Technical SEO vs on-page SEO: which one is costing you organic conversions?
Organic traffic has stalled, or it arrives and leaves without acting, and the next budget call is whether to fix the site’s infrastructure or its pages. How to tell from your own data which layer is failing, what each one controls today, and how to order the fixes by conversions.
- Start from the symptom, not the discipline. Whether a page is missing from the index, passed over on the results page or abandoned after the click tells you which layer to fix.
- Indexing is a hard gate. A commercial page Google has not indexed earns nothing, and the same rule decides eligibility for AI Overviews and AI Mode.
- Speed matters, relevance matters more. Google ranks the most relevant page even with sub-par page experience, but a 0.1 second mobile gain went with 8.4% more retail conversions in a Deloitte study.
- Order the work by customers lost. Estimate each fix in conversions per organic visit on the page types it touches. The biggest gain is often after the click, on neither SEO list.
Ask for an SEO audit and you get two long lists, one technical and one on-page, each sorted by severity and neither sorted by what it costs you in customers. The two disciplines answer different questions. Technical SEO decides whether search engines can crawl, render and index a page, and whether it loads and responds fast enough. On-page SEO decides whether the page matches what the searcher wanted and earns the click. Which one to fix first depends on where your organic visitors drop out, and Search Console and your analytics already show that.
There is a third layer most comparisons leave out. Once a visitor lands, what they see decides whether the visit becomes a demo request, a purchase or a booking. Neither SEO list covers it, and in the worked example later in this article it is where the largest gain sits.
The conversion event in that example is the demo request. If you run an online store, read purchase and revenue per visitor instead; the table in section 5 gives the event for each type of business. Off-page SEO, mainly links from other sites, is a separate discipline and outside the scope here.
What each one controls
Technical SEO works on the site: how pages are crawled, rendered and indexed, how fast they load and respond, whether they work on a phone and over HTTPS, and whether structured data describes them correctly. On-page SEO works on each page: the content and its match to the query, the title, the meta description, the headings and the images.
Crawling, rendering, indexing, speed, mobile, HTTPS, structured data
Shows in: Page indexing and Core Web Vitals reportsIntent match, content, title, meta description, headings, images
Shows in: impressions and click-through by queryThe offer, proof and next step the visitor sees after landing
Shows in: conversions per organic visit
The elements in each list are the same ones this article covered when it was first published. What is new is the column on the right: the symptom each one produces when it fails, which is how you find it without running a full audit.
| Element | Layer | What it decides | Symptom when it fails |
|---|---|---|---|
| Crawl and index controls (robots.txt, noindex, canonical) | Technical | Whether the page can appear at all | Page missing from the Page indexing report |
| Site speed and Core Web Vitals | Technical | Whether the page loads and responds fast enough | Engagement gap concentrated on mobile |
| Mobile rendering | Technical | What Google indexes, since it indexes the mobile version | Content visible on desktop that does not rank |
| HTTPS and accessibility | Technical | Browser trust and use by people with assistive technology | Security warnings, errors in the HTTPS report |
| Structured data | Technical | Eligibility for rich results, not a guarantee | Errors in the rich result reports |
| Content and intent match | On-page | Whether the page answers the query that reached it | Impressions for the wrong queries, or none for the right ones |
| Title element and meta description | On-page | How the result reads on the results page | Click-through below your own average at that position |
| Headings and structure | On-page | How fast readers and systems find the answer | Short visits on pages that rank well |
| Images and alt text | On-page | Context for image search and screen readers | No image search traffic, accessibility errors |
Some facts inside the lists have changed since then. Google finished moving to mobile-first indexing in October 2023, and Interaction to Next Paint replaced First Input Delay as a Core Web Vital in March 2024. Section 3 has both.
Find the failing layer from the symptom
Each layer fails in its own way, and the failure is visible before anyone opens an audit tool. Work through the questions below for the pages that matter commercially, such as pricing, product, comparison and category pages, not for the whole site.
