Guidde SEO Teardown Update: 3,197 Pages, 4,405 Broken Links, and 1,586 Orphaned Pages

Originally published from a March 8, 2026 crawl. Updated with pages, broken-links, orphan-pages, and internal-link recommendations exported on August 7, 2026. Analysis powered by redCacti.

Guidde has built a large content footprint around video documentation, product comparisons, playbooks, integrations, and educational content.

Our March crawl found 2,055 URLs. Five months later, the August pages export contained 3,197 crawled URLs, including 3,148 successful pages and 47 URLs returning HTTP 404.

Growth, however, is only one part of the story.

The August crawl also found:

• 4,405 broken-link records
• 1,586 successful pages with no incoming internal links
• 8,945 potential internal-link recommendations in the separate recommendations export
• 2,090 successful pages without a canonical URL
• 1,531 successful pages without a meta description
• 48,625 images without alt text across successful pages

The useful lesson is not that a large site has a long list of errors. It is that different problems require different units of analysis. Broken links should be grouped by destination. Orphan pages should be evaluated by section and business value. Metadata gaps should be fixed through templates where possible.

This teardown shows how those choices change the work.

August 2026 Crawl Summary

The pages export contained seven older rows last crawled in March. To keep the August sitewide analysis consistent, the table above uses the 3,197 rows crawled on August 7. The orphan count includes only successful August pages.

What Changed Since March

The successful page count increased by 56.7%. Broken-link records increased by 35.7%.

Looking only at the larger broken-link total would therefore give an incomplete picture. Broken-link records per crawled URL fell from about 1.58 in March to 1.38 in August. At the same time, the share of successful pages classified as orphaned increased from 22.3% to 50.4%.

These are crawl findings, not Google index or traffic data. They show what the crawler could reach and how the discovered pages connect to one another. They do not prove that every orphaned page is indexed, valuable, or intended to rank.

Finding 1: The Site Added Pages Faster Than It Added Internal Connections

The largest August section was /knowledge-hub/, with 1,490 successful pages. Of those, 1,381 had no incoming internal links.

That means 92.7% of the section was classified as orphaned in the crawl.

This is particularly important because these are not mostly thin pages. The median /knowledge-hub/ page contained 5,852 words. The section also accounted for 5,293 potential internal-link recommendations.

Publishing more content does not automatically create a stronger search footprint. A page can be comprehensive and still receive little support if no relevant page links to it.

What a reader should do

Do not start by linking all 1,381 pages. First divide them into four groups:

1. Pages targeting terms that matter to the business
2. Pages already receiving impressions, clicks, or backlinks
3. Pages that overlap heavily with another page
4. Pages that are outdated, low value, or not intended for search

Add internal links to the first two groups. Consolidate genuine overlap. Archive or exclude pages that should not compete for organic traffic.

The value of an orphan report is not the count. It is the prioritization queue it creates.

Finding 2: Five Broken Destinations Produce 91.6% of the Records

The August broken-link export contained 4,405 records, but only 128 unique broken destinations.

Five destinations produced 4,035 records, or 91.6% of the entire export.

The /old-home-3 destination appeared 438 times in March and 1,578 times in August. Its increase of 1,140 occurrences accounts numerically for 98.4% of the net increase in broken-link records.

This does not mean every record came from one sitewide template. It means the recurrence is concentrated enough to investigate shared components, content-generation rules, or repeated authoring patterns before creating page-level tickets.

What a reader should do

Group a broken-link export by destination and count affected source pages. Then inspect a small sample of source pages for the leading destination.

• If the link appears in the same navigation or reusable component, fix the component once.
• If the link was generated by a content template, fix the generation rule and repair existing output.
• If the destination moved, point links directly to the correct page.
• If a retired page has a true replacement, use a targeted redirect.
• If there is no replacement, remove the link.

After deployment, crawl again. A completed engineering task is not the same as a verified SEO fix.

Finding 3: Broken Conversion URLs Deserve More Urgency Than Editorial 404s

The list of 47 URLs returning HTTP 404 included paths such as /signup, /sign-up, /start-free, /try-free, /free-trial, /request-demo, and /book-demo.

Some of these may be intentionally retired variants. Their presence still matters because a broken link near a conversion point can cost more than a broken citation inside an old article.

The broken-link export recorded 39 occurrences of /signup across 24 source pages. That is much smaller than the /release-notes count, but its commercial importance may be higher.

What a reader should do

Use two scores when prioritizing broken links:

1. Scale: how many pages or users encounter the problem
2. Business impact: whether the link affects signup, demo, pricing, product education, or another important journey

A frequency-only queue can place a high-value conversion problem below a harmless high-volume link.

Finding 4: Metadata Gaps Follow Section Patterns

Across the 3,148 successful August pages:

• 2,090 had no canonical URL
• 1,531 had no meta description
• 3,061 had no detected schema type
• 2,042 had no Open Graph title
• 2,042 had no Twitter title

These gaps are not distributed evenly.

This pattern points toward template-level work. It would be inefficient to write 1,489 separate tickets for knowledge-hub canonicals if one template controls the field.

Not every page needs rich structured data, and a missing schema type is not automatically an SEO defect. Canonicals also require context before implementation. The practical takeaway is to inspect templates and indexability rules, not blindly fill every empty field.

What a reader should do

Create a section-level matrix with four questions:

1. Is this page type intended to appear in search?
2. Does it have a self-referencing canonical or another deliberate canonical target?
3. Does it have unique search-facing title and description fields?
4. Is there a valid schema type that accurately describes the page?

