Under the hood
White Hat Technical SEO: Crawling, Canonicals and Speed, Plus Two Traps Dressed as Clever Engineering
Technical SEO is where a vendor can bill for anything, since you cannot see it. Google's requirements fit in one sentence. Most of the rest is ordinary housekeeping, a little of it matters enormously, and two tricks can put the site itself in violation.
November 9, 2026 · 11 min read
The short answer
White hat technical SEO means keeping a site readable: Googlebot is not blocked, pages return a 200 status, content is indexable, duplicates point to one preferred URL, and pages load well. Those are mostly housekeeping, not policy. The policy traps are hidden text and cloaking, which means showing search engines different content from visitors. JavaScript sites are allowed, though rendering adds a delay.
On this page
- Three requirements, and an audit with 214 red flags
- Settings that do the opposite of what people assume
- Duplicate content is a filing problem, not a crime
- Speed: published thresholds, modest claims
- Hidden text and cloaking: the two that can cost you the site
- JavaScript sites are slower to read, not suspect
- Reading a technical invoice without learning to code
Three requirements, and an audit with 214 red flags
The audit arrives as a PDF. Two hundred and fourteen issues, most of them in red. Missing alt text on a decorative swirl. A title that is four characters too long. The cover letter suggests a monthly program to address these critical findings. Before you sign, read what Google says it needs. Its technical requirements, last updated December 18, 2025, come to three: Googlebot is not blocked, the page works (Google receives an HTTP 200), and the page has indexable content. Its Search Essentials overview, updated December 10, 2025, adds that most sites pass without even realizing it.
That does not make the rest worthless. It does mean a technical fault is usually not a sin. A broken sitemap is not a violation; it is a broken sitemap. "White hat" only becomes the relevant phrase at the traps further down this page. For the mechanics of each item, read our technical SEO guide. What follows is narrower: which settings are misunderstood, where the honest line runs, and how to read an invoice. The white hat SEO hub has the wider argument, and our spam policy guide lists every policy.
Settings that do the opposite of what people assume
The most expensive technical mistakes are small and sincere. Someone wants a page out of Google, so they block it in robots.txt, which stops Google reading the noindex that would have removed it. Google's robots.txt introduction, updated December 10, 2025, says the file is "not a mechanism for keeping a web page out of Google." Its technical requirements page says to use noindex and allow crawling, because a blocked URL might still appear in results.
| What you want | The right control | The usual mistake |
|---|---|---|
| A page out of the index | noindex, with crawling allowed, or a password | A robots.txt block, which can hide the noindex |
| One preferred URL among duplicates | A redirect or rel="canonical" | Relying on the sitemap, which Google calls a weak signal |
| The status of one page | URL Inspection | Staring at the Page indexing report on a 40-page site |
| Google to find a page | A link from another page, using an anchor with an href | A button that only runs a script |
| Nothing odd after a redesign | A check of robots.txt, noindex tags and redirects on launch day | Assuming the staging settings left with the staging site |

Keep the indexing numbers in proportion. Google's Page indexing help, read November 7, 2026, says not to expect every URL to be indexed, that non-indexed URLs can be fine for reasons such as a noindex tag or a duplicate, and that a site with fewer than 500 pages probably does not need that report at all. A list of "excluded" URLs is not a list of problems until someone shows which pages you wanted indexed. If a real page is missing, start with this walk-through, and crawl errors has its own.
Duplicate content is a filing problem, not a crime
"Duplicate content penalty" may be the most profitable phrase in the trade. Google's starter guide, last updated December 10, 2025, says duplicate content under multiple URLs is fine and will not cause a manual action, though it is inefficient. The same passage says copying someone else's content "is a different story." So a product at three addresses, because a filter added a parameter, needs tidying. It does not need a rescue.
The tidying has a ranking of its own. Google's canonicalization guide, updated July 10, 2026, calls redirects and rel="canonical" annotations strong signals and sitemap inclusion a weak one, and says using more than one raises the chance your preferred URL appears. Note the word chance. A canonical tag is a strong request and not a command. Our canonical tags guide shows the syntax, and redirects covers moving pages. When two town pages differ by a name, that is a separate matter: the doorway policy cares about purpose, and a canonical tag will not fix a page that has no reason to exist.
Speed: published thresholds, modest claims
Speed has real numbers, which is rare in this business. Google's web.dev guidance, read November 6, 2026, sets "good" Core Web Vitals at Largest Contentful Paint of 2.5 seconds or less, Interaction to Next Paint of 200 milliseconds or less and Cumulative Layout Shift of 0.1 or less, at the 75th percentile of page loads. Check against those. A perfect score from a testing tool on one page, from one desk, is not the same thing.