- Is the page indexed? Check the Page indexing report and URL inspection in Search Console. “Crawled, currently not indexed” or “Excluded by noindex” on a pricing or comparison page is a technical problem, and no change to the copy will fix it.
- Does it earn impressions for the queries it targets? Filter the Performance report to the page. An indexed page with no impressions for its target queries usually has an intent problem: the content answers a different question. It can also lack the links to compete, which is off-page work.
- Is the click-through rate in line with your own pages at that position? Compare against your own site, not a published curve. Results pages with AI Overviews, ads or shopping units lower click-through for everyone on them.
- Do visitors stay, on mobile as well as desktop? Compare engaged sessions by device for the same page. A gap concentrated on mobile points to speed, responsiveness or layout. Low engagement on both devices points to a mismatch between the query and the page.
- Do engaged visitors convert? If they read and still do not act, the problem sits after SEO: the offer, the proof or the call to action does not fit what the visit came for.
A sitewide drop. When organic traffic falls across many pages at once, check for a core update, manual actions and recent site changes before diagnosing page by page. The article on why organic traffic drops covers that case.
When technical SEO is the bottleneck
Technical problems are absolute at the top and gradual at the bottom. A page that is not indexed earns nothing, however good it is. A page that loads slowly still ranks and still converts, only less.
Google’s page experience documentation, updated on 22 September 2026, sets the order of priority: “Google Search always seeks to show the most relevant content, even if the page experience is sub-par.” Speed and stability help where many pages are equally relevant. They do not replace relevance.
The thresholds Google publishes for a good experience are a Largest Contentful Paint within 2.5 seconds, an Interaction to Next Paint under 200 milliseconds and a Cumulative Layout Shift under 0.1. INP replaced First Input Delay as a Core Web Vital on 12 March 2024, so an audit that still reports FID is measuring the wrong thing.
Mobile is no longer a separate check. Google announced on 31 October 2023 that the move to mobile-first indexing was complete. The version Google indexes is the one a phone sees, so content hidden or missing on mobile is, for ranking purposes, missing.
The conversion case for speed comes from Milliseconds make millions, a study commissioned by Google and run by 55 and Deloitte across 37 European and American brand sites. A 0.1 second improvement in mobile speed metrics went with conversion rates 8.4% higher on retail sites and 10.1% higher on travel sites. On lead generation sites, 21.6% more users progressed to the form submission page.
The data covers 30 days at the end of 2019, on mobile only, with speed metrics Google has since replaced, such as First Meaningful Paint. It gives the direction of the effect. The size on your site comes from comparing the pages you fix before and after, or better, from a holdout.
Structured data sits on the technical list with a caveat Google states in its structured data guidelines, updated 10 July 2026: correct markup makes a rich result possible, not certain. Treat it as eligibility, not as a lever you can plan click-through on.
The same technical gate now applies to AI search. Google’s guide to generative AI features, updated 10 July 2026, says a page must be indexed and eligible to be shown with a snippet to appear in those features, and that no special files or markup are needed. A page that fails the first question in section 2 is invisible to AI Overviews and AI Mode as well.
When on-page SEO is the bottleneck
On-page problems rarely stop a page from ranking altogether. They show up as the wrong traffic, or as the right traffic that scrolls past the result.
The title element and meta description shape the result, but Google writes the final version. Its documentation on snippets, updated 20 April 2026, says snippets “are primarily created from the page content itself”, and that Google sometimes uses the meta description when it describes the page more accurately. A clear opening paragraph does more for the result than a polished meta description on a page that buries the answer.
Compare plans side by side, see what each includes, and book a demo to get a quote for your team size.
- Title linkusually from the title element
- Snippetprimarily from page content, sometimes the meta description
The visitor sees what the result said. Next step: book a demo for a quote.
Prices are three scrolls down. The click counted; the visit is lost.
