Homepage reviews usually go like this: one person likes the headline, another does not, and the loudest voice wins. Nobody can say what the page actually fails to tell a stranger. A messaging audit swaps opinions for a checklist and a score, so you can see what is missing and fix it in order.
Use this audit when your homepage gets visits but not demos and you suspect the message is the problem. It takes an afternoon: twenty checks across five areas, a simple scoring sheet, a five second test with strangers and a short list of what to fix first. Copy the checks into your own sheet as you go.
Key takeaways
Audit what a stranger understands, not what the team intended.
Check five areas in order: who, what, why you, proof and next step.
Score each point pass, partial or fail, then fix the lowest area first.
Run a five second test with people outside the company before you rewrite anything.
Fix the headline and subhead before the design. A good layout cannot rescue a vague promise.
What is a SaaS website messaging audit?
A messaging audit is a structured review of what your site says, to whom and how clearly. It checks the words, not the visual design, and ends in a ranked list of fixes.
A messaging audit is not a content audit. A content audit reviews your blog and pages for traffic, duplicates and rankings. A messaging audit, sometimes called a SaaS website audit for copy, reviews whether the pages tell a stranger who the product is for and why to choose it. You may need both, but they fix different problems.
An audit also differs from a redesign brief. A brief describes what you want. An audit finds what is wrong. Doing the audit first often shrinks the redesign, because many problems turn out to be copy problems.
If you have not yet decided who the product is for, start with Positioning an AI Agent Product. An audit can only measure messaging against a decision.
What is the Five Layer Audit?
The Five Layer Audit checks a page against the five questions a buyer asks in order. It gives each question four checks.
Layer | The buyer’s question | Checks |
|---|---|---|
1. Who | Is this for me? | Buyer named, role or company type clear, Not For line optional |
2. What | What does it do? | Headline states the outcome, subhead names the mechanism, category is clear |
3. Why you | Why this and not another? | Difference stated, alternative named, claim is specific |
4. Proof | Can I believe it? | Evidence near each claim, named customers, numbers sourced |
5. Next step | What do I do now? | One main action, matching the buyer’s readiness, no dead ends |
The order matters. A page can have strong proof and still fail layer 1, and then the proof never gets read.
What are the 20 checks?
Score each check pass (2), partial (1) or fail (0). Twenty checks gives a maximum of 40.
Layer 1: Who
The headline or subhead names the buyer or their situation.
A stranger could say what kind of company this serves.
The page avoids “teams of all sizes” and “every business.”
A reader can tell if the product is not for them.
Layer 2: What
The headline states an outcome, not a category label.
The subhead says how, in verbs a buyer would use.
A first time visitor can name the product category.
The hero shows the product or a real output.
Layer 3: Why you
The page names what you replace or beat.
Each claim could not be pasted onto a competitor.
Jargon (AI, platform, solution) is explained or replaced with the job.
One idea leads the page, not five.
Layer 4: Proof
Every number has a source or a customer.
Named customers or specific stories appear above the fold or close to it.
Claims about accuracy or speed say how they were measured.
Logos are real customers, not partners or press.
Layer 5: Next step
One primary action is clear on every screen.
The action matches the buyer’s stage (try, see, talk).
The button says what happens next.
The page has no dead ends, such as a pricing page with no action.
Also read: AI Homepage Messaging Mistakes: 7 Fixes From Three Homepage Teardowns
How do you score the audit?
Add the points, then read the score by layer, not just in total. A layer with fewer than five of eight points is a priority.
Total score | What it suggests |
|---|---|
32 to 40 | The messaging works. Test small changes. |
22 to 31 | Clear gaps. Fix the lowest layer first. |
12 to 21 | The page likely confuses buyers. Rewrite before redesigning. |
0 to 11 | The positioning may be undecided. Pause and decide it first. |
These bands are our rule of thumb from reviewing pages, not a measured standard. The by layer view is the useful part. If layers 1 and 2 are weak, nothing below them will be read.
Pro tip: Have two people score separately, then compare. The checks they disagree on are the ones the team has not defined.
How do you run the five second test?