And keep the claim small. Google's content guidance, updated October 5, 2026, says core ranking systems look to reward a good page experience and that owners should not focus on only one or two aspects of it. Nobody should promise you rankings for a speed score. The honest reason to fix a slow site is the visitor standing in a parking lot on a phone. Ours is a Core Web Vitals service; the free guide may be all you need.
Hidden text and cloaking: the two that can cost you the site
Here the phrase finally earns its keep. Google's spam policies, last updated August 28, 2026, define cloaking as presenting different content to users and search engines with the intent to manipulate rankings and mislead users, including inserting text or keywords only when the requester is a search engine. Hidden text and link abuse is content placed solely to manipulate search engines and not easily seen by people: white text on white, text behind an image, CSS that moves text off-screen, font size or opacity of 0, a link on one small character such as a hyphen.
Google says these are not violations
- Accordions, tabs, slideshows, sliders and tooltips
- Text meant only for screen readers
- A paywall, if Google can see the full gated content and the site follows its Flexible Sampling guidance
- Redirects for moving a site, consolidating pages, or after login
Google names these in its policies
- Keywords served only to the search engine's request
- A town list in the footer, in the footer's color
- Off-screen or zero-opacity text placed for rankings
- Sending mobile visitors, and only them, to a spam domain
The awkward part: you may have these without having done anything. Google says hackers often use cloaking to make a hack harder for the owner to detect, and that a hacked redirect can depend on referrer, user agent or device, so a URL can redirect when clicked from Google Search but not when visited directly. The owner types the address, sees the normal site, and tells the customer they are imagining things. Google's April 13, 2026 post on back button hijacking adds that some instances may originate from a site's included libraries or advertising platform. Here is how to check, using free tools.
- 1
Search your own site and click from the results
Use a phone, on cellular data, and arrive from Google rather than typing the address. A hacked redirect may only fire that way. Also search your site name plus words you would never publish, such as pharmacy or casino.
- 2
Compare what Google got with what you see
In URL Inspection, read the HTML Google received for a key page against your browser's copy. Words in one and not the other deserve a question.
- 3
Read the Security issues report
Google's report help, read November 7, 2026, says it covers hacked content, malware and unwanted software, and social engineering. A clean site shows a green check mark.
- 4
Check the owners list
Google's permissions help, read the same day, says an owner you do not recognize might mean the site was hacked.
- 5
Clean first, optimize later
If you find something, stop all other SEO work. Malware removal comes before anything else, and penalty recovery covers what follows if a notice appears.
JavaScript sites are slower to read, not suspect
A site built in React is not a shady site. Google's JavaScript SEO basics, updated March 4, 2026, describes three phases: crawling, rendering and indexing. A page may wait in the rendering queue for a few seconds or longer before a headless Chromium runs its scripts. If the words depend on that step, they arrive later, and if the script fails they may not arrive. That is a timing and reliability problem. It breaks no policy. Our JavaScript SEO guide shows how to read the rendered HTML.
Google still says server-side rendering or pre-rendering "is still a great idea," because it makes a site faster for users and crawlers and not all bots can run JavaScript. We build on Next.js, which prerenders, so count us as an interested party. In fairness: a standard WordPress site also sends its text in the HTML, so rendering alone is no reason to leave one. WordPress vs Next.js has the fuller comparison. The AI crawler side of the question is on JavaScript and AI crawlers.
Reading a technical invoice without learning to code
Work you can see
- "Removed noindex from 12 service pages" with the 12 listed
- "Redirected 40 old URLs," with the map attached
- "Compressed hero images; LCP on these pages moved from this to that"
- A note of what was checked and found fine, so you do not pay twice
Filler to question
- An issue count going down, with no list of what changed
- "Technical optimization: ongoing," month after month
- A third-party health score: Google says such tools have no access to its ranking data
- Urgent warnings about crawl budget on a small site, with nothing to show for them
Google's hiring page, updated June 5, 2026, suggests asking "Will you share with me all the changes you make to my site, and provide detailed information about your recommendations and the reasoning behind them?" and says an SEO with FTP access should be willing to explain every change. For an audit, it says to grant read access to Search Console, not write access. Those two sentences protect you better than any tool. Questions to ask an agency has the rest, and what to demand in a report covers the monthly side.
- Most sites: a morning with Search Console and a free site audit or health check finds what matters. A one-off repair is often all you need, and the six-month checklist keeps it from recurring.
