A conversion funnel is the sequence a visitor moves through between landing on your page and buying. On a multi-page site the stages are obvious. On a single landing page most of the funnel happens inside one screen, which makes it easy to assume there is nothing to measure between arrival and signup.
There is. Scrolling to pricing. Playing the demo video. Expanding an FAQ. Starting the form. Each is a step short of the purchase that signals intent, and the term for those steps is a micro conversion.
The macro conversion is the one you actually want: the signup, the sale, the booked call.
The intermediate steps matter because on most pages the macro conversion is too rare to learn from. If your page gets 200 visitors a month and converts at 4%, you get 8 events. You cannot run a test on 8 events. You can, though, watch what the other 192 people did before they left.
Why the end of the funnel is a measurement problem
Statistical significance needs volume. A page with 200 monthly visitors and a handful of conversions will take months to produce a trustworthy answer about a headline change.
Meanwhile every one of those 200 visitors leaves a trail. Some read to the pricing section and stopped. Some bounced from the hero. Some watched 40 seconds of the video. Those events arrive in the hundreds rather than the single digits, which makes them usable now.
They are a proxy, and proxies mislead. Optimising for scroll depth can produce a page that people read and never buy from. Treat micro conversions as diagnosis rather than as the target.
What to instrument
Ordered by how much they tell you per unit of effort:
CTA clicks by section. If your page has three CTAs, knowing which one people press tells you where conviction arrives. This is the highest-value thing on the list.
Scroll depth to a named section. Not "50% of the page", which means nothing across pages of different lengths. Scroll to the pricing block, or to the testimonials.
Form starts against form completions. A visitor who typed into the first field and left told you the form is the problem. A visitor who never touched it told you the copy above it is.
Video plays and watch time, if a video carries weight on the page.
Time to first interaction. Long times mean people are reading. They can also mean people are confused.
Where the ceiling on this sits
Our own crawl gives a sense of how far most companies are from needing any of this. We scored 675 live B2B SaaS landing pages from Y Combinator's directory in July 2026.
30.7% had no call to action in the page body, excluding header, nav and footer. The median page carried 2 CTA elements and 1 distinct CTA label. The median form-field count was 0.
A page with no in-body CTA has no micro conversions to instrument, because it never asks for anything. Fix that before you add analytics.
The link between the steps and the sale
Unbounce's data across roughly 41,000 landing pages found conversion falling as CTA links multiply: 13.5% at one link, 11.9% at two to four, 10.5% at five or more. They also found 68.2% of pages carried more than five.
That gradient is why per-CTA click tracking earns its place. If your page has six links and one of them takes 80% of the clicks, you have found both your real CTA and the five you can delete.
A caution on reading these numbers
Micro conversions are cheap to collect and easy to over-read.
A page can improve on every micro metric and sell less. Scroll depth rises when copy gets vague and people hunt for the answer. Time on page rises when a form confuses people.
Use micro conversions to form a hypothesis about where attention breaks. Confirm with the macro number, even if confirming takes a month.
Getting started without new tooling
If your page is hosted somewhere that reports views, unique visitors and CTA clicks, you already have the two most useful signals. Add a source breakdown and you can tell whether a weak conversion rate is a page problem or a traffic problem.
That distinction saves more wasted effort than any other measurement on this page. A page converting at 1% from a broad social post and 9% from search is a working page with a targeting problem.
Sources
- Unbounce, Conversion Benchmark Report. Approximately 41,000 landing pages, 464M pageviews, 57M conversions.
- LeadzLander, The State of SaaS Landing Pages (2026). 800 attempted, 675 scored, collected 2026-07-25, seed 20260725.