “Finance, reimagined” |
| Subhead | The pain in the buyer’s words and how the product removes it | A list of features |
| Call to action | The payoff of the click: “Get a free close-time review” | “Learn more,” “Let’s talk,” or two equal buttons |
| Proof line | Two to five customer logos or one operating number | “Trusted by leading companies” with no names |
| Product image | A real screen showing the outcome, large enough to read | An abstract illustration or a tiny dashboard |
Editorial examples, not tested copy. Write the headline in the words customers use. Aaron Epstein, a YC group partner, suggests asking existing users how they would describe the product and reusing their phrasing. In YC’s design review with Pete Koomen, co-founder of Optimizely, the pair argue for one clear call to action and against elements that compete with it.
The FounderNexus session view on the first screen: name the audience, lead with the specific pain, and make the button say what the visitor gets. The strongest example in the session was a bookkeeping startup whose real hook was the founder’s complaint about an accountant who stopped answering. That complaint made a better headline than the category. The same session pushed logos and hard operating numbers up the page: an average response time in minutes says more than a satisfaction percentage. Judgment from a FounderNexus session, not a survey.
Write it the way your buyer reads
In Unbounce’s SaaS readability data, pages written at a 5th to 7th grade reading level converted at 12.9%. Pages written at a professional level converted at 2.1%. The best-converting SaaS pages ran 250 to 725 words.
Read the page aloud to someone outside your company. Mark each word they ask about. Replace it, or define it in the next sentence. Keep technical detail for the pages a technical buyer opens on purpose, such as docs, security, and integrations.
Test with strangers before you test with traffic. Ask five people from your target role to look at the first screen for five seconds, then ask what the company does and who it is for. If they cannot answer, rewrite the headline before you buy any ads.
Use proof you can count
| Proof | What it tells the buyer | Use it when |
|---|---|---|
| Customer logos | Companies like theirs pay you | You have permission and two or more logos from the target segment |
| Operating number | The product does the job at a measurable level | You can pull it from your own data, with the period and sample stated |
| Named customer quote | A peer had the same pain and it went away | The quote names the result and the person’s role |
| Short case study | How the outcome happened, step by step | A buyer needs a story to forward internally |
| Security and compliance page | You pass the review before procurement asks | You sell into companies with a security team |
| Satisfaction score | Little by itself | Only with a named survey, sample, and date |
TrustRadius found that 35% of buyers said customer reviews on the vendor’s site would make them more likely to buy. Put proof under the first screen, and repeat one piece of it next to the call to action.
Show the product, then the price
TrustRadius surveyed 2,185 technology buyers in February 2022. 81% wanted to find pricing information on their own, and 54% first looked for it during initial research. Asked what would make them more likely to buy, 71% named pricing on the website and 70% named access to a demo or free trial. Buyers who could not find a price split three ways: 43% said they would delay and keep researching, 41% would book a sales call to get it, and 16% would stop considering the product.
| Pricing choice | Fits when | What to show |
|---|---|---|
| Full public price | Self-serve or a simple plan with few variables | Plans, what each includes, and the charge metric |
| Starting price | Deals vary by seats, usage, or modules | “Plans from $X a month” plus what moves the number |
| Price after a short qualifier | Enterprise deals with a security review and custom terms | What the buyer gets on the call, how long it takes, and typical contract shape |
Put a product screen, a short recorded walkthrough, or a sandbox before the pricing section. For the charge metric itself, see seat vs usage vs outcome pricing for AI SaaS.
Keep one call to action and a short form
| Weak button | Stronger button | Why |
|---|---|---|
| Learn more | See how the close works in 3 minutes | Names the payoff and the time cost |
| Contact us | Book a 20-minute workflow review | Says what happens on the call |
| Get started | Start a 14-day trial, no card | Removes the main risk |
| Request a demo | Get a demo on your own data | Promises their own problem on screen |
Editorial examples, not tested copy. Ask only for what you need to route the request: name, work email, company, and one question about the problem. Collect the rest on the call. Repeat the same call to action at the end of the page rather than adding a second one.
Unbounce found that 79% of visits to SaaS landing pages came from mobile devices, with mobile and desktop average conversion rates close together (6.4% and 6.2%). Check the first screen and the form on a phone before you ship.
Measure qualified demos by source
| Metric | Definition | Use it to decide |
|---|---|---|
| Visitor to call to action | Clicks or form submits ÷ unique visitors, by source | Whether the first screen works for that audience |
| Form completion | Submits ÷ form starts | Whether the form asks too much |
| Request to qualified demo | Requests that pass your written qualified line ÷ all requests | Whether the page attracts your buyer |
| Qualified demo to opportunity | Demos that end on a dated next step ÷ qualified demos | Whether the page set the right expectation |
| Time to first reply | Hours from request to your first answer | Whether follow-up loses warm buyers |
Editorial example, not a benchmark: a Seed company gets 2,000 visits in a month, half from founder email and intros. The page produces 50 demo requests, a 2.5% rate. 18 pass the qualified line and 7 end on a dated next step. The headline change worth testing is the one that raises the 18, not the 50.