Fix the template used by valuable, indexable sections first. Validate a sample before rolling the change across thousands of URLs.

Finding 5: Image Alt Work Should Be Prioritized by Page Type

The August successful pages contained 59,212 images. Of these, 48,625 had no alt text, producing aggregate alt coverage of 17.9%.

The coverage also varied by section:

• /blog/: 4.9%
• /playbooks/: 5.1%
• /knowledge-hub/: 21.8%
• /tool-comparison/: 21.0%
• /gallery/: 42.7%

Adding alt text to every image is not the right instruction. Decorative images should generally use empty alt attributes. Informative screenshots and diagrams need descriptions that communicate their function or content.

What a reader should do

Start with image-heavy, high-traffic templates and classify images as decorative, linked, or informative. Fix the component for recurring decorative and linked images. Write useful alt text for informative visuals that help readers understand the page.

Do not use filenames or keyword lists as alt text.

Finding 6: Internal-Link Suggestions Need Editorial Filters

The recommendations export contained 8,945 potential links across 1,818 source pages and 1,589 target pages. Of those recommendations, 6,962 had a similarity score of at least 80%.

That is a useful discovery queue, not an instruction to insert 8,945 links.

High textual similarity can identify topical relationships, but it does not know whether a link helps the reader, whether the target competes with the source, or whether the same destination is already overused.

What a reader should do

Prioritize recommendations where all four conditions are true:

1. The target page has clear business or search value.
2. The source and target satisfy different search needs.
3. The proposed anchor fits naturally in the source paragraph.
4. The link helps a reader take a logical next step.

redCacti recommends internal links. It does not insert them automatically, so every suggestion can be reviewed before implementation.

What Guidde Is Doing Well

The crawl is not only a list of gaps.

• All 3,148 successful August pages had a title.
• The site has substantial educational depth, including 2,337 successful pages with at least 1,000 words.
• The blog had meta descriptions on all 372 successful pages and canonicals on 370.
• Only 22 of 553 tool-comparison pages were orphaned.
• All 75 playbook pages had detected schema, canonicals, and meta descriptions.

These strengths matter because they show where established templates are already working. The goal should be to extend reliable patterns to weaker sections, not apply one generic SEO checklist to every URL.

Recommended Order of Work

1. Protect conversion journeys

Audit links pointing to signup, trial, demo, pricing, and other commercial destinations. Fix broken paths and standardize the preferred URL variants.

2. Repair the repeated broken-link sources

Investigate /release-notes, /old-home-3, and the three malformed /knowledge-hubS/ destinations. Together they represent 91.6% of broken-link records.

3. Triage the orphaned knowledge hub

Use search demand, Search Console data, backlinks, conversions, freshness, and content overlap to select the pages worth reconnecting. Do not create links to every orphan merely to reduce the count.

4. Fix indexation fields by template

Start with knowledge-hub and tool-comparison canonicals. Confirm the intended indexation behavior before deployment.

5. Review the best internal-link recommendations

Apply business and editorial filters, then add links where they improve navigation and topical context.

6. Improve image accessibility through components

Prioritize informative screenshots and repeated image components on important templates.

7. Schedule a verification crawl

Track resolved and newly introduced issues. Compare both totals and normalized measures such as broken records per crawled page and orphan rate.

Seven Transferable Lessons for Large Content Sites

1. Track the denominator. A larger crawl can produce more errors even when the error density improves.
2. Group broken links by destination. Thousands of rows can originate from a handful of shared causes.
3. Add business impact to technical priority. A broken signup link can matter more than a higher-volume editorial 404.
4. Treat publishing and discoverability as separate steps. Content without incoming links may remain isolated regardless of its length.
5. Fix recurring fields at the template level. Section patterns are usually more actionable than sitewide averages.
6. Use recommendation tools to create a review queue. Similarity is evidence of relevance, not permission to add a link.
7. Recrawl after implementation. Verification is the only way to know whether the issue was actually removed.

These lessons are valuable because each one changes a decision. They help a team decide what to fix, where to fix it, and how to measure whether the work succeeded.

Methodology and Limitations

This analysis used redCacti exports from March 8 and August 7, 2026.

The August inputs were:

• Pages export: 3,197 rows crawled on August 7
• Broken-links export: 4,405 records
• Orphan-pages export: 1,593 records, including 1,587 last crawled on August 7
• Internal-link recommendations export: 8,945 records

Sitewide August percentages use the 3,148 successful pages in the August pages export. The orphan count of 1,586 refers to successful pages marked as orphaned in that same export. The separate orphan export includes one additional August record and older retained rows, so its raw total should not be used as the denominator.

The March comparison uses the previously analyzed March export of 2,055 crawled URLs, including 2,009 successful pages.

The crawl reports technical observations at a point in time. It does not include Guidde’s analytics, Search Console data, backlink data, conversion data, private sitemaps, or internal content strategy. A crawler’s orphan classification means no incoming internal link was found within the crawl scope. It does not prove that a URL is absent from XML sitemaps, external links, or Google’s index.

Website responses and content may have changed after the crawl date.

Run the Same Analysis on Your Website

redCacti can crawl a site, group broken links, identify pages without incoming internal links, and recommend contextually relevant internal-link opportunities.

It does not change your website automatically. You decide which fixes and links to implement.

Add your website and run a free crawl

Already monitoring a growing site? Learn how scheduled broken-link monitoring and the orphan page finder help turn recurring crawl data into a prioritized work queue.