How to show a client site errors and broken pages

By Ranjit Bhinge, founder of Blur Studio · Updated · 9 min read

Short answer: A client-ready error report has three parts: a list of broken pages found, a label on each saying fix, redirect or ignore, and a plain sentence on what you did. You can build it free from Google Search Console, Google's free tool that shows how its search engine sees your site, plus a crawl of your own links with a tool such as Screaming Frog, a desktop program that follows every link on a site. Google itself says 404 errors are not necessarily a problem, so a report that lists every one in red will scare the client for no reason.

Ranjit Bhinge

You need to show a client proof of what is broken on their site and what you fixed, in a form they will actually read.

I'm Ranjit, and I run Blur Studio: 15+ years building software, 120 apps taken live, and clients who have gone on to raise $10M+. I build AI systems and processes that help businesses do a lot more with a lot less time. I also work differently from most of what else is out there, more on that further down this page. Say hello.

The short method: pull the "Not found (404)" and "Soft 404" lists from Search Console, crawl the site for broken internal links, decide for each URL whether it needs a redirect, a content fix or nothing, and put the result on one page. Do it monthly so the client sees the trend, not just a scary list on the day something breaks.

This page is written from an agency or freelancer view, but it works just as well if you are the owner explaining your own site to a boss or partner. It covers what each error type means, where to find it, a decision table that tells you which errors matter, a client-readable template, and what a hosting platform's logs can and cannot add. For the repair side of one common cause, see how to redirect old links after a redesign.

This page may contain affiliate links from which Blur Studio earns commissions.

The error types, in plain words

An HTTP status code is a three-digit number a web server sends back with every page, saying how the request went. Clients do not need the numbers, but you do.

What the client seesCodeWhat it meansDoes it need fixing
"Page not found"404The address does not exist (any more)Only if you link to it, list it in a sitemap, or it had visitors and has a close replacement
Friendly "not found" page that loads as normalSoft 404 (the server sends a normal 200 code)Google sees an empty or error-style page that claims to be fineYes: return a real 404, or add real content if the page should exist
Page jumps to another address301 or 308 (permanent), 302, 303 or 307 (temporary)A redirectUsually fine; check long chains
"Something went wrong"500 rangeThe server failedYes, urgently, and tell the host
Form or checkout fails with no messageOften noneNot a status code problem at allYes. See the forms not sending checklist

What Google says, as of November 2026, in the Search Console help page for the page indexing report: 404 responses "are not necessarily a problem, if the page has been removed without any replacement". Fix the 404s you link to yourself or list in a sitemap, and if a page moved, send a 301 to the new location. A soft 404 is a friendly "not found" message that returns a normal response code, and Google's fix is to return a real 404 for pages that do not exist or add real content to pages that should. Google's crawl documentation also says all 4xx codes except 429 are treated alike, indexed URLs returning 4xx are removed from the index, and Google follows up to 10 redirect hops. Temporary redirects (302, 303, 307) are a weaker signal than permanent ones (301, 308).

So do not tell a client that "404s hurt your rankings site-wide". That is not what Google says. Tell them which ones cost them leads.

Where to find the errors

Search Console: what Google found

In Search Console, open Pages (the page indexing report) and look at the reasons list. The ones to export for a client are "Not found (404)", "Soft 404", "Page with redirect" (usually no action, since Google says only the redirect target is indexed) and "Crawled - currently not indexed" (common, and it matters only for pages you care about). Click a reason, then a URL, and use URL Inspection to see where Google found the link. That referrer tells you whether the fault is your own link or someone else's. Use "Validate fix" after you repair a group. The report has limits: it reflects Google's crawl, not every visitor, and it is not instant.

A crawler: what a visitor would hit

Run a crawler across the site to find broken links Google may not have reached yet. Screaming Frog's free version crawls up to 500 URLs, which covers most small business sites, and its paid licence is listed on its pricing page in pounds per year (check the vendor page for the current figure). Export the "client error 4xx" list and the "inlinks" for each, so you know which page holds the bad link.

