Segment 1 of 11

Technical SEO

82/10014 of 17 checks passed

Whether search engines can reach, read and index your pages at all.

3 remarks to fix here

Search engines arrive at your shop with no instructions at all
Check T03
The site publishes a sitemap
What it means if this fails
Nothing lists which pages exist, so pages nothing links to may never be found.
What you would do about it
Publish a sitemap at /sitemap.xml listing every page you want found, and add a Sitemap line to robots.txt pointing at it.
Severity
high
This is costing you money right now, or it is something a stranger could take advantage of today. It is not a warning about the future.
Why it matters
Start here. A high finding is usually the difference between a bill that grows quietly and one that stops growing, or between a door that is open and one that is shut.
Under an hour
/:1
Three addresses every crawler asks for before reading anything are missing from your site. There is no robots file telling them what to read, no sitemap listing which pages exist, and nothing written for AI assistants. Your pages are reachable by following links, so they will be found eventually. Pages nobody links to may never be found at all.
Fix it yourself
My home page is missing the files search engines look for first. Add a robots file served at /robots.txt that allows everything except the basket and account pages and points at my sitemap, a sitemap served at /sitemap.xml listing every page including each product, and an llms.txt describing what Harper Goods sells and linking the shop, about and contact pages.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
Your shop page tells search engines not to list it
Check T10
No page that should be found is hidden from search engines
What it means if this fails
A page tells search engines not to list it at all.
What you would do about it
Remove the noindex robots meta tag from the pages named so search engines can list them again. Keep it only on private pages such as the basket and account screens.
Severity
high
This is costing you money right now, or it is something a stranger could take advantage of today. It is not a warning about the future.
Why it matters
Start here. A high finding is usually the difference between a bill that grows quietly and one that stops growing, or between a door that is open and one that is shut.
Under an hour
/shop:9
This page carries a noindex instruction, which is a direct request to leave it out of search results entirely. It is almost always left behind from when a site was being built and nobody wanted a half-finished page found. It is doing exactly what it says: your shop cannot appear in any result, no matter what anyone searches for.
Fix it yourself
My /shop page is telling search engines to ignore it. Remove the noindex instruction from /shop so the page can be listed normally, and check that no other customer-facing page has it either. Only the basket and account pages should stay hidden.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
The sitemap exists but the robots file does not mention it, so it has to be guessed at
Check T04
robots.txt points at the sitemap
What it means if this fails
The sitemap exists but the robots file does not mention it, so it has to be guessed at.
What you would do about it
Add a Sitemap line to robots.txt giving the full address of the sitemap.
Severity
low
Real, but small. Either the effect is minor at your current size, or it only shows up in narrow circumstances.
Why it matters
Fix it if it is quick and you are already in that file. Nothing bad happens if you leave it.
What we found: robots.txt has no Sitemap: line.
Fix it yourself
On harpergoods.com: Add a Sitemap line to robots.txt giving the full address of the sitemap. What we found: robots.txt has no Sitemap: line.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.

14 checks this site passed

The site has a robots.txt
Check T01
The site has a robots.txt
What it means if this fails
The site has no robots file, so crawlers arrive with no instructions.
What you would do about it
Add a robots.txt at the site root that allows crawling and points at the sitemap.
robots.txt does not block the whole site
Check T02
robots.txt does not block the whole site
What it means if this fails
The site's robots file tells every search engine to stay out.
What you would do about it
Remove the site-wide Disallow rule from robots.txt so search engines can read the site. Keep any narrower rules for private areas.
Every page says which address is the real one
Check T06
Every page says which address is the real one
What it means if this fails
Pages do not say which address is the real one, so the same page can be counted several times.
What you would do about it
Add a canonical link to the head of every page, each one pointing at that page's own address.
No page points that address at a different page
Check T07
No page points that address at a different page
What it means if this fails
Pages point their canonical address at a different page, which asks search engines to ignore them.
What you would do about it
Correct the canonical link on the pages named so each points at its own address rather than at another page.
Every page declares a mobile viewport
Check T08
Every page declares a mobile viewport
What it means if this fails
Pages do not declare a mobile viewport, so phones render them at desktop width.
What you would do about it
Add a viewport meta tag with width=device-width and initial-scale=1 to the head of every page.
Every page says what language it is in
Check T09
Every page says what language it is in
What it means if this fails
Pages do not declare their language, so search engines guess which country to show them in.
What you would do about it
Set a lang attribute on the html element of every page, matching the language the page is written in.
No noindex is being sent in the response headers
Check T11
No noindex is being sent in the response headers
What it means if this fails
The server sends a noindex header, which hides the page and is invisible in a browser.
What you would do about it
Stop the server sending an X-Robots-Tag noindex header with these pages.
The insecure address redirects to the secure one
Check T12
The insecure address redirects to the secure one
What it means if this fails
The site answers on http as well as https, so there are two copies of every page.
What you would do about it
Permanently redirect every http address to its https equivalent so the site is served on one protocol only.
A missing address answers with a proper 404
Check T13
A missing address answers with a proper 404
What it means if this fails
Addresses that do not exist answer as though they do, so search engines index pages that are not there.
What you would do about it
Make addresses that do not exist return a 404 status instead of 200, so search engines stop indexing pages that are not there.
Every link we followed led to a real page
Check T14
Every link we followed led to a real page
What it means if this fails
A link on the site leads to a page that is not there.
What you would do about it
Fix or remove the links named so nothing on the site points at a page that is not there.
The site lives at one address, not two
Check T15
The site lives at one address, not two
What it means if this fails
Both the www and the bare address serve the site, so every signal it earns is split in two.
What you would do about it
Pick one host, www or bare, and permanently redirect the other to it so the site lives at one address.
No page names two of them at once
Check T17
No page names two of them at once
What it means if this fails
A page declares two canonical addresses, so search engines ignore both and decide for themselves.
What you would do about it
Leave one canonical link on each page and delete the duplicates.
Nothing on a secure page is loaded insecurely
Check T18
Nothing on a secure page is loaded insecurely
What it means if this fails
A secure page loads something over an insecure address, which every browser blocks outright, so that part of the page is missing for everyone but the person who built it.
What you would do about it
Change every http:// address in the page's own tags to https://. If a host genuinely has no secure version, move the file onto your own site and serve it from there.
Language versions point at each other
Check T19
Language versions point at each other
What it means if this fails
The site is published in more than one language and nothing tells a search engine the versions are the same page, so they compete with each other instead of being offered to the right reader.
What you would do about it
In the <head> of every translated page, add one <link rel="alternate" hreflang="xx" href="…"> for each language including the page itself, plus one with hreflang="x-default" pointing at the version for everybody else.