Show a stranger the homepage for five seconds, then hide it and ask what they remember. The method was named by Christine Perfetti in the mid 2000s. Its five second figure comes mostly from practitioner experience, and a Smashing Magazine case study found ten seconds worked better for complex pages, so use ten if your homepage is busy.
Find five people who match your buyer but do not work with you.
Show the page for five seconds, then close it.
Ask: What does this company do? Who is it for? What would you do next?
Record the answers word for word.
Pass if four of five name the category and the buyer.
The wording of their answers is useful in itself. People who describe your product in their own words give you the language the page should use.
How do you audit with real pages?
Open the page at desktop and phone width and read it as a stranger would, from the top.
Capture screenshots and note the date, since pages change.
Read only the first screen aloud. Does it pass layers 1 and 2?
Scroll once. Where does proof first appear?
Click the main button. Does the next step match the promise?
Check the same on pricing, a use case page and a product page.
Audit three pages, not only the homepage. Messaging often drifts between the homepage and the pages visitors land on from search or ads.
What should you fix first?
Fix the lowest layer first, and within it the headline and subhead. A lower layer problem makes upper layer fixes useless.
Weakest layer | First fix |
|---|---|
Who | Name the buyer in the headline or subhead |
What | Rewrite the headline as an outcome in the buyer’s words |
Why you | Replace shared claims with one specific difference |
Proof | Add a named customer or sourced number next to the main claim |
Next step | Reduce to one primary action and rename the button |
Pro tip: Change one layer at a time and run the five second test again. If you change five things, you will not know which one helped.
What are the most common messaging audit mistakes?
Auditing opinions. Use the score, not preference.
Only reviewing the homepage. Visitors land elsewhere.
Testing with colleagues. They already know the product.
Fixing proof first. Proof for an unclear claim goes unread.
Rewriting without a baseline. Save the old score and screenshots.
Skipping a retest. Run the checks again after the changes.
Also read: Product Messaging for AI Startups, AI Landing Page Design.
To see messaging and design built together, read our ORYN AI agent monitoring landing page case study, or browse all of our case studies.
Keep reading: Product Messaging for AI Startups: What It Is and When to Hire Help, AI Homepage Messaging Mistakes: 7 Fixes From Three Homepage Teardowns, Digital Health SaaS Messaging: How to Write for Clinicians, Admins and Patients and How to Show AI Accuracy and Trust on a Landing Page
Want us to score your homepage?
Send us your homepage and we will return a scored audit with the three highest impact fixes. Book a messaging audit · See our work
SaaS messaging audit FAQs
A messaging audit reviews what a website says, to whom and how clearly, then ranks the fixes. It checks words and structure, not visual style. A good audit scores the buyer, the offer, the difference, the proof and the next step, so changes are driven by evidence rather than opinion.
Score the page against a checklist covering who it is for, what it does, why you, proof and next step. Add a five second test with strangers. Fix the weakest layer first, starting with the headline and subhead, then retest after changes.
A five second test shows a page to a stranger for five seconds, then hides it and asks what the company does, who it is for and what they would do next. It reveals whether the headline and hero communicate quickly, without any explanation from you.
Audit after any change to your product, buyer or positioning, and at least once a year. Rerun the five second test after major page changes. Messaging drifts as pages are added, so check three pages, not only the homepage.
Yes. Kraftbase is a product messaging and UI/UX design studio for AI and B2B SaaS companies. We score your pages against a structured checklist, then rewrite and redesign in order of impact, so the words and layout are fixed together.
Start with three: the homepage, your pricing page and the page most visitors reach from search or ads. That sample shows whether the promise holds across the buyer’s path. Add product and use case pages afterwards if scores vary widely.
On our 40 point sheet, 32 or above suggests the messaging works, and 22 to 31 shows clear gaps. These bands are our rule of thumb, not an industry standard. Compare scores over time and by layer rather than treating one number as a verdict.
About the author
Dhaval Agrawal is the founder and CEO of Kraftbase, a product messaging and UI/UX design studio in Vadodara, India, that helps AI and B2B SaaS teams in the US, UAE and India explain and design what they’ve built.
Sources
Five Layer Audit and scoring bands: Kraftbase framework, not a published standard
Messaging audit
Want us to score your homepage?
Send us your homepage and we will return a scored audit with the three highest impact fixes.
