SEO Specialisms · Technical
The engineering layer of search. Crawl budget, rendering, indexing rules, Core Web Vitals and structured data for websites in Dubai and across the UAE, audited against Google's own documentation.
Technical SEO is the work of making sure a search engine can reach your pages, render them the way a browser does, understand what each one is for, and store it in the index, quickly enough that a real person does not give up first. That is the whole scope. It does not include keyword research, writing or earning links. Technical SEO in Dubai is therefore scoped by page count and platform rather than by keyword list, which is why two sites of similar size can need very different amounts of work. Technical SEO in Dubai is usually bought either as a one off audit before a bigger investment, or as the first phase of a longer engagement, because none of the other work pays back on a site Google cannot process.
This page is deliberately narrow. If you want the broader programme, with keyword research, content and reporting alongside the engineering, that is our SEO services page instead. What follows is the engineering layer on its own: crawling, rendering, indexing, site architecture, speed and structured data, each checked against what Google publishes about how Search works rather than against folklore.
One framing worth fixing at the start, because it changes how you read every section below. Google describes its pipeline in three stages: crawling, where a program discovers and fetches URLs; indexing, where it processes and tries to understand a page and stores it; and serving, where it picks results for a query. Technical SEO acts on the first two. Nothing you do in this discipline influences the third directly. It removes obstacles. It does not manufacture demand, and an agency that presents technical fixes as a route to fixed positions is describing something that does not exist.
Stage one
Before anything else can go right, a crawler has to discover the URL, be allowed to request it, and get a usable response.
Google finds URLs from links it already knows, from sitemaps you submit, and from redirects. A page with no internal links pointing at it and no sitemap entry may simply never be found, which is the most common cause of a page that has been live for months and has never appeared in search.
The robots.txt file at your root decides which paths a well behaved crawler may request. It is a fetching instruction, not an indexing one, which is the distinction that catches most teams out. A blocked page can still appear in results, without a description, if other sites link to it.
The crawler needs a fast, stable status code. Intermittent server errors, long response times, or an aggressive firewall rule that treats a crawler as an attacker will all reduce how much of your site gets fetched, and none of them show up in a normal browser test.
robots.txt
Robots exclusion was standardised as RFC 9309 in 2022, after decades as a convention. It is a small file and a large liability. The single most expensive error we find on Dubai websites is a Disallow: / left in place from a staging environment after launch, which tells compliant crawlers not to fetch anything on the site.
The second most expensive is more subtle. Teams block a directory in robots.txt intending to keep it out of the index, then discover the URLs listed in search anyway. Google’s own guidance is explicit that robots.txt controls crawling, not indexing. To keep a page out of the index you must let the crawler fetch it and find a noindex instruction, which is impossible if you have blocked it. The two rules cancel each other out.
Blocking CSS and JavaScript is the third. It was once common advice and it is now actively harmful, because the renderer needs those files to see the page.
Crawl economics
This is the area where the most effort is wasted on sites that do not need it.
Crawl budget is the number of URLs a crawler will fetch from your site in a period. Google has been consistent that most sites do not need to think about it. If you have a few hundred pages and they get indexed, the topic is not your problem and time spent on it is time not spent on something that is. A technical SEO audit in Dubai should say plainly when crawl budget is not your constraint rather than filling pages with it.
It becomes real on large catalogues, and in Dubai that usually means property portals, car marketplaces, recruitment boards and retailers with faceted navigation. The pattern is always the same: filters generate URLs combinatorially. A category with six filters, each with eight values, can generate hundreds of thousands of addresses, nearly all of them near identical. A crawler spends its allowance on those and never reaches the pages that matter.
Diagnosing this properly needs server log files, not a crawling tool. A log tells you what search engines actually requested, how often, and what they got back. A crawler tells you what is theoretically reachable. Those are different questions, and only the first one answers where the budget went. Getting logs out of a UAE shared host sometimes takes a support ticket, and it is worth the wait.
The fixes are unglamorous and effective: canonical tags pointing filtered views at the clean category, parameters excluded at the robots level once you are certain they carry no unique content, internal links that do not point at filtered URLs in the first place, and a sitemap listing only the pages you actually want fetched.
Stage two
This is where modern sites fail most often, and where the failure is least visible to the people who own the site.
If your website is built with React, Angular, Vue or any framework that assembles the page in the browser, there is a step between fetching and indexing that did not exist on older sites. Google processes JavaScript in three phases: it crawls the URL, queues the page for rendering, and then renders and indexes it. The rendering phase is not instant and it does not always reproduce what you see in Chrome.