Segment 2 of 11

Page Speed

100/1007 of 7 checks passed

Whether your pages can load quickly, judged on what makes them slow.

Nothing to do here.

Seven checks this site passed

Scripts do not hold the page up before it draws
Check V01
Scripts do not hold the page up before it draws
What it means if this fails
Scripts in the head have to finish downloading and running before anything appears, so the visitor sees a blank page for as long as that takes.
What you would do about it
Add the `defer` attribute to every <script src=…> in the <head>, or move those tags to just before </body>. Nothing on a marketing page needs to run before the page is drawn.
Pictures say how big they are before they arrive
Check V02
Pictures say how big they are before they arrive
What it means if this fails
Images with no width and height leave the browser no idea how much room to keep, so the text jumps down the page as each one loads and people tap the wrong thing.
What you would do about it
Give every <img> a width and a height attribute matching the file's real pixel dimensions. Keep the CSS sizing as it is; these two attributes only tell the browser the shape to reserve.
Pictures further down wait until they are needed
Check V03
Pictures further down wait until they are needed
What it means if this fails
Every image on the page is downloaded before the visitor has scrolled to any of them, so the ones they can actually see arrive last.
What you would do about it
Add loading="lazy" to every <img> below the first screenful. Leave it off the one image at the top: that one should load immediately.
No picture is heavy enough to hold the page up on its own
Check V04
No picture is heavy enough to hold the page up on its own
What it means if this fails
At least one image on the site is large enough that a phone on a normal connection spends seconds downloading it, which is time the visitor spends looking at nothing.
What you would do about it
Save the pictures named below at the size they are actually displayed, in WebP, and aim for under 200 KB each. Most photo tools export WebP directly; the visible quality difference at these sizes is nil.
Pictures use a format that is not twice the size it needs to be
Check V05
Pictures use a format that is not twice the size it needs to be
What it means if this fails
The photographs are served as PNG or JPEG, which every browser in use can already beat by roughly a third with no visible difference.
What you would do about it
Re-save the photographs as WebP and update the <img src> to match. If you need a fallback, wrap them in a <picture> element with the WebP first and the original underneath.
The page warns the browser about the other servers it needs
Check V06
The page warns the browser about the other servers it needs
What it means if this fails
The page loads fonts or scripts from other companies' servers without telling the browser in advance, so it discovers each one late and pays for the introduction before anything can be drawn.
What you would do about it
For each third-party host the page loads from, add <link rel="preconnect" href="https://that-host" crossorigin> in the <head>, above the tag that uses it.
The page itself is not enormous before anything else loads
Check V07
The page itself is not enormous before anything else loads
What it means if this fails
The page's own markup is large enough to be a noticeable download on a phone, before a single image or script is fetched.
What you would do about it
Ask your builder to find what is making the page this large. It is usually the same content included twice, once as HTML and once as a JSON blob for the JavaScript, or a very long page that should be split.

Segment 3 of 11

On-Page SEO

74/10014 of 19 checks passed

What each page says about itself: titles, summaries, headings and pictures.