The larger on-page failure is intent. A page that ranks for a commercial query with an informational article, or for an informational query with a sales page, earns clicks it cannot serve. Read the queries the page actually earns impressions for, group them by intent, and check that the page type fits the largest group.
| Query wording | Intent | Page that fits | Realistic next step |
|---|---|---|---|
| “what is”, “how to” | Informational | Guide or explainer | A related page, a report, a return visit |
| “X vs Y”, “X alternatives”, “best X for” | Commercial | Comparison or alternatives page | Demo, trial or product page |
| “X pricing”, “X demo”, “buy X” | Transactional | Pricing, product or booking page | Demo request, purchase or booking |
Common mistake: rewriting titles to lift click-through with a promise the page does not keep. Click-through rises, engagement and conversions fall, and the gain disappears from the report that matters. The title has to describe the page the visitor lands on.
Rank the fixes by conversions, not by audit score
An audit ranks problems by how many URLs they touch. Your budget should rank them by how many customers they cost. That takes one number per page type: conversions per organic visit, here demo requests divided by organic visits.
In this example, comparison and pricing pages take 15.0% of organic visits and produce 72.1% of demo requests. A fix that touches them is worth more than a fix that touches the blog, whatever its severity score. Three issues on the same site, estimated the same way:
Indexing, technical. 18 alternatives pages are “crawled, currently not indexed”. The 20 indexed comparison pages average 250 organic visits a month at 1.2%. Assume the new ones earn half that: 18 × 125 = 2,250 visits, × 1.2% = 27 demo requests a month.
Snippet, on-page. The pricing page sits at position 3.1 with 24,000 impressions and a 1.5% click-through rate: 360 clicks. The site’s own pages at positions 3 to 4 average 4.0%. At that rate it gets 960 clicks, 600 more, × 3.0% = 18 demo requests.
Conversion layer, after SEO. Comparison pages convert at 1.2%. The best one on the site converts at 1.8%. Lifting the group to 1.8% on 5,000 visits gives 90 demo requests instead of 60, 30 more.
The audit’s longest item, 1,900 blog images without alt text, has no conversion estimate at all. It is worth fixing for accessibility and image search, in routine maintenance, not first. The largest estimate, the conversion layer, appears on neither SEO list. This is the same reading used to find organic pages that rank and do not convert, applied to the order of the work.
An audit ranks problems by how many URLs they touch. Your budget should rank them by how many customers they cost.
The event you divide by depends on the business. Rank page types by one event at a time, never by a sum of them.
| Business | Divide organic visits into | Keep separate from |
|---|---|---|
| E-commerce | Revenue (revenue per organic visitor) | Add-to-cart events, newsletter sign-ups |
| B2B SaaS | Qualified demo or trial requests | Content downloads |
| Travel | Bookings | Brochure or quote requests |
| Healthcare and clinics | Booked appointments | Phone number clicks |
| Education | Enrollment enquiries | Prospectus downloads |
Page types with a few hundred organic visits a month or fewer give rates too noisy to rank by. On a small site, fix indexing and obvious technical faults on commercial pages first, and switch to conversion-led ordering once each page type has enough visits to compare.
How Pathmonk runs both layers and the one after them
Pathmonk SEO runs the technical and on-page work continuously from each site’s own first-party data. The SEO case study page describes the technical layer in one line: “Meta descriptions, Core Web Vitals, broken links and thin pages are fixed automatically.” Every draft also goes through an on-page check scored out of 100, with each finding marked OK or FIX and the correction next to it.
Across Pathmonk SEO customers, the average lift was 35% in organic traffic and 63% in organic impressions in the first three months, measured in their own Search Console and Pathmonk data. Those are acquisition figures: they say more of the right people arrive, not what they do next.
For the layer after the click, Pathmonk personalization adapts the page a visitor lands on in real time to the source of the visit, search included, so it opens on the promise the visit was made with. The lift is measured on a 50/50 control on your own traffic. The hub article on combining SEO and CRO covers how that works for organic pages.