Server logs or host reports: what really happened

A server log lists every request and its status code, including requests from people who followed an old link or a bookmark. This is the only place you see broken-link hits that came from outside Google. Ask the host for 30 days of logs, filter to status 404 and 500, then count by URL. Many hosts keep only a short window, so save the counts monthly. A log also includes bot requests (probes for admin pages, for example), which are usually noise. Sort by URL and ignore the ones that never existed on your site, such as login paths for software you do not use.

Your platform's own error view

Some website platforms have a built-in 404 report. If yours does, use it, and note that its window and retention are up to the vendor. On Ploy, an AI website builder and host, Ploy's analytics feature page says it logs requests at the edge (the servers closest to the visitor), and its analytics docs describe a Bots card for AI crawlers and agent visits. Ploy's docs do not describe an error or 404 report, or a log retention period, so check with Ploy before promising a client one. Ploy's docs say its routing rules can set redirects (the default is a permanent 301; check the docs for other types), and its migration guidance says to contact support before migration day if you need more than 20 rules.

Decision table: fix, redirect or ignore

SituationActionWhat to tell the client
URL had visitors or links and a close replacement page exists301 redirect to that replacement"Old address now forwards to the new page."
URL is linked from your own pages or sitemapFix the link, or remove it from the sitemap"We corrected 12 internal links."
URL was removed on purpose and has no replacementLeave as a real 404, or 410 if you want to say it is gone for good"This page was retired. No action needed."
URL never existed (typos, bot probes)Ignore"Random addresses nobody uses."
Soft 404Return a real 404 status, or add real content"The error page told Google the page was fine. We fixed the signal."
Redirect chain over 3 hopsPoint the first address straight to the final page"We shortened the route."
Many old URLs sent to the home pageRedirect each to the closest page instead"Old pages now forward to the matching new ones."

One caution: do not send every old address to the home page. Google's site-move guide says not to redirect many old pages to one irrelevant page such as the new home page. If a page has no close replacement, a 404 is the right answer.

The one-page client template

Copy this into a document or email. Keep it to one page.

  1. Header: client name, site, period (for example, 1 to 30 November), date.
  2. Summary sentence: "X broken addresses turned up. Y needed action, Z did not. All Y are fixed or scheduled."
  3. Table: columns "Page (plain name)", "Problem", "Visitors affected", "What we did", "Status".
  4. Server errors: "Number of 500 errors this month: N. Dates and cause if known."
  5. Forms and leads: "Test form sent on [date], arrived in [inbox] at [time]."
  6. Next month: two lines on what you will recheck.

Label each row the way the client says it ("Pricing page", not "/p/3829"). Show counts, not code. If you have many items, group by cause: "old links after the redesign", "typos in the menu", "pages retired on purpose".

A worked example with made-up numbers

A roofing company's site has 140 pages. After a redesign in October, the agency runs the checks in November.

SourceFound (made up)Result after sorting
Search Console, Not found (404)46 URLs31 old service pages with a close match: redirect. 9 retired: leave. 6 never existed: ignore
Search Console, Soft 4044 URLsAll 4 were empty city pages: add real content to 2, remove 2
Crawl of the site18 broken internal linksAll fixed in the menu and footer
Server log, 500 errors7 on one dayHost confirmed a plugin crash, rolled back

The report to the client says: "68 problems turned up: 46 addresses that gave a not-found error, 4 pages that looked fine to Google but were empty, and 18 broken links on our own pages. 31 old addresses now forward to the right page, 18 links on the site were corrected, 2 thin pages were filled out and 2 were removed. 15 retired or never-used addresses need no action. One short outage on 12 November was caused by a plugin and was rolled back." That is plain, short and checkable. It makes no claim that rankings or leads will rise.

Test plan with pass and fail rules

  1. Collect. Pass: you have Search Console lists, a crawl export and 30 days of status counts. Fail: the host keeps no logs, so ask for them or set a reminder to save monthly.
  2. Sort every URL. Pass: each row has one of fix, redirect, leave or ignore. Fail: any row is "unknown". Inspect it first.
  3. Fix and retest. Pass: after each redirect, the old address loads the new page in a private browser window and returns a permanent redirect. Fail: you see a chain or loop, so correct the rule.
  4. Validate in Search Console. Pass: the validation starts and the count falls over the following weeks. Fail: it fails, so inspect a sample URL for the remaining cause. Google does not give a fixed timeline.
  5. Forms. Pass: a test submission reaches the real inbox. Fail: it does not. Do not send the report until you have fixed it.
  6. Client read test. Pass: the client can say in one sentence what happened. Fail: they ask what a 404 is, so rewrite with plain names.

