I measured 24 pages Google ranks for landing page advice. Four of them put the form above the fold.

One goal per page. Form near the top. Cut the ways out. You have read that paragraph before, probably this week, and I have said it to clients with a straight face for years.

Then something obvious occurred to me. The pages Google ranks first for that advice are themselves pages. Nobody scores them on it. Nobody checks.

So I took the top results for four teaching queries about PPC landing pages, kept one page per domain so a single publisher could not swamp the sample, downloaded all twenty-four and counted what was actually on them.

Half of them have no form at all

Twelve of twenty-four carry a form somewhere. The other twelve have none.

An article does not owe anyone a form, so on its own that proves nothing. It is worth saying out loud anyway, because most of those twelve spend a whole section explaining where your form belongs.

Four out of twenty-four put it near the top

I looked at the first fifteen thousand characters of body markup and asked whether a form appears in there. Four pages did.

Twenty told me to do it. Four did it.

I want to be careful here, because markup order is a proxy and CSS can float a form up from anywhere. It cuts both ways, and I would not defend the number four to a decimal place. What I would defend is the shape: the practice is close to universal in the writing and close to absent in the shipping.

Check your own header before you touch anything else

Count the links in your header and nav. Right now, on the page you are about to run traffic to.

The median in my sample is 24. One page had 97. The advice everyone repeats, including me, is to strip the exits down to almost nothing, and the pages teaching that ship two dozen of them.

Then look at what your form asks for. Median in my sample is two visible fields, ignoring hidden ones, and the longest form I found has six.

That surprised me. The warnings you read are always about forms with ten or fifteen fields, and there was not one of those anywhere in the sample. Either the industry fixed that years ago and kept writing about it, or it was never as common as the advice implies. Both are worth knowing before you spend an afternoon cutting fields nobody has.

Speed, where the average lies to you

Median response was 619 milliseconds. Perfectly fine.

One page took 17.25 seconds. Three were over two seconds.

If I had averaged those twenty-four numbers I would have got something around one and a half seconds and reported that the field is a bit slow. It is not a bit slow. It is fast, with a couple of pages that are broken, and those are two different problems with two different fixes.

An hour I would not spend again

I tried to see whether pages with a form ranked differently from pages without one. Twenty-four results across four queries is nowhere near enough to say anything about that, which I knew going in, and I checked anyway.

The number I keep coming back to

Twenty-four links in the header. Not the form position, not the speed.

Every one of those is a documented, deliberate way for someone who was ready to convert to go and look at something else instead. And they are on the pages written by people who tell you to remove them.

I do not think anyone is being dishonest. I think a header is inherited from the site template, nobody treats it as part of the landing page, and the advice stops at the fold. Mine did too, until I counted.

Questions people ask about this