5 remarks to fix here

Nobody sharing your contact page sees anything but the address
Check O05
Every page has a description
What it means if this fails
Pages carry no summary, so the search engine writes its own from a sentence off the page.
What you would do about it
Add a meta description of roughly 25 words to each page named, saying what that page is for.
Severity
high
This is costing you money right now, or it is something a stranger could take advantage of today. It is not a warning about the future.
Why it matters
Start here. A high finding is usually the difference between a bill that grows quietly and one that stops growing, or between a door that is open and one that is shut.
Under an hour
/contact:6
This page has no summary and no social preview. The summary is the grey line under your link in search results, and without one the search engine writes its own by taking a sentence off the page. The preview is what messaging apps read to build a card, so a customer sending your link to a friend sends a bare address with no picture and no explanation.
Fix it yourself
My /contact page has no description and no social preview. Add a description of about 25 words to /contact saying who we are and what someone can get in touch about, and add a preview image, title and description so the link shows a proper card when it is shared.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
Your journal page arrives completely empty
Check O13
Pages arrive with their content already in them
What it means if this fails
A page arrives empty and is only filled in after JavaScript runs.
What you would do about it
Render the content of the pages named into the HTML the server sends, so the words are there before any JavaScript runs.
Severity
high
This is costing you money right now, or it is something a stranger could take advantage of today. It is not a warning about the future.
Why it matters
Start here. A high finding is usually the difference between a bill that grows quietly and one that stops growing, or between a door that is open and one that is shut.
Half a day
/journal:14
This page sends almost no words. Everything a visitor reads is drawn afterwards by code running in their browser, which works perfectly for a person and leaves nothing for a search engine that reads the page as it arrives. The same is true for the assistants people now ask for recommendations. To them your journal is a blank sheet with a script attached.
Fix it yourself
My /journal page arrives empty because all the writing appears only after JavaScript runs. Make /journal send its real content in the page itself: the post titles, dates and opening paragraphs should be in the page before any script runs, so a search engine reading it sees the same words a visitor does.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
Three of your pages are all called Harper Goods
Check O03
No two pages share a title
What it means if this fails
Several pages share one title, so search engines cannot tell them apart.
What you would do about it
Give each of the pages named its own title describing what is on that page specifically.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
Under an hour
/about:5
Your about, delivery and stockists pages carry exactly the same title. The title is the blue line someone reads before deciding whether to click, so all three look like the same page to a person choosing between them, and a search engine deciding which one answers a question has nothing to tell them apart with. It usually resolves that by showing only one.
Fix it yourself
The /about page shares its title with two other pages. Give /about a title that says what is actually on it, something like Who makes Harper Goods candles, and make sure the delivery and stockists pages each get their own distinct title too.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
Your delivery page never says what it is about
Check O08
Every page has a main heading
What it means if this fails
A page has no main heading, so its subject is left to be guessed at.
What you would do about it
Add a single h1 heading at the top of each page named, saying what the page is about.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
Under an hour
/delivery:11
This page has no main heading. A heading is the one line a machine treats as the subject of the page, and yours has none, so a search engine and an assistant both have to work it out from the paragraphs. Delivery and returns is one of the things people most often ask an assistant about a shop, and this is the page that should answer it.
Fix it yourself
My /delivery page has no main heading. Add one at the top of /delivery that says plainly what the page covers, something like Delivery and returns at Harper Goods, and keep the existing text underneath it.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
Pages have no social preview, so a shared link arrives as a bare address with no picture
Check O14
Shared links carry a preview
What it means if this fails
Pages have no social preview, so a shared link arrives as a bare address with no picture.
What you would do about it
Add OpenGraph and Twitter card tags, a title, a description and an image, to the pages named so shared links show a preview.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
What we found: no OpenGraph or Twitter card tags on 4 pages: /contact, /delivery, /stockists, /journal.
Fix it yourself
On harpergoods.com: Add OpenGraph and Twitter card tags, a title, a description and an image, to the pages named so shared links show a preview. What we found: no OpenGraph or Twitter card tags on 4 pages: /contact, /delivery, /stockists, /journal.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.

14 checks this site passed

Pictures say how big they are before they arrive
Check V02
Pictures say how big they are before they arrive
What it means if this fails
Images with no width and height leave the browser no idea how much room to keep, so the text jumps down the page as each one loads and people tap the wrong thing.
What you would do about it
Give every <img> a width and a height attribute matching the file's real pixel dimensions. Keep the CSS sizing as it is; these two attributes only tell the browser the shape to reserve.
Every page has a title
Check O01
Every page has a title
What it means if this fails
A page has no title, so search engines invent one from whatever they find.
What you would do about it
Give every page a title element that says what that page is about.
No page title is a placeholder
Check O02
No page title is a placeholder
What it means if this fails
A page title is a placeholder rather than something a person would search for.
What you would do about it
Replace the placeholder titles named with ones that say what the business offers, phrased the way someone would search for it.
Titles are short enough to be shown in full
Check O04
Titles are short enough to be shown in full
What it means if this fails
Page titles are long enough that search engines cut them off mid sentence.
What you would do about it
Shorten the titles named to 60 characters or fewer so search results show them in full.
No two pages share a description
Check O06
No two pages share a description
What it means if this fails
Several pages share one description, so their search results all read the same.
What you would do about it
Write a separate meta description for each of the pages named.
Descriptions are a usable length
Check O07
Descriptions are a usable length
What it means if this fails
Descriptions are too short or too long to be shown as written.
What you would do about it
Rewrite the descriptions named to between 70 and 160 characters so they are shown as written.
Headings step down one level at a time
Check O09
Headings step down one level at a time
What it means if this fails
Heading levels are skipped, so the outline of the page reads as broken to a machine.
What you would do about it
Fix the heading order on the pages named so levels step down one at a time, with no jump from h1 straight to h3.
Every image says what it shows
Check O10
Every image says what it shows
What it means if this fails
Images carry no alt text, so search engines and screen readers cannot tell what they show.
What you would do about it
Add alt text to the images named, describing what each one shows. Leave alt empty only for images that are purely decorative.
Every page carries enough content to be worth reading
Check O11
Every page carries enough content to be worth reading
What it means if this fails
Pages carry so little text that there is nothing for a search engine to judge them on.
What you would do about it
Add real content to the pages named: at least 150 words saying what the page is for and what someone gets from it.
Every page links onward into the site
Check O12
Every page links onward into the site
What it means if this fails
A page has no links to the rest of the site, so anyone who lands there has nowhere to go.
What you would do about it
Add links from the pages named to the rest of the site, so a visitor who lands there has somewhere to go.
Pages inside the site say where they sit in it
Check O16
Pages inside the site say where they sit in it
What it means if this fails
Pages below the top level carry no breadcrumb data, so search results show a bare address where they could show the path through your site.
What you would do about it
Add a BreadcrumbList JSON-LD block to every page below the top level, listing each step from the home page down to that page, with a name and a URL for each.
The site has an icon to show beside its name
Check O17
The site has an icon to show beside its name
What it means if this fails
There is no icon declared, so the browser tab, the bookmark and the search result all show a blank square where every competitor shows a mark.
What you would do about it
Add a favicon: put an icon.png or favicon.ico at the site root and declare it with <link rel="icon" href="/icon.png"> in the <head>. 32×32 is enough.
Products are described in machine-readable terms
Check O18
Products are described in machine-readable terms
What it means if this fails
The things you sell are described only in words a person reads, so a search engine cannot show a price, a picture and availability beside your listing the way it does for your competitors.
What you would do about it
Add Product JSON-LD to every product page: name, image, description, and an "offers" block with price, priceCurrency and availability.
Product markup carries a price and whether it is in stock
Check O19
Product markup carries a price and whether it is in stock
What it means if this fails
The product markup exists but leaves out the price or whether the item is in stock, which are the two things a search result shows and the two a shopper decides on.
What you would do about it
In each product's "offers" block, fill in price, priceCurrency (as "EUR" or "USD") and availability (as "https://schema.org/InStock" or "OutOfStock"). Keep them matching what the page says.

