Rick’s Landing Page Audit Rubric

A useful landing-page audit shows what was observed, which principle it tests, why the issue matters, how confident the finding is, and what to change first. It does not promise a conversion lift from a screenshot or a language model's opinion.

The eight dimensions

The weights are a prioritization device, not a scientific conversion model. A page can score well and still fail because the offer is unwanted, traffic is mismatched, or the product cannot deliver the promise.

1. Offer clarity

The first screen should let the intended buyer explain what the product is, who it helps, the concrete outcome, and the next step. Clever language is not automatically bad, but it should not require several paragraphs before the category and value become legible.

Strong evidence includes a visible product, artifact, workflow, result example, or precise mechanism. Weak evidence includes abstract claims such as “transform,” “revolutionize,” or “unlock growth” without a defined object or action.

2. Audience and problem fit

The page should name a recognizable situation rather than address “businesses” or “everyone.” Look for the trigger that makes the visitor search now, the cost of the current workaround, and the constraints that determine fit.

3. Primary action

Audit the label, destination, surrounding promise, and consistency of the primary CTA. “Learn more” often hides the next step; “Run the free readiness check” declares it. Multiple CTAs can coexist when they represent a clear ladder, but they should not compete at the same visual priority.

4. Proof and specificity

Proof must be traceable. A current metric needs a definition, period, source, and retrieval date. A case needs permission and context. A tool result needs visible evidence and limitations. Testimonials without provenance and counters without definitions reduce trust rather than increase it.

5. Objection and risk handling

Answer what the product requires, excludes, costs, stores, accesses, and expects from the customer. Explain trial, refund, cancellation, or handover terms accurately. Risk reversal should not promise outcomes the seller cannot control.

6. Information hierarchy

The page should move from the visitor’s job to the offer, evidence, process, boundaries, and next step. Repeated slogans, several product names for the same offer, and large sections that do not answer a buying question create cognitive load.

7. Friction and accessibility

Count visible form fields, check labels, required-state clarity, error behavior, input types, privacy context, mobile layout, focus states, keyboard operation, and contrast. Automated review can identify HTML signals; dynamic behavior and screen-reader quality still need manual testing.

8. Measurement readiness

Define the product-qualified action before changing the page. Record a baseline, ship one coherent hypothesis, verify instrumentation, and retest over a meaningful window. A visual improvement without a declared outcome is design activity, not an experiment.

Finding format

Every useful finding should include:

1. observed page evidence;
2. rubric dimension;
3. user or business risk;
4. recommended change;
5. confidence and limitation;
6. verification method after the change.

The free Roast is an automated first pass. Deep Roast is a separate paid expert page-teardown product and should not be confused with a whole-business diagnostic. Neither replaces customer research, accessibility testing, analytics, or a controlled experiment.