For context only, Unbounce reported a SaaS median of 3.8% across its hosted campaign pages and 11.6% or higher for the top quarter. Those pages count any conversion and are not homepages. There is no reliable public benchmark yet for visitor-to-qualified-demo conversion on a Seed or Series A B2B site, so track your own by source for two or three months.
Rewrite the page in order
Landing page rewrite
- Name one buyer. The role or industry that pays you today.
- Collect their words. Ask customers how they describe the product and the pain (Epstein).
- Write the headline. Audience and outcome in one line.
- Write the button. The payoff of the click and the time it takes.
- Pick the proof. Logos and one operating number with its period.
- Show the product. A real screen or a short walkthrough before the price.
- Decide the price display. Full, starting at, or after a qualifier (TrustRadius).
- Cut the copy. Plain words, 250 to 725 words on the page (Unbounce).
- Test with five strangers. Five seconds on the first screen, then ask what it does.
- Track qualified demos. By source, monthly, with the first reply time.
Mistakes that cost the warm visitor
Leading with the category. “Modern accounting” tells the buyer nothing about their problem. Name the pain.
Writing for investors. Market size and vision statements belong in the deck.
Two equal buttons. “Book a demo” beside “Start free” splits the click. Pick the one that matches your sales motion.
Hiding the product. Abstract art and stock photos hide what the buyer would use.
A price wall with no context. Show the outcome and the product first.
Buying traffic before the page works. Fix the first screen on warm traffic, then pay for cold.
Judging the page on raw conversion rate. Fifty unqualified requests cost more founder time than ten good ones.
Pair with messaging, the funnel, and search
For an AI product, decide what the headline leads with using how to market an AI product without leading with AI. Move each qualified demo into customer development as your early sales funnel. If you have no customers yet, run the page as a test with how to validate demand before you build. Split writing hours between converting pages and category answers with AEO vs SEO for B2B SaaS. Set the charge metric with seat vs usage vs outcome pricing. Orient the rest of the motion from the GTM hub.
Sources
- Nielsen Norman Group, Scrolling and Attention. Therese Fessenden, 15 April 2018. Eyetracking study, more than 130,000 fixations from 120 participants on 1920×1080 screens. About 57% of page-viewing time above the fold; 74% in the first two screenfuls; 81% in the first three; more than 65% of above-the-fold viewing time in the top half of the viewport.
- Unbounce, SaaS conversion rate benchmarks. Conversion Benchmark Report 2024, data from 23 July 2023 to 23 July 2024 (methodology: more than 41,000 landing pages, 464 million unique visitors, 57 million conversions). SaaS median 3.8% against 6.6% for all industries. Median by channel: email 16.9%, paid search 4.1%, paid social 2.9%, display 0.3%. 79% of SaaS landing page visits on mobile; average conversion mobile 6.4%, desktop 6.2%. 5th to 7th grade copy 12.9% against professional-level 2.1%. Best-converting word count 250 to 725.
- Unbounce, What’s a good conversion rate?. Based on the 2024 report. Top 25% of SaaS landing pages convert at 11.6% or above.
- TrustRadius, 2022 B2B Buying Disconnect: The Age of the Self-Serve Buyer. Survey of 2,185 technology buyers, February 2022. 81% want to find pricing on their own; 54% first look for pricing in initial research. More likely to buy if: pricing on the website 71%, demo or free trial access 70%, customer reviews on the site 35%. Without pricing: 43% delay and research, 41% book a sales call, 16% stop considering.
- 6sense, 2024 Buyer Experience Report. 9 October 2024. Buyers nearly 70% through the process before engaging sellers; 81% have a preferred vendor at first contact.
- Y Combinator, Design Review: How to convert more visitors into customers. Greg Kumparak, 26 October 2023. Aaron Epstein with Pete Koomen: one clear call to action, fewer competing elements, faster path to the product’s value, less signup friction.
- Glance, Video critique of YC startup websites. Summary of a YC design review of W22 startup websites with Aaron Epstein and Zain Ali. Visitors ask what this is and whether it is for them; reuse customers’ own words in the headline; keep the page skimmable.
- FounderNexus. Founder session, live teardown of early-stage B2B sites, May 2026. Name the audience; lead with the specific pain; buttons that say what the visitor gets; logos and operating numbers over satisfaction percentages; product before price; at pre-seed the site keeps deals that warm channels start; visitor identification tools only once traffic exists. Not a survey. No closed-session numbers.
Related
- Go-to-market
- How to market an AI product without leading with AI
- Customer development as your early sales funnel
- AEO vs SEO for B2B SaaS startups
- Seat vs usage vs outcome pricing for AI SaaS
- How to validate demand before you build
Founders who have rewritten an early B2B site after a live teardown will pressure-test your headline, your proof, and your call to action in a FounderNexus session.