Homepage reviews usually go like this: one person likes the headline, another does not, and the loudest voice wins. Nobody can say what the page actually fails to tell a stranger. A messaging audit swaps opinions for a checklist and a score, so you can see what is missing and fix it in order.
Use this audit when your homepage gets visits but not demos and you suspect the message is the problem. It takes an afternoon: twenty checks across five areas, a simple scoring sheet, a five second test with strangers and a short list of what to fix first. Copy the checks into your own sheet as you go.
Key takeaways
Audit what a stranger understands, not what the team intended.
Check five areas in order: who, what, why you, proof and next step.
Score each point pass, partial or fail, then fix the lowest area first.
Run a five second test with people outside the company before you rewrite anything.
Fix the headline and subhead before the design. A good layout cannot rescue a vague promise.
What is a SaaS website messaging audit?
A messaging audit is a structured review of what your site says, to whom and how clearly. It checks the words, not the visual design, and ends in a ranked list of fixes.
A messaging audit is not a content audit. A content audit reviews your blog and pages for traffic, duplicates and rankings. A messaging audit, sometimes called a SaaS website audit for copy, reviews whether the pages tell a stranger who the product is for and why to choose it. You may need both, but they fix different problems.
An audit also differs from a redesign brief. A brief describes what you want. An audit finds what is wrong. Doing the audit first often shrinks the redesign, because many problems turn out to be copy problems.
If you have not yet decided who the product is for, start with Positioning an AI Agent Product. An audit can only measure messaging against a decision.
What is the Five Layer Audit?
The Five Layer Audit checks a page against the five questions a buyer asks in order. It gives each question four checks.
Layer | The buyer’s question | Checks |
|---|---|---|
1. Who | Is this for me? | Buyer named, role or company type clear, Not For line optional |
2. What | What does it do? | Headline states the outcome, subhead names the mechanism, category is clear |
3. Why you | Why this and not another? | Difference stated, alternative named, claim is specific |
4. Proof | Can I believe it? | Evidence near each claim, named customers, numbers sourced |
5. Next step | What do I do now? | One main action, matching the buyer’s readiness, no dead ends |
The order matters. A page can have strong proof and still fail layer 1, and then the proof never gets read.
What are the 20 checks?
Score each check pass (2), partial (1) or fail (0). Twenty checks gives a maximum of 40.
Layer 1: Who
The headline or subhead names the buyer or their situation.
A stranger could say what kind of company this serves.
The page avoids “teams of all sizes” and “every business.”
A reader can tell if the product is not for them.
Layer 2: What
The headline states an outcome, not a category label.
The subhead says how, in verbs a buyer would use.
A first time visitor can name the product category.
The hero shows the product or a real output.
Layer 3: Why you
The page names what you replace or beat.
Each claim could not be pasted onto a competitor.
Jargon (AI, platform, solution) is explained or replaced with the job.
One idea leads the page, not five.
Layer 4: Proof
Every number has a source or a customer.
Named customers or specific stories appear above the fold or close to it.
Claims about accuracy or speed say how they were measured.
Logos are real customers, not partners or press.
Layer 5: Next step
One primary action is clear on every screen.
The action matches the buyer’s stage (try, see, talk).
The button says what happens next.
The page has no dead ends, such as a pricing page with no action.
Also read: AI Homepage Messaging Mistakes: 7 Fixes From Three Homepage Teardowns
How do you score the audit?
Add the points, then read the score by layer, not just in total. A layer with fewer than five of eight points is a priority.
Total score | What it suggests |
|---|---|
32 to 40 | The messaging works. Test small changes. |
22 to 31 | Clear gaps. Fix the lowest layer first. |
12 to 21 | The page likely confuses buyers. Rewrite before redesigning. |
0 to 11 | The positioning may be undecided. Pause and decide it first. |
These bands are our rule of thumb from reviewing pages, not a measured standard. The by layer view is the useful part. If layers 1 and 2 are weak, nothing below them will be read.
Pro tip: Have two people score separately, then compare. The checks they disagree on are the ones the team has not defined.
How do you run the five second test?