- Ongoing SEO: technical upkeep sits inside our white hat SEO service, from $575 a month, depending on complexity, setup and budget, with no long contract, per our SEO pricing.
- Keeping it maintained: hosting and maintenance starts at $55 a month.
- Site past repair: a rebuild is fixed by page count, from $1,500 for up to 10 pages to $5,750 for 51 to 70 pages, with financing at 0% and 25% down. A rebuild is rarely the first answer, and you should be suspicious of anyone who says it is.
Questions owners ask about white hat SEO
Nearly all of it. Fixing crawl errors, redirects, canonicals and speed breaks no policy; those jobs are housekeeping and Google asks for them in its own documents. The exceptions are the tricks wearing a technical costume: cloaking, hidden text and sneaky redirects, all named in Google's spam policies, updated August 28, 2026. If a technique only works when customers and search engines see different things, it is not housekeeping.
No, according to Google's starter guide, updated December 10, 2025: duplicate content under multiple URLs is fine and will not cause a manual action, though it is inefficient. Clean it up with a redirect or rel="canonical" so Google shows your preferred address. Copying other people's content is a separate issue, and Google says so in the same passage.
No. Google's spam policies, updated August 28, 2026, say accordions, tabs, slideshows, sliders, tooltips and text meant only for screen readers do not violate its policies. Design choices that help a reader are fine. The policy targets content placed solely to manipulate search engines and not to be easily seen by people, such as white text on a white background.
Yes. Google says hackers often use cloaking to make a hack harder for the owner to detect, and that hacked redirects can depend on referrer, user agent or device. So your site may look normal when you type its address. Click through from Google on a phone, compare the HTML in URL Inspection, and read the Security issues report. If you find it, treat it as a security job first.
Nobody honest will promise that. Google says its core ranking systems look to reward a good page experience, and that owners should not focus on only one or two aspects of it. Fix speed because your visitors feel it, and use Google's published thresholds as the target. A speed repair is worth doing whether or not rankings move, and it should be sold that way.
No. Google says robots.txt is not a mechanism for keeping a page out of Google, and that a blocked URL might still appear in results. Use noindex and allow crawling, or password-protect the page, which is the right choice for anything truly private. This fix page covers the opposite problem, when robots.txt is blocking pages you want found.
Keep reading
- Under the hoodTechnical SEO: The Checklist That Decides Whether Google Can Read Your SiteTechnical SEO in plain English: robots.txt, sitemaps, canonicals, redirects, status codes, rendering and speed, plus a one-afternoon audit you can run free.
- Month by monthWhite Hat SEO Checklist: Six Months of Steps, in Order, for an Owner With Search ConsoleA white hat SEO checklist in order, month by month, that an owner can run with Search Console: accounts, indexing, titles, local, links and upkeep.
- The monthly workWhite Hat SEO Services: One Month Opened Up, Line by Line, With the Evidence Each Line LeavesWhat ethical SEO services contain in a month, what you can open and inspect, what we will not do, and our price: from $575 a month, no long contract.
- The arithmeticWhite Hat SEO Pricing: Why the Cheap Offer Is Cheap, Where the Money Goes and When to Spend NothingWhat ethical SEO costs and why: the arithmetic of the $99 backlink offer, where an honest month's money goes, a budget test, and when to spend nothing.
- Earned, or something elseWhite Hat Link Building: What Earns a Link, Where the Scheme Starts and What Can Honestly Be PromisedWhite hat link building explained: why a stranger links to you, where Google's link spam policy starts, what a provider can promise, and how to check it.
- Local, without the mailboxWhite Hat Local SEO: Real Addresses, Real Reviews and a Business Name Nobody StuffedWhite hat local SEO in plain terms: Google Business Profile rules field by field, honest service areas, reviews nobody arranged, and who holds the keys.
- Everything in White Hat SEOSee the section
From the Insights desk
- Why Your Trade Pages Rank on Page ThreeSearch Console says Google shows your service page, just on page three. Six reasons contractor pages stall in the twenties and thirties, what Google's own guidance says about each, and a one-month plan to move one trade up.
- Your Site Ranks on Google and Is Invisible to AI. Here Is How to Check.Google's crawler runs your JavaScript. Many others do not — they read the HTML your server sent and move on. If your content is assembled in the browser, you are not ranked badly in an AI answer; you are not in the running. Two tests, no tools.
Send the audit you were given and we will sort it
Forward the PDF or your site address. A person at DataCram replies with what matters, what is filler and what you can fix alone, in plain language. Repair work is quoted after a look at the site.
Which of these sounds like you? (pick any)
Rather talk now? Call (210) 346-0848.