Landing Page Conversion Analyzer

What this tool does

CM8-374 finds which landing pages are losing visitors, split by where they came from and what device they used, and ranks them by the money you would recover rather than by which has the worst percentage.

Everything runs inside this single file — no account, no upload, no network request of any kind.

What one row is

One row is one page, one source, one device, one period. That is deliberately granular, because the interesting differences are almost always inside a page rather than between pages: the same pricing page converts entirely differently on mobile than on desktop, and to paid traffic than to organic.

Start with your ten most-visited pages split by device. That is usually enough to find the largest problem on the site.

Comparing like with like

The most important setting here, and the one that stops the analysis producing stupid advice.

A guide converting at 0.8% and a pricing page converting at 8% are not competitors. Ranking them together produces a list recommending you delete your best content — which is genuinely what happens when conversion analysis is done on a whole site at once.

Each row is tagged with what the visitor came for: ready to buy, comparing options, learning, or a general entry point. By default each page is compared against the best page doing the same job, which is a target somebody has already proved is reachable on your own site.

The other two options are the site average, which is blunter, and the page's own best device or source, which is the strictest and most useful comparison when you already know which page to fix.

Rank by money, not by percentage

This is what changes the order of work. A page converting at 1% with two hundred sessions has a terrible rate and is worth almost nothing to fix. A page converting at 4% against a 5% comparison, with forty thousand sessions, is a much better page and a far larger prize.

Set the value per conversion to gross profit if you have it. Revenue works, and overstates every prize by your cost of goods.

The mobile gap

Present on nearly every site, priced on almost none of them. Mobile typically carries most of the traffic and converts at half to two-thirds the desktop rate, and the gap is usually treated as a fact of nature rather than as the largest single opportunity on the site.

The tiles price it directly: what mobile matching desktop would be worth across the mobile sessions you already have. In the sample it is comfortably the biggest number on the page.

Some of the gap is real — people on phones are often earlier in their thinking. Not all of it is, and the way to find out is to look at what the page actually does on a phone. In the sample, the pricing page puts its pricing table below three screens of feature icons.

Bounce, and what it hides

The second figure is the useful one and it is rarely calculated. It separates two completely different problems:

• High bounce, good rate if they stayed — the page works for the right people and the wrong people are arriving. That is a traffic problem: fix the advert, the keyword or the link.
• Low bounce, poor rate if they stayed — people are interested and the page is not converting them. That is a page problem: the offer, the form, the proof, the price.

Bounce definitions vary between analytics tools and some have abandoned the metric entirely. Whichever definition you use, use it consistently; the comparison between rows matters more than the absolute figure.

Load time

Record what a real visitor experiences — a phone on a mobile connection — rather than the figure from a fast machine on an office network. The two commonly differ by a factor of three.

The load chart groups sessions into speed bands and shows the conversion rate of each. The relationship is usually visible, and it is not proof of anything: heavy pages are often heavy because of the images and video that persuade people, so the slow pages and the persuasive pages are frequently the same pages.

Treat a gap here as a hypothesis worth testing, not a cause established. The A/B test calculator is the tool for settling it.

The same page, different visitors

The source table shows how each traffic source behaves overall. Expect email and direct to convert several times better than paid or social, because those people already know you — that is not a judgement on the paid campaigns.

What is worth acting on is a source converting badly on a page built for it. A paid campaign sending traffic to a page that converts other sources well is usually a message mismatch: the advert promised something the page does not immediately deliver.

Too few sessions

Below the session threshold, a conversion rate is noise. Twenty-one conversions from a hundred and forty sessions looks like a spectacular 15% and could easily be 8% or 22%.

Those rows are shown in the table with all their figures, marked too few sessions to judge, excluded from the rankings, and never used as a comparison for anything else. The sample includes one for exactly this reason.

What to do about a leaking page

In rough order of how often it works:

• Match the page to the promise. Whatever the advert, link or search result said should be the first thing visible.
• Move the action up. On mobile especially, anything below the second screen may as well not exist.
• Ask for less. Every form field costs conversions — though see the A/B test tool for what shortening a form does to lead quality.
• Answer the objection. Price, contract length, whether it works with what they already have.
• Make it faster, particularly on mobile, and particularly where the page is over twice your slow threshold.

Then test the change rather than assuming it worked. Landing page changes are the one area where before-and-after comparisons are most often confounded by everything else that moved that week.

What this cannot tell you

• Why a page converts badly. It finds the leak; the cause needs somebody to look at the page, ideally on a phone, ideally as a stranger.
• Anything about causation between load time, bounce and conversion.
• Whether the conversions are worth having. A page optimised into cheap, unqualified leads will look excellent here.
• What happens after the conversion.
• Whether the prize is achievable. It assumes a page could reach what a comparable page already reaches, which is a reasonable target and not a promise.

Printing and sharing

The Report tab prints the tiles, charts and both tables with a title block you fill in. The ranked prize chart is the one to take to a prioritisation meeting — it converts an argument about which page looks worst into an ordered list of money.

Saving your work

Pages are held in this browser, on this computer, and stay there between visits. Use the backup button to write a JSON file you control.

Accuracy & disclaimer

Every figure here comes from your analytics, with whatever definitions that system uses. The tool compares and prices; it has no view on why a page behaves as it does.