The practical consequence is a site that looks perfect to everyone in the business and is close to invisible in search. The test is simple and worth doing before you commission any other SEO work at all: fetch your own page with JavaScript disabled, or use the URL Inspection tool in Search Console and read the rendered HTML it returns. If your headings, body copy and internal links are absent from that output, nothing else on this page will help you until it is fixed. Rendering is the single most common reason technical SEO work in Dubai is commissioned in the first place.
The remedies are architectural, which is why this conversation belongs with your developers early rather than late. Server side rendering produces the full HTML on the server. Static generation pre builds pages at deploy time. Both put the content in the initial response. Retrofitting either onto a live single page application is substantial work, and it costs far less to decide before a web development project starts than eighteen months after launch. If you are planning a rebuild, the platform choice and the rendering strategy are the two decisions with the longest shadow.
Stage three
Most websites publish the same content at more than one address without anyone intending it. The homepage answers at four variants of the domain. A product sits under two categories. A tracking parameter creates a new URL every time a campaign runs. A print view duplicates an article. Each of these is a separate address serving substantially the same content, and Google has to pick one to show.
You can make that choice instead of leaving it to be guessed. A canonical link element nominates the preferred address, and consolidates the signals pointing at the duplicates onto it. Google treats it as a strong hint rather than a directive, and it ignores canonicals that contradict other evidence, so the tag has to agree with your sitemap, your internal links and your redirects. When those four disagree, the tag is the first thing discounted. Reconciling them across templates is routine technical SEO work on any Dubai site with more than a few hundred URLs.
The related decision is what should not be indexed at all. Thank you pages, internal search results, filtered views, tag archives and staging subdomains rarely deserve a place in the index. The correct instrument is a noindex robots meta tag on a page the crawler is allowed to fetch, never a robots.txt block.
Architecture
Internal linking is usually filed under content. The parts that decide whether a page is discovered and how authority moves through a site are engineering.
The number of clicks from your homepage to a given page is a reasonable proxy for how readily it will be crawled and how much weight it carries. Pages buried six levels down, reachable only through pagination, tend to be fetched rarely. Flattening that is usually a navigation and hub page exercise rather than a URL change.
A crawler follows anchor elements with an href attribute. A div with a click handler, a button that calls a router, or a menu that only populates after a user interaction may be perfectly usable and completely invisible. This is common in Dubai on sites built from a design tool export. Technical SEO catches it because a crawl finds pages a visitor can reach and a crawler cannot.
An XML sitemap should list canonical, indexable URLs that return 200, and nothing else. Sitemaps generated automatically by a plugin routinely include redirected URLs, noindexed pages and 404s, which wastes the crawl and weakens the signal.
Every migration leaves them. A URL redirecting to a URL redirecting to a URL costs crawl time and loses a little signal at each hop. Chains longer than two should be collapsed to point straight at the destination, and redirect loops should be found before a crawler does.
Arabic and English
A large share of Dubai businesses publish in both Arabic and English, and a fair number add Russian, Hindi or Chinese. That creates a technical problem specific to this market: several pages carrying the same meaning at different addresses, competing with each other unless you declare the relationship.
Hreflang annotations tell Google that these URLs are alternates of one another and which audience each serves. Get them right and an Arabic speaker in Dubai is shown the Arabic page. Get them wrong and the annotations are ignored wholesale.
The rule that is broken most often is reciprocity. If the English page names the Arabic page as its alternate, the Arabic page must name the English page back. A one directional declaration is discarded. Every page in a set must also list itself, and the set needs an x-default for anyone the other entries do not fit.
Two further points matter locally. Language and region are separate: ar is Arabic anywhere, ar-AE is Arabic for the United Arab Emirates, and using the wrong one narrows or widens your reach unintentionally. And hreflang is a targeting signal, not a ranking one. It influences which version is shown, never whether the set ranks. For bilingual businesses this is the highest value technical SEO fix available in Dubai, because it decides which language a searcher is shown.
Speed
Three metrics, each describing a different way a page can feel broken to the person using it.
Core Web Vitals are Google’s attempt to put numbers on page experience. There are three, and they measure loading, responsiveness and visual stability. They are reported on field data, meaning measurements collected from actual Chrome users, and this distinction matters more in the UAE than almost anywhere, for a reason covered below.
How long until the biggest element in the viewport has rendered. Usually a hero image, a video poster or a headline. Slow LCP is most often an oversized uncompressed image, a slow server