When another approach is better

  • If the client needs uptime numbers or a contractual service level, use a dedicated uptime monitor and the host's status page. Ploy's docs do not publish a full uptime or SLA figure, so see Ploy hosting speed and uptime and ask for the terms in writing.
  • If the problem is leads that never arrive, the status codes will not show it. Use the forms checklist.
  • If the site has under 20 pages, a manual click-through once a month may be enough.
  • If a hosting platform's log has no 404 view and you cannot get raw logs, Search Console plus a crawler still covers most of the need.

How I'd approach this

I would run Search Console and a crawl on the same day each month, spend 30 minutes sorting, and send the one-page template the same week. I would put the redirects in place before I tell the client anything, so the report ends with "fixed" rather than "found". On a platform like Ploy, I would use its routing rules for the redirects and its analytics for traffic, and I would still keep Search Console and the crawl for the error list. If you want ongoing care, see keeping a website healthy after launch, and if you want a custom log dashboard or someone to repair what the logs reveal, Blur Studio builds that, see what a custom build includes.

First-week plan. Day 1: export the four Search Console reasons and run the crawl. Day 2: sort every URL with the decision table. Day 3: set the redirects and fix internal links. Day 4: send a test form and check the inbox. Day 5: click Validate fix in Search Console and write the one-page report. Day 7: send it to the client and book the same slot next month.

For the next step, see Website Heatmaps: How to See Why Visitors Leave.

Related guide: A Site Health Audit You Can Hand to a Client.

Questions people ask

Do 404 errors hurt my SEO?

Google says 404 responses are not necessarily a problem if a page was removed with no replacement. They matter when you link to the page yourself, list it in a sitemap, or it had visitors and links and has a close replacement. Check Google's page indexing help for current wording.

What is a soft 404?

A soft 404 is a page that looks like a not-found page to visitors but sends a normal success code. Google's fix is to return a real 404 for pages that do not exist, or add real content to pages that should exist.

How do I find broken pages without a developer?

Use Search Console's Pages report for what Google found and a free crawl with Screaming Frog (up to 500 URLs on the free version) for broken internal links. For a one-off fix of old links, see how to redirect old links after a redesign.

Should I redirect every 404 to the home page?

No. Google's site-move guide advises against redirecting many old pages to one irrelevant page such as the home page. Redirect to the closest matching page, or leave a real 404 where there is no match.

How long does it take Google to clear fixed errors?

Google gives no fixed timeline. Use the Validate fix button in Search Console and check back over the following weeks. Googlebot may keep trying known URLs for a while.

Can Ploy produce an error report for my client?

Ploy's feature page says it logs requests at the edge, and its analytics docs describe a Bots card. The docs do not describe an error or 404 report or a log retention period, so ask Ploy before you promise one. For plans see the Ploy pricing page.

How often should I send an error report?

Monthly is enough for most small sites, and sooner after a redesign or move. Include forms: send a test submission and note where it arrived, as in the forms not sending checklist.

What if the host will not give me logs?

Use Search Console and a crawler, which need no host access, and set a monthly reminder to save any status counts the host does show. If logs matter for a client contract, ask for them in writing before you sign.

More in this topic: Websites: simple stats and reports for owners and clients (9 guides) →

Sources

Ready for the next step?

A 30 minute call. If you do not need me, I will tell you.

Ranjit Bhinge You talk to me, Ranjit. I reply within a few hours.

Not sure what you need? Tell me what is going on. I will point you the right way, even if it is not me.

Blur Studio · 30 minute call

Book a 30 minute call

You talk to me, Ranjit. If you do not need me, I will tell you.

  • We talk about your business and what is going on.
  • You leave with a clear next step, even if it is not me.
  • I reply to every message within a few hours.
Loading the calendar…