Why SaaS Websites Get Traffic but No Signups

Traffic is not the problem. That is usually the hardest thing to accept, because traffic is the number that is easy to look at and easy to buy more of.

If people are arriving and not signing up, something between the click and the button is failing. This article is about finding out which thing — before you spend a quarter redesigning the wrong part.

The short answer

There are six reasons a SaaS site gets traffic and no signups. In rough order of how often we find them:

1. The visitor never understood the product. They left before the page said what it was.
2. They understood it and did not believe it. Claims outnumbered proof.
3. They believed it and it was not for them. The traffic was never qualified.
4. They wanted it and could not tell what the next step cost. The CTA hid its terms.
5. They took the next step and the form stopped them. Friction at the boundary.
6. They signed up and you never counted it. The tracking is wrong.

These produce different symptoms, and the point of the rest of this article is telling them apart. Guessing is expensive: fixing messaging when the real fault is traffic quality moves nothing, and you will conclude that design does not work.

Start by splitting the traffic

Before touching the page, split conversion rate by source. This one step resolves more cases than any amount of staring at the design.

That fourth row is the only one where the page is genuinely the suspect. The others are upstream, and no redesign fixes them.

The third row is the most commonly misread. A site that converts direct traffic and nothing else is not a good site — it is a site that only works for people who already knew what you did before they arrived. That is a rung-one failure, and it is invisible to a founder testing their own homepage, because the founder cannot un-know the product.

Telling the six apart

Each failure leaves a different fingerprint. Here is how to read them.

That last row deserves more attention than it gets. We have seen a "conversion problem" that was a consent banner blocking analytics for most European visitors, and another that was a signup event firing on a page that had been renamed six months earlier. Before you redesign anything, confirm the number is real by counting rows in your own database.

Cause one: they never understood it

This is the most common and the least visible to the people who built the product.

The research is worth being precise about, because a great deal of what circulates on this topic is invented. Nielsen Norman Group, reporting a Microsoft Research study by Chao Liu and colleagues that analysed over two billion visits across 205,873 pages, found that dwell time follows a Weibull distribution with negative aging: visitors are most likely to leave immediately, and the first ten seconds decide most departures. The curve only flattens after about thirty seconds. You can read the original write-up here.

Ten seconds is not enough time for atmosphere. It is enough time for a category.

The fix is almost always to make the headline plainer than feels comfortable. "Fraud detection for online stores" over "The intelligence layer for commerce." Differentiation belongs in the line underneath, once the visitor knows what kind of thing they are looking at. The full sequence is in our guide to SaaS website design.

How to be sure it is fixed: show it to five people outside your company for five seconds each. If they cannot name the category, it is not fixed, however much the team likes it.

Cause two: they did not believe it

A useful way to think about this: every claim a page makes opens a question, and proof is what closes it. "Set up in five minutes" opens what's the catch? "Enterprise-grade security" opens says who?

A page carries proof debt when it has opened more questions than it has closed. Unpaid proof debt does not read as ambition. It reads as risk — and a visitor deciding whether to put their company's data into your product is running a risk assessment whether or not you designed the page as one.

This is why polished pages fail. Visual polish increases the number and confidence of claims without adding anything that settles them.

The audit

Print the page. Highlight claims in one colour and proof in another. Count.

Most pages we audit come in around four claims to every one piece of proof. Under two-to-one is healthy.

Then find the single biggest unanswered objection. Ask three people who did not buy why they did not. The answer is usually the same sentence, and it is usually not on the page anywhere.

Where proof goes

Next to the claim it settles — not collected into a testimonials band halfway down. A quote about how fast onboarding was belongs beside the onboarding claim. A carousel of five unrelated quotes answers no specific question, which is why it is so often the most-scrolled-past section on a SaaS site.

Cause three: the traffic was never qualified

If the page is clear and proven and still nothing converts, the visitors may simply not be buyers.

