I get asked to look at landing pages constantly. For years my feedback was a list of observations: the headline is vague, the form is long, the testimonials are generic. Useful, unstructured, and impossible for anyone else on my team to reproduce.
So I built a sequence. Nine questions, always in the same order, because the order is what makes it a diagnostic instead of a critique. A page that fails question two cannot be fixed by anything in questions five through nine, and most teams spend their time on questions five through nine.
One: can a stranger say what this is in five seconds
Open the page, look at the top section only, close it, and say out loud what the company does and who it is for.
If you cannot, nothing else matters. This is the failure mode in roughly a third of the pages I review, and it is almost always caused by a headline written to sound impressive rather than to be understood. "Reimagining workflow for modern teams" fails. "Scheduling software for restaurants with hourly staff" passes.
The fix is boring and effective: state what it is, then be clever in the subhead if you must.
Two: does the page match the promise that brought them
Whatever the ad, the email, or the search result said, the page must visibly continue it. Same words, same offer, same framing.
A visitor who clicked on "free shipping over 50 dollars" and lands on a page whose hero talks about craftsmanship has to do work to reconnect the two ideas, and a meaningful share will not.
I check this by opening the traffic source and the page side by side. The mismatch is usually obvious within seconds and it is usually caused by two different people writing the two assets months apart. If organic search is the source, remember that the promise a searcher saw was the title and description, not the H1.
Three: is there one action, or several
Count the distinct things the page asks the visitor to do. Navigation links count. A newsletter box counts. A chat widget counts. A footer full of links counts.
Dedicated landing pages with full site navigation consistently underperform. Every additional path is a chance to leave.
One page, one action, repeated. Secondary information can exist. Secondary actions should not.
Four: is the value proposition specific or categorical
Categorical: "save time and money." Specific: "cuts your monthly close from nine days to three."
Categorical claims are invisible because every competitor makes them. Specific claims are risky, which is precisely why they are persuasive. The reader assumes you would not commit to a number you could not defend.
If the client cannot produce a specific claim, that is a business problem surfacing as a copy problem, and the honest response is to go find out what actually changes for customers rather than to write around it.
Five: what objection is unaddressed
List every reason someone would not buy. Price, switching cost, trust, timing, whether it works for their situation, what happens if it fails.
Now find each objection on the page. Typically half are missing entirely.
The strongest landing pages address objections explicitly rather than avoiding them. A section that says "this is probably not right for you if" outperforms almost every other block I have tested, because it converts a skeptical reader into a self selected one. The same honesty mechanism I described in trust signals and conversion.
Six: is the proof specific and attributable
Three testimonials with a first name and a job title are worth less than one with a full name, a company, a photograph, and a specific outcome.
The pattern that kills credibility is the perfect testimonial. "This changed everything for our business" reads as manufactured. "We were skeptical about the price and it took two months to pay for itself" reads as real, and it handles an objection at the same time.
Proof includes more than testimonials: named clients, real numbers, third party coverage, credentials. The psychology underneath is in social proof psychology.
Seven: what is the friction cost of the action
Count the fields. Count the clicks. Count the decisions.
Then ask whether each field is required now or could be collected later. Most forms ask for information the sales process does not need until the second conversation, and every field is a chance to abandon.
The other half of friction is uncertainty about what happens next. "Submit" tells the visitor nothing. "Get the audit within two business days" tells them the shape of the commitment they are making, which is the actual thing they are hesitant about.
Eight: does the page read at the speed people actually read
Not a readability score for its own sake, but a genuine check on whether the sentences are doing work. Long paragraphs on a phone are skipped entirely. Dense jargon signals that the reader is not the intended audience.
I read pages aloud, mostly to catch the sections where a writer got tangled. When a paragraph is noticeably harder to say out loud than its neighbors, it is nearly always because the writer was uncertain about the argument and hid it in complexity.
Nine: what does it look like on a phone at 375 pixels
Last, because it is a check rather than a strategy question, and non negotiable because it is where most of your traffic is.
The specific failures I find over and over: hero text pushed below the fold by an oversized image, a form that requires horizontal scrolling, a sticky header eating a third of the viewport, buttons too close to be tapped accurately, and a testimonial carousel nobody swipes.
Load performance belongs here too. A page that takes five seconds on a mobile connection has a conversion problem that no copy edit fixes, and the fixes are in Core Web Vitals in practice.
How to use the sequence
Do not fix everything. The order exists so you can stop early.
If the page fails questions one through four, stop and rewrite the top of the page. Do not touch the form, the testimonials, or the button color. Those changes will not register against a page that has not communicated what it is.
If it passes one through four and fails five through seven, the problem is trust and friction, and that is where the testing budget goes.
If it passes everything and still underperforms, the problem is upstream: the traffic is wrong, or the offer is wrong. A page cannot fix a bad offer, which is why I usually go to offer design next rather than running another test.
Test the top, not the bottom
The last thing I will say is about testing discipline. Teams test buttons because buttons are easy to test and the results are clean.
But in the tests I have run and reviewed, changes to the headline and the primary value proposition produce effect sizes many times larger than anything below the fold. The reason teams avoid them is that headline changes require agreeing on what the company actually does, which is a harder meeting than picking a color.
Before you run anything, check whether you have the traffic to detect the effect at all with an A/B test calculator. Most small business landing pages do not have the volume to resolve a 5 percent difference in any reasonable timeframe, which means the correct strategy is not testing at all. It is making one large, well reasoned change based on the nine questions above and watching what happens over a quarter. That is less scientific and, at low traffic, considerably more honest.