Segment 4 of 11

Generative Engine Optimization

79/10011 of 14 checks passed

Whether an AI that writes an answer has anything of yours it can use.

3 remarks to fix here

A page arrives empty and is only filled in after JavaScript runs
Check O13
Pages arrive with their content already in them
What it means if this fails
A page arrives empty and is only filled in after JavaScript runs.
What you would do about it
Render the content of the pages named into the HTML the server sends, so the words are there before any JavaScript runs.
Severity
high
This is costing you money right now, or it is something a stranger could take advantage of today. It is not a warning about the future.
Why it matters
Start here. A high finding is usually the difference between a bill that grows quietly and one that stops growing, or between a door that is open and one that is shut.
What we found: 1 page sends almost no readable text and builds everything with JavaScript: /journal.
Fix it yourself
On harpergoods.com: Render the content of the pages named into the HTML the server sends, so the words are there before any JavaScript runs. What we found: 1 page sends almost no readable text and builds everything with JavaScript: /journal.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
Also counts toward On-Page SEO, where it is written up in full.
No page carries structured data, so nothing states in machine terms what this business is
Check G03
There is structured data somewhere on the site
What it means if this fails
No page carries structured data, so nothing states in machine terms what this business is.
What you would do about it
Add JSON-LD structured data to the site, starting with an Organization or LocalBusiness block on the home page.
Severity
high
This is costing you money right now, or it is something a stranger could take advantage of today. It is not a warning about the future.
Why it matters
Start here. A high finding is usually the difference between a bill that grows quietly and one that stops growing, or between a door that is open and one that is shut.
What we found: no JSON-LD structured data on any of the 9 pages we read.
Fix it yourself
On harpergoods.com: Add JSON-LD structured data to the site, starting with an Organization or LocalBusiness block on the home page. What we found: no JSON-LD structured data on any of the 9 pages we read.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
There is no llms.txt, so assistants have nothing written for them about this business
Check G01
The site publishes an llms.txt
What it means if this fails
There is no llms.txt, so assistants have nothing written for them about this business.
What you would do about it
Publish an llms.txt at the site root describing the business in plain sentences and linking the pages that matter.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
What we found: no /llms.txt.
Fix it yourself
On harpergoods.com: Publish an llms.txt at the site root describing the business in plain sentences and linking the pages that matter. What we found: no /llms.txt.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.

Eleven checks this site passed

Every page carries enough content to be worth reading
Check O11
Every page carries enough content to be worth reading
What it means if this fails
Pages carry so little text that there is nothing for a search engine to judge them on.
What you would do about it
Add real content to the pages named: at least 150 words saying what the page is for and what someone gets from it.
Pages inside the site say where they sit in it
Check O16
Pages inside the site say where they sit in it
What it means if this fails
Pages below the top level carry no breadcrumb data, so search results show a bare address where they could show the path through your site.
What you would do about it
Add a BreadcrumbList JSON-LD block to every page below the top level, listing each step from the home page down to that page, with a name and a URL for each.
Products are described in machine-readable terms
Check O18
Products are described in machine-readable terms
What it means if this fails
The things you sell are described only in words a person reads, so a search engine cannot show a price, a picture and availability beside your listing the way it does for your competitors.
What you would do about it
Add Product JSON-LD to every product page: name, image, description, and an "offers" block with price, priceCurrency and availability.
Product markup carries a price and whether it is in stock
Check O19
Product markup carries a price and whether it is in stock
What it means if this fails
The product markup exists but leaves out the price or whether the item is in stock, which are the two things a search result shows and the two a shopper decides on.
What you would do about it
In each product's "offers" block, fill in price, priceCurrency (as "EUR" or "USD") and availability (as "https://schema.org/InStock" or "OutOfStock"). Keep them matching what the page says.
The structured data says who the business is
Check G04
The structured data says who the business is
What it means if this fails
The structured data never says who this business is, only what the pages are.
What you would do about it
Add an Organization, LocalBusiness or Person block to the structured data so it states who the business is, not only what each page is.
None of the structured data is broken
Check G05
None of the structured data is broken
What it means if this fails
A structured data block is broken, so search engines discard it and the work counts for nothing.
What you would do about it
Fix the structured data blocks named so they are valid JSON and can be read.
Pages are structured into lists and tables an answer can lift
Check G06
Pages are structured into lists and tables an answer can lift
What it means if this fails
Content is unbroken prose, so there is nothing an AI answer can quote as a step or a comparison.
What you would do about it
Break the key content into lists or tables, so an assistant writing an answer has a step or a comparison it can lift.
The site points agents at its machine-readable version
Check M01
The site points agents at its machine-readable version
What it means if this fails
The site publishes a plain-text version for assistants but never mentions it in the response itself, so an assistant only finds it by guessing the address.
What you would do about it
Send a Link header on every page pointing at your llms.txt, so an assistant is told where the plain-text version of this site lives instead of having to guess: Link: </llms.txt>; rel="alternate"; type="text/plain".
Agents are told what this site lets them do
Check M05
Agents are told what this site lets them do
What it means if this fails
This site has an API but publishes no description an assistant can read on its own, so an assistant can talk about the business without being able to do anything with it.
What you would do about it
Publish a server card at /.well-known/mcp/server-card.json describing what an assistant may do here, or a skills index at /.well-known/agent-skills/index.json listing the tasks it can carry out.
The people behind the site are described in machine terms
Check X05
The people behind the site are described in machine terms
What it means if this fails
No structured data describes a person or an organisation, so nothing links this site to anyone real.
What you would do about it
Add Person or Organization structured data naming who is behind the site.
Structured data claims those profiles as the same business
Check F02
Structured data claims those profiles as the same business
What it means if this fails
The site links to its profiles but never declares they are the same business, so search engines must guess.
What you would do about it
List those profile addresses in the structured data using sameAs, so search engines connect them to this business.