Two patterns account for most of it:

Ranking for research intent. You rank for "what is customer data platform" and get students, competitors and analysts. Useful for awareness, terrible for conversion. This is not a failure — it is just the wrong page for that visitor. Send informational traffic to an article, not a product page.

Ads promising something the page does not. The ad says "free forever", the page opens with "book a demo". The mismatch is not dishonesty; it is usually two people writing in two tools. It converts like dishonesty though.

The fix is upstream: change the targeting, or change the page's promise to match the traffic you are actually getting. If one campaign converts far better than the rest, it has found the message — take its promise and put it on the page.

Cause four: the CTA hid its terms

A call to action has to answer three things: what happens when I click, what it costs, and what I am committing to.

The pattern: hesitation at the button is nearly always an unanswered cost. Naming the cost — even when the cost is real, like "card required" — converts better than hiding it, because it moves the discovery from after the click to before it, and a person who clicks knowing the terms is a person who meant it.

One primary action per page. Not per section. A page offering a free trial, a demo and a downloadable guide with equal weight is asking the visitor to choose your sales process for you.

Cause five: the form stopped them

Everything above is about the page. This is about the boundary.

The single most reliable lever is removing fields. Each required field is a reason to stop, and most exist because someone once wanted the data, not because anything is done with it. Company size, phone number and "how did you hear about us" are the usual suspects.

Some rules that hold up:

• Ask only for what you need to create the account. Everything else can be asked after, inside the product, where the person is already committed.
• Do not ask for a card on a free trial unless you have decided you want fewer, more qualified trials — a legitimate choice, but make it deliberately.
• Validate as they type, not on submit. Discovering three errors at once after filling everything in is where people quit.
• Never clear the form on error. This still happens, and it is close to unforgivable.
• Offer SSO if your buyers are technical. For developer tools it often outperforms every copy change you could make.

Watch field-level drop-off before guessing. The field people abandon on is usually not the one the team expects.

Cause six: you never counted it

Confirm the number is real before acting on it.

• Compare analytics signups against rows in your own database for the same week.
• Check the consent banner is not blocking measurement for a large share of visitors.
• Confirm the signup event still fires — page renames and framework upgrades break these silently.
• Check bot filtering. Automated traffic inflates sessions and deflates conversion rate, making a healthy page look broken.

What to fix first

In order, because the order saves money:

1. Confirm the measurement is real. A day's work, and it sometimes ends the project.
2. Split by source. Another day. Tells you whether it is a page problem at all.
3. Fix comprehension. Cheapest change with the largest effect. Usually a headline and a subheading.
4. Pay down proof debt. One named customer with a real number beats a month of copy edits.
5. Fix the CTA's terms. An afternoon.
6. Strip the form. An afternoon, and often the largest single jump.
7. Only then consider a redesign. By this point you will know whether you need one, and what it has to do.

Most teams start at seven. That is why redesigns so often deliver nothing — they change everything except the thing that was wrong, and they destroy whatever was quietly working. If you are heading that way, read our framework for redesigning without making it worse first.

Frequently asked questions

What is a good conversion rate for a SaaS website?

There is no honest single number, and any article giving you one without naming its source and sample is guessing. It varies enormously by price point, motion, traffic mix and how conversion is defined — a free-trial signup and an enterprise demo request are not the same event. The useful benchmark is your own site last quarter, split by source.

Should I run A/B tests to find the problem?

Usually not at first. A/B testing needs traffic volume most early-stage SaaS sites do not have, and it compares two guesses rather than telling you what is wrong. Diagnose first with the methods above; test later, once you have enough traffic for a result to mean anything.

How long before a fix shows up in the numbers?

Comprehension and CTA changes show up almost immediately — within a week for most sites, because they affect every visitor. Proof and positioning changes take longer, since they affect people with longer consideration cycles. Give any change at least one full sales cycle before judging it.

My homepage converts fine but my landing pages do not. Why?

Usually because a landing page inherited the homepage's structure. A homepage serves several audiences and can afford breadth; a landing page serves one promise made in an ad, and every element that does not support that promise is working against it. The section-by-section anatomy covers the difference.