A homepage has an impossible brief. It has to serve the customer checking their account, the journalist looking for a press contact, the candidate looking at jobs, the existing customer looking for support, and the stranger who just clicked an ad about one specific thing.
It serves all of them adequately and none of them well. For organic traffic that is a fine trade. For paid traffic it is a per-click cost, because the stranger who clicked your ad had exactly one question and your homepage opened with a summary of the company.
The ten second test, applied to message match
Nielsen Norman Group's survival analysis of page visits found that most abandonment happens in the first ten seconds. That window buys roughly 20 words of reading, and the visitor spends it deciding whether this page is an answer to whatever the link promised.
So the test is mechanical. Write down the promise the ad made. Open the page, count to ten, close it, and write down what the page offered. If those two sentences are not close to identical, you found the leak.
Illustrative breakdown of the ten second window documented by Nielsen Norman Group, not a measured study. The window itself is theirs. The allocation is our framing.
Where the first ten seconds go on each destination. Homepage: work out what company this is: 3s. Homepage: find the relevant section: 4s. Homepage: left with 3s to decide: 3s. Landing page: the promise, restated: 2s. Landing page: 8s left to decide: 8s.
What a landing page needs
Less than most landing page templates suggest. The failure mode of a purpose-built landing page is a wall of sections that exist because a template had slots.
- 01A headline that restates the ad's promise in the ad's own words
- 02One action. Two competing calls to action is a decision the visitor did not ask to make
- 03Proof appropriate to the claim, which for an early product means being specific about what is built rather than borrowing logos
- 04The obvious objection, answered. Price, commitment, or what happens after the form
- 05Nothing else. No global navigation offering fourteen ways to leave
The navigation point is the one people resist and it is straightforward. A landing page exists to serve a single decision. A header menu is a list of alternative decisions, presented at the exact moment you want one.
The version of this that is honest
One warning, because this advice curdles easily. There is a style of landing page that strips navigation, hides the price, invents urgency and treats the visitor as an obstacle. It converts in the short term and it is a poor trade for anything with a renewal.
The distinction we would draw is between removing distraction and removing information. A single call to action removes distraction. A hidden price removes information the visitor needs, and they will find it eventually, in a worse mood.
For measurement, the thing to instrument is the first funnel step rather than the final conversion. If a campaign's traffic never reaches step 2, no amount of checkout optimisation will help, and the landing page is where the money went. You get about ten seconds has the underlying research, and put a price on every funnel step covers how to price that first step properly.