Segment 5 of 11

AI Overviews

56/1005 of 9 checks passed

Whether you turn up inside the answer box above the search results.

4 remarks to fix here

A page arrives empty and is only filled in after JavaScript runs
Check O13
Pages arrive with their content already in them
What it means if this fails
A page arrives empty and is only filled in after JavaScript runs.
What you would do about it
Render the content of the pages named into the HTML the server sends, so the words are there before any JavaScript runs.
Severity
high
This is costing you money right now, or it is something a stranger could take advantage of today. It is not a warning about the future.
Why it matters
Start here. A high finding is usually the difference between a bill that grows quietly and one that stops growing, or between a door that is open and one that is shut.
What we found: 1 page sends almost no readable text and builds everything with JavaScript: /journal.
Fix it yourself
On harpergoods.com: Render the content of the pages named into the HTML the server sends, so the words are there before any JavaScript runs. What we found: 1 page sends almost no readable text and builds everything with JavaScript: /journal.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
Also counts toward On-Page SEO, where it is written up in full.
No page carries structured data, so nothing states in machine terms what this business is
Check G03
There is structured data somewhere on the site
What it means if this fails
No page carries structured data, so nothing states in machine terms what this business is.
What you would do about it
Add JSON-LD structured data to the site, starting with an Organization or LocalBusiness block on the home page.
Severity
high
This is costing you money right now, or it is something a stranger could take advantage of today. It is not a warning about the future.
Why it matters
Start here. A high finding is usually the difference between a bill that grows quietly and one that stops growing, or between a door that is open and one that is shut.
What we found: no JSON-LD structured data on any of the 9 pages we read.
Fix it yourself
On harpergoods.com: Add JSON-LD structured data to the site, starting with an Organization or LocalBusiness block on the home page. What we found: no JSON-LD structured data on any of the 9 pages we read.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
No page carries FAQ markup, so questions the business already answers cannot be shown as answers
Check A01
The site marks up its questions and answers
What it means if this fails
No page carries FAQ markup, so questions the business already answers cannot be shown as answers.
What you would do about it
Mark up the questions and answers the site already publishes as FAQPage structured data.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
What we found: no FAQPage or QAPage structured data on any page we read.
Fix it yourself
On harpergoods.com: Mark up the questions and answers the site already publishes as FAQPage structured data. What we found: no FAQPage or QAPage structured data on any page we read.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
No heading is phrased as a question, so nothing on the site matches how people search out loud
Check A02
Some headings are the questions people actually ask
What it means if this fails
No heading is phrased as a question, so nothing on the site matches how people search out loud.
What you would do about it
Phrase some headings as the questions people actually ask, and answer each one directly underneath.
Severity
low
Real, but small. Either the effect is minor at your current size, or it only shows up in narrow circumstances.
Why it matters
Fix it if it is quick and you are already in that file. Nothing bad happens if you leave it.
What we found: not one heading across 8 pages is phrased as a question.
Fix it yourself
On harpergoods.com: Phrase some headings as the questions people actually ask, and answer each one directly underneath. What we found: not one heading across 8 pages is phrased as a question.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.

Five checks this site passed

Pages are structured into lists and tables an answer can lift
Check G06
Pages are structured into lists and tables an answer can lift
What it means if this fails
Content is unbroken prose, so there is nothing an AI answer can quote as a step or a comparison.
What you would do about it
Break the key content into lists or tables, so an assistant writing an answer has a step or a comparison it can lift.
Questions are answered immediately underneath
Check A03
Questions are answered immediately underneath
What it means if this fails
Questions are asked as headings and then answered several paragraphs later, so no answer can be lifted.
What you would do about it
Put a short, direct answer in the first paragraph under each question heading, before any of the detail.
Pages say when they were written or updated
Check A04
Pages say when they were written or updated
What it means if this fails
Nothing says when the content was written, so a search engine cannot tell whether it is still true.
What you would do about it
Show a visible date on the content and a dateModified in the structured data, so a reader can tell it is current.
The main content is marked out from the furniture
Check E03
The main content is marked out from the furniture
What it means if this fails
Nothing separates the content from the navigation, so a machine has to guess where the page begins.
What you would do about it
Wrap the main content of each page named in a main element so a machine can tell it from the navigation.
An assistant asking for plain text gets plain text
Check M02
An assistant asking for plain text gets plain text
What it means if this fails
An assistant that asks for a plain-text version of a page is sent the full HTML instead, so it has to read a page built for a browser and work out which parts are the content.
What you would do about it
Serve a plain-text version of each page when the request asks for one. If your site is behind Cloudflare, turn on Markdown for Agents in the dashboard. Otherwise set your server to reply with Content-Type: text/markdown and the page's text when a request sends Accept: text/markdown, and keep sending HTML to everyone else.

Segment 6 of 11

Answer Engine Optimization

38/1003 of 8 checks passed

Whether your pages answer questions directly enough to be quoted.