Key takeaways
- Technical SEO decides whether a page can be found and how well it loads. On-page SEO decides whether it matches the query and earns the click.
- Diagnose by symptom: not indexed, no impressions, low click-through, low engagement or no conversions. Each points to one layer.
- A page that is not indexed earns nothing, and no on-page work fixes it. The same gate applies to Google’s AI features.
- Relevance comes before page experience. Google shows the most relevant content even when its page experience is sub-par.
- Good Core Web Vitals: LCP within 2.5 s, INP under 200 ms, CLS under 0.1. INP replaced FID in March 2024.
- Snippets come primarily from page content, so the opening of the page matters more than the meta description.
- Order fixes by conversions per organic visit on the page types they touch, not by how many URLs an audit flags.
- The conversion layer after the click is on neither SEO list, and it is often where the largest gain sits.
FAQs on technical and on-page SEO
What is the difference between technical SEO and on-page SEO?
Technical SEO covers the site’s infrastructure: crawling, rendering, indexing, speed, mobile rendering, HTTPS and structured data. On-page SEO covers each page’s content and HTML elements: how well the content matches the query, the title, the meta description, the headings and the images. Technical work makes a page eligible to rank; on-page work makes it the right result for a query.
Which matters more, technical SEO or on-page SEO?
Neither in general. It depends on where your pages fail. If commercial pages are not indexed, technical work comes first, because nothing else can help them. If they are indexed but earn the wrong queries or few clicks, on-page work comes first. Order the fixes by the conversions they are likely to recover.
Is page speed a ranking factor?
Google says its core ranking systems look to reward a good page experience and recommends good Core Web Vitals: LCP within 2.5 seconds, INP under 200 milliseconds and CLS under 0.1. It also says it shows the most relevant content even when page experience is sub-par. Speed helps when many pages are equally relevant, and it affects conversions whatever the ranking.
Do meta descriptions still matter?
They affect how a result reads, not whether the page is relevant. Google says snippets are primarily created from page content and that it sometimes uses the meta description when it describes the page more accurately. Write the meta description for the click, and make the opening of the page answer the query.
Does structured data improve rankings?
Structured data makes a page eligible for rich results. Google’s guidelines state that correct markup does not guarantee the feature will appear. Treat it as eligibility, not as a ranking lever.
Do I need different SEO for AI Overviews and AI Mode?
No. Google’s guide to generative AI features, updated 10 July 2026, says a page must be indexed and eligible to be shown with a snippet to appear in them, and that no special files, markup or rewriting are needed. The same technical and on-page basics apply.
How do I know if my organic traffic problem is technical?
Check the Page indexing report and URL inspection in Search Console for the pages that matter. If they are not indexed, or the indexed version is missing content, the problem is technical. If they are indexed and earn impressions but few clicks or conversions, look at on-page elements and at what the page offers after the click.
- Google Search Central: Understanding page experience in Google Search results, updated 22 September 2026 (developers.google.com)
- Google Search Central: Understanding Core Web Vitals and Google search results, updated 10 December 2025 (developers.google.com)
- web.dev: Interaction to Next Paint is officially a Core Web Vital, 12 March 2024 (web.dev)
- Google Search Central Blog: Mobile-first indexing has landed, 31 October 2023 (developers.google.com)
- Google Search Central: How to write meta descriptions (snippets), updated 20 April 2026 (developers.google.com)
- Google Search Central: General structured data guidelines, updated 10 July 2026 (developers.google.com)
- Google Search Central: Guide to optimizing for generative AI features on Google Search, updated 10 July 2026 (developers.google.com)
- web.dev: Milliseconds make millions, study by 55 and Deloitte for Google, data from late 2019, page updated 24 June 2020 (web.dev)
- Pathmonk: SEO case study, customer figures read on 11 November 2026 (pathmonk.com)