Show a stranger the homepage for five seconds, then hide it and ask what they remember. The method was named by Christine Perfetti in the mid 2000s. Its five second figure comes mostly from practitioner experience, and a Smashing Magazine case study found ten seconds worked better for complex pages, so use ten if your homepage is busy.
Find five people who match your buyer but do not work with you.
Show the page for five seconds, then close it.
Ask: What does this company do? Who is it for? What would you do next?
Record the answers word for word.
Pass if four of five name the category and the buyer.
The wording of their answers is useful in itself. People who describe your product in their own words give you the language the page should use.
How do you audit with real pages?
Open the page at desktop and phone width and read it as a stranger would, from the top.
Capture screenshots and note the date, since pages change.
Read only the first screen aloud. Does it pass layers 1 and 2?
Scroll once. Where does proof first appear?
Click the main button. Does the next step match the promise?
Check the same on pricing, a use case page and a product page.
Audit three pages, not only the homepage. Messaging often drifts between the homepage and the pages visitors land on from search or ads.
What should you fix first?
Fix the lowest layer first, and within it the headline and subhead. A lower layer problem makes upper layer fixes useless.
Weakest layer | First fix |
|---|---|
Who | Name the buyer in the headline or subhead |
What | Rewrite the headline as an outcome in the buyer’s words |
Why you | Replace shared claims with one specific difference |
Proof | Add a named customer or sourced number next to the main claim |
Next step | Reduce to one primary action and rename the button |
Pro tip: Change one layer at a time and run the five second test again. If you change five things, you will not know which one helped.
What are the most common messaging audit mistakes?
Auditing opinions. Use the score, not preference.
Only reviewing the homepage. Visitors land elsewhere.
Testing with colleagues. They already know the product.
Fixing proof first. Proof for an unclear claim goes unread.
Rewriting without a baseline. Save the old score and screenshots.
Skipping a retest. Run the checks again after the changes.
Also read: Product Messaging for AI Startups, AI Landing Page Design.
To see messaging and design built together, read our ORYN AI agent monitoring landing page case study, or browse all of our case studies.
Keep reading: Product Messaging for AI Startups: What It Is and When to Hire Help, AI Homepage Messaging Mistakes: 7 Fixes From Three Homepage Teardowns, Digital Health SaaS Messaging: How to Write for Clinicians, Admins and Patients and How to Show AI Accuracy and Trust on a Landing Page
Want us to score your homepage?
Send us your homepage and we will return a scored audit with the three highest impact fixes. Book a messaging audit · See our work
SaaS messaging audit FAQs
A messaging audit reviews what a website says, to whom and how clearly, then ranks the fixes. It checks words and structure, not visual style. A good audit scores the buyer, the offer, the difference, the proof and the next step, so changes are driven by evidence rather than opinion.
Score the page against a checklist covering who it is for, what it does, why you, proof and next step. Add a five second test with strangers. Fix the weakest layer first, starting with the headline and subhead, then retest after changes.
A five second test shows a page to a stranger for five seconds, then hides it and asks what the company does, who it is for and what they would do next. It reveals whether the headline and hero communicate quickly, without any explanation from you.
Audit after any change to your product, buyer or positioning, and at least once a year. Rerun the five second test after major page changes. Messaging drifts as pages are added, so check three pages, not only the homepage.
Yes. Kraftbase is a product messaging and UI/UX design studio for AI and B2B SaaS companies. We score your pages against a structured checklist, then rewrite and redesign in order of impact, so the words and layout are fixed together.
Start with three: the homepage, your pricing page and the page most visitors reach from search or ads. That sample shows whether the promise holds across the buyer’s path. Add product and use case pages afterwards if scores vary widely.
On our 40 point sheet, 32 or above suggests the messaging works, and 22 to 31 shows clear gaps. These bands are our rule of thumb, not an industry standard. Compare scores over time and by layer rather than treating one number as a verdict.
About the author
Dhaval Agrawal is the founder and CEO of Kraftbase, a product messaging and UI/UX design studio in Vadodara, India, that helps AI and B2B SaaS teams in the US, UAE and India explain and design what they’ve built.
Sources
Five Layer Audit and scoring bands: Kraftbase framework, not a published standard
Messaging audit
Want us to score your homepage?
Send us your homepage and we will return a scored audit with the three highest impact fixes.