5 remarks to fix here

No page carries structured data, so nothing states in machine terms what this business is
Check G03
There is structured data somewhere on the site
What it means if this fails
No page carries structured data, so nothing states in machine terms what this business is.
What you would do about it
Add JSON-LD structured data to the site, starting with an Organization or LocalBusiness block on the home page.
Severity
high
This is costing you money right now, or it is something a stranger could take advantage of today. It is not a warning about the future.
Why it matters
Start here. A high finding is usually the difference between a bill that grows quietly and one that stops growing, or between a door that is open and one that is shut.
What we found: no JSON-LD structured data on any of the 9 pages we read.
Fix it yourself
On harpergoods.com: Add JSON-LD structured data to the site, starting with an Organization or LocalBusiness block on the home page. What we found: no JSON-LD structured data on any of the 9 pages we read.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
A page has no main heading, so its subject is left to be guessed at
Check O08
Every page has a main heading
What it means if this fails
A page has no main heading, so its subject is left to be guessed at.
What you would do about it
Add a single h1 heading at the top of each page named, saying what the page is about.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
What we found: no h1 heading on 1 page: /delivery.
Fix it yourself
On harpergoods.com: Add a single h1 heading at the top of each page named, saying what the page is about. What we found: no h1 heading on 1 page: /delivery.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
Also counts toward On-Page SEO, where it is written up in full.
No page carries FAQ markup, so questions the business already answers cannot be shown as answers
Check A01
The site marks up its questions and answers
What it means if this fails
No page carries FAQ markup, so questions the business already answers cannot be shown as answers.
What you would do about it
Mark up the questions and answers the site already publishes as FAQPage structured data.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
What we found: no FAQPage or QAPage structured data on any page we read.
Fix it yourself
On harpergoods.com: Mark up the questions and answers the site already publishes as FAQPage structured data. What we found: no FAQPage or QAPage structured data on any page we read.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
No heading is phrased as a question, so nothing on the site matches how people search out loud
Check A02
Some headings are the questions people actually ask
What it means if this fails
No heading is phrased as a question, so nothing on the site matches how people search out loud.
What you would do about it
Phrase some headings as the questions people actually ask, and answer each one directly underneath.
Severity
low
Real, but small. Either the effect is minor at your current size, or it only shows up in narrow circumstances.
Why it matters
Fix it if it is quick and you are already in that file. Nothing bad happens if you leave it.
What we found: not one heading across 8 pages is phrased as a question.
Fix it yourself
On harpergoods.com: Phrase some headings as the questions people actually ask, and answer each one directly underneath. What we found: not one heading across 8 pages is phrased as a question.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
There is no FAQ or help page, so the questions customers ask are not answered anywhere findable
Check E01
There is a page dedicated to answering questions
What it means if this fails
There is no FAQ or help page, so the questions customers ask are not answered anywhere findable.
What you would do about it
Add a page that answers the questions customers ask before they buy, and link it from the main navigation.
Severity
low
Real, but small. Either the effect is minor at your current size, or it only shows up in narrow circumstances.
Why it matters
Fix it if it is quick and you are already in that file. Nothing bad happens if you leave it.
What we found: no FAQ, help or support page in the sitemap or among the pages we read.
Fix it yourself
On harpergoods.com: Add a page that answers the questions customers ask before they buy, and link it from the main navigation. What we found: no FAQ, help or support page in the sitemap or among the pages we read.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.

Three checks this site passed

Headings step down one level at a time
Check O09
Headings step down one level at a time
What it means if this fails
Heading levels are skipped, so the outline of the page reads as broken to a machine.
What you would do about it
Fix the heading order on the pages named so levels step down one at a time, with no jump from h1 straight to h3.
Questions are answered immediately underneath
Check A03
Questions are answered immediately underneath
What it means if this fails
Questions are asked as headings and then answered several paragraphs later, so no answer can be lifted.
What you would do about it
Put a short, direct answer in the first paragraph under each question heading, before any of the detail.
The main content is marked out from the furniture
Check E03
The main content is marked out from the furniture
What it means if this fails
Nothing separates the content from the navigation, so a machine has to guess where the page begins.
What you would do about it
Wrap the main content of each page named in a main element so a machine can tell it from the navigation.

Segment 7 of 11

LLM Optimization

86/1006 of 7 checks passed

Whether assistants are allowed to read you, and find something worth reading.

1 remark to fix here

There is no llms.txt, so assistants have nothing written for them about this business
Check G01
The site publishes an llms.txt
What it means if this fails
There is no llms.txt, so assistants have nothing written for them about this business.
What you would do about it
Publish an llms.txt at the site root describing the business in plain sentences and linking the pages that matter.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
What we found: no /llms.txt.
Fix it yourself
On harpergoods.com: Publish an llms.txt at the site root describing the business in plain sentences and linking the pages that matter. What we found: no /llms.txt.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.

Six checks this site passed

Assistants are allowed to read the site
Check L01
Assistants are allowed to read the site
What it means if this fails
The robots file blocks the crawlers AI assistants use. This is sometimes a deliberate choice, so say what it costs rather than calling it a mistake: nothing on this site can be quoted in an assistant's answer while it stands.
What you would do about it
Decide whether assistants may read this site. If they may, remove the robots.txt rules blocking their crawlers so these pages can be quoted in AI answers.
The site points agents at its machine-readable version
Check M01
The site points agents at its machine-readable version
What it means if this fails
The site publishes a plain-text version for assistants but never mentions it in the response itself, so an assistant only finds it by guessing the address.
What you would do about it
Send a Link header on every page pointing at your llms.txt, so an assistant is told where the plain-text version of this site lives instead of having to guess: Link: </llms.txt>; rel="alternate"; type="text/plain".
An assistant asking for plain text gets plain text
Check M02
An assistant asking for plain text gets plain text
What it means if this fails
An assistant that asks for a plain-text version of a page is sent the full HTML instead, so it has to read a page built for a browser and work out which parts are the content.
What you would do about it
Serve a plain-text version of each page when the request asks for one. If your site is behind Cloudflare, turn on Markdown for Agents in the dashboard. Otherwise set your server to reply with Content-Type: text/markdown and the page's text when a request sends Accept: text/markdown, and keep sending HTML to everyone else.
The API says where it is documented
Check M03
The API says where it is documented
What it means if this fails
This site has an API, but nothing at the standard address tells an agent where to find its documentation, so software cannot discover what the API offers without a person reading the docs first.
What you would do about it
Publish an API catalogue at /.well-known/api-catalog as application/linkset+json, listing each API with links to its specification and its documentation.
An agent can find out how to sign in to the API
Check M04
An agent can find out how to sign in to the API
What it means if this fails
This site has an API behind a login, and nothing at the standard addresses tells an agent how to authenticate, so software has no way to get a token without a person reading the documentation.
What you would do about it
Publish your sign-in details where software looks for them: /.well-known/oauth-authorization-server (or /.well-known/openid-configuration) with your endpoints, and /.well-known/oauth-protected-resource naming which server issues tokens for this API.
Agents are told what this site lets them do
Check M05
Agents are told what this site lets them do
What it means if this fails
This site has an API but publishes no description an assistant can read on its own, so an assistant can talk about the business without being able to do anything with it.
What you would do about it
Publish a server card at /.well-known/mcp/server-card.json describing what an assistant may do here, or a skills index at /.well-known/agent-skills/index.json listing the tasks it can carry out.

Segment 8 of 11

E-E-A-T

78/1007 of 9 checks passed

The signals that say a real business with real people stands behind this.

2 remarks to fix here

What this score does not include. This scores the trust signals your pages carry. Whether readers and search engines find you authoritative is decided off your site, and we do not measure that.

There is no about page, so nothing on the site says who runs this business
Check X01
There is a page saying who is behind this
What it means if this fails
There is no about page, so nothing on the site says who runs this business.
What you would do about it
Add a page saying who runs the business, and link it from the navigation.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
What we found: no about, team or company page among the pages we read or in the sitemap.
Fix it yourself
On harpergoods.com: Add a page saying who runs the business, and link it from the navigation. What we found: no about, team or company page among the pages we read or in the sitemap.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
Articles carry no author, so nothing says who is qualified to have written them
Check X04
Articles say who wrote them
What it means if this fails
Articles carry no author, so nothing says who is qualified to have written them.
What you would do about it
Credit an author on each article, and describe them as a Person in the structured data.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
What we found: 2 of 2 article pages name no author: /journal/first-pour, /journal/winter-scents.
Fix it yourself
On harpergoods.com: Credit an author on each article, and describe them as a Person in the structured data. What we found: 2 of 2 article pages name no author: /journal/first-pour, /journal/winter-scents.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.

Seven checks this site passed

There is a postal address a machine can read
Check N02
There is a postal address a machine can read
What it means if this fails
The address is not in the structured data, so a search engine has to guess it from the page text and often gets it wrong or ignores it.
What you would do about it
Add a "address" property to your business structured data, as a PostalAddress with streetAddress, addressLocality, postalCode and addressCountry filled in.
The structured data says who the business is
Check G04
The structured data says who the business is
What it means if this fails
The structured data never says who this business is, only what the pages are.
What you would do about it
Add an Organization, LocalBusiness or Person block to the structured data so it states who the business is, not only what each page is.
There is a way to make contact
Check X02
There is a way to make contact
What it means if this fails
There is no contact page and no address or number on the site, so there is no way to reach a person.
What you would do about it
Publish a way to reach a person: a contact page, an email or telephone link, or a postal address in the structured data.
The site publishes its legal pages
Check X03
The site publishes its legal pages
What it means if this fails
There is no privacy policy or terms page, which readers and search engines both read as a missing basic.
What you would do about it
Publish a privacy policy and a terms page, and link them from the footer.
The people behind the site are described in machine terms
Check X05
The people behind the site are described in machine terms
What it means if this fails
No structured data describes a person or an organisation, so nothing links this site to anyone real.
What you would do about it
Add Person or Organization structured data naming who is behind the site.
The site shows what customers say about it
Check X06
The site shows what customers say about it
What it means if this fails
There are no testimonials or reviews anywhere, so nothing but the site vouches for the site.
What you would do about it
Publish customer testimonials or case studies, and mark them up as Review structured data.
Content shows when it was last looked at
Check X07
Content shows when it was last looked at
What it means if this fails
Nothing on the site is dated, so a reader cannot tell whether any of it is still current.
What you would do about it
Show a published or updated date on each article.

Segment 9 of 11

Local SEO

100/1006 of 6 checks passed

Whether somebody searching nearby can find you, ring you and get to you.

Nothing to do here.

Six checks this site passed

The site says in machine terms that it is a local business
Check N01
The site says in machine terms that it is a local business
What it means if this fails
Nothing on the site tells a search engine this is a business with a place, so it is never a candidate for the map results that sit above the ordinary ones.
What you would do about it
Add a JSON-LD block to the home page with "@type" set to the kind of business this is — LocalBusiness, or a more specific one like Pharmacy, Restaurant or Dentist — carrying the name, the address, the telephone number and the opening hours.
There is a postal address a machine can read
Check N02
There is a postal address a machine can read
What it means if this fails
The address is not in the structured data, so a search engine has to guess it from the page text and often gets it wrong or ignores it.
What you would do about it
Add a "address" property to your business structured data, as a PostalAddress with streetAddress, addressLocality, postalCode and addressCountry filled in.
Opening hours are published where a machine can read them
Check N03
Opening hours are published where a machine can read them
What it means if this fails
Nowhere says when you are open in a form a search engine can use, so the result for your business cannot show whether somebody can come now.
What you would do about it
Add "openingHours" or "openingHoursSpecification" to your business structured data. The short form looks like "Mo-Fr 09:00-18:00", and it belongs on the page as text as well.
A telephone number can be dialled by tapping it
Check N04
A telephone number can be dialled by tapping it
What it means if this fails
The telephone number is written as text rather than as a link, so somebody reading on a phone has to memorise it and switch apps to ring you.
What you would do about it
Wrap the telephone number in a link: <a href="tel:+301234567890">. Keep the readable spacing in the visible text and put the digits with the country code in the href.
There is a way to get directions
Check N05
There is a way to get directions
What it means if this fails
Nothing on the site links to a map, so somebody who has decided to visit has to copy your address into another app and hope it resolves.
What you would do about it
Put a link to your business on Google Maps next to your address, on the contact page and in the footer. The share link from your own Maps listing is the one to use.
The site says which places it serves
Check N06
The site says which places it serves
What it means if this fails
Nothing says which town or area you cover, so a search for your trade in a nearby place has no reason to return you.
What you would do about it
Add "areaServed" to your business structured data listing the towns or the region you cover, and say the same thing in words on the page: a sentence naming the places you work in.

Segment 10 of 11

Programmatic SEO

Not applicable

Whether you have pages built from a template to cover many searches at once.

Programmatic SEO is about sections built from a template, which starts to matter once a site runs to a few dozen pages. This one is smaller than that, so there is nothing here to score.

Segment 11 of 11

Off-Page SEO

71/1005 of 7 checks passed

What your own site says about your presence everywhere else.

2 remarks to fix here

What this score does not include. We do not count backlinks or press mentions. Nothing in the response your site sends reveals who links to you, so this scores only the off-site presence your own pages declare.

There are no case studies, so nothing on the site shows the work rather than describing it
Check F05
The site publishes work it has done for customers
What it means if this fails
There are no case studies, so nothing on the site shows the work rather than describing it.
What you would do about it
Publish case studies or customer stories showing work already done.
Severity
medium
This is not hurting you badly today, but it gets worse on its own as more people use your app, or it leaves you exposed if something else goes wrong.
Why it matters
Worth doing once the high ones are done. These are the findings that turn into high ones after a few good months of growth.
What we found: no case study, customer story or testimonial page.
Fix it yourself
On harpergoods.com: Publish case studies or customer stories showing work already done. What we found: no case study, customer story or testimonial page.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.
Nothing points a visitor at a place they could read or leave an independent review
Check F03
The site points at somewhere it can be reviewed
What it means if this fails
Nothing points a visitor at a place they could read or leave an independent review.
What you would do about it
Link to wherever this business can be reviewed independently.
Severity
low
Real, but small. Either the effect is minor at your current size, or it only shows up in narrow circumstances.
Why it matters
Fix it if it is quick and you are already in that file. Nothing bad happens if you leave it.
What we found: no links to G2, Capterra, Trustpilot, Product Hunt, Yelp or Google reviews.
Fix it yourself
On harpergoods.com: Link to wherever this business can be reviewed independently. What we found: no links to G2, Capterra, Trustpilot, Product Hunt, Yelp or Google reviews.
Paste this into whatever you built the site with. It names the page and what the result should be, so the builder does not have to guess.

Five checks this site passed

The site shows what customers say about it
Check X06
The site shows what customers say about it
What it means if this fails
There are no testimonials or reviews anywhere, so nothing but the site vouches for the site.
What you would do about it
Publish customer testimonials or case studies, and mark them up as Review structured data.
The site links to its own profiles elsewhere
Check F01
The site links to its own profiles elsewhere
What it means if this fails
The site links to no profile anywhere else, so nothing connects it to the business's presence online.
What you would do about it
Link the business's own profiles elsewhere from the site, in the footer.
Structured data claims those profiles as the same business
Check F02
Structured data claims those profiles as the same business
What it means if this fails
The site links to its profiles but never declares they are the same business, so search engines must guess.
What you would do about it
List those profile addresses in the structured data using sameAs, so search engines connect them to this business.
The site shows where it has been covered
Check F04
The site shows where it has been covered
What it means if this fails
There is no press or coverage page, so nothing on the site shows it has been written about elsewhere.
What you would do about it
Add a page collecting where the business has been written about.
Outbound links credit the sources they use
Check F06
Outbound links credit the sources they use
What it means if this fails
The site links out to nothing at all, which reads as a closed page rather than part of a subject.
What you would do about it
Link out to the sources and tools the content references, where it helps the reader.

Segment	Score	Passed	Status	Priority
Technical SEOTechnical	82	14 of 17	Good	High
Page SpeedSpeed	100	7 of 7	Good	Low
On-Page SEOOn-Page	74	14 of 19	Needs work	Do first
Generative Engine OptimizationGEO	79	11 of 14	Needs work	Do first
AI OverviewsAIO	56	5 of 9	Needs major work	Do first
Answer Engine OptimizationAEO	38	3 of 8	Needs major work	High
LLM OptimizationLLMO	86	6 of 7	Good	Low
E-E-A-TE-E-A-T	78	7 of 9	Needs work	Medium
Local SEOLocal	100	6 of 6	Good	Low
Programmatic SEOpSEO	—	—	Not applicable	—
Off-Page SEOOff-Page	71	5 of 7	Needs work	Medium

How your site scored

What to do, in order

Do these first

Then these

When you have time

14 checks this site passed

Seven checks this site passed

14 checks this site passed

Eleven checks this site passed

Five checks this site passed

Three checks this site passed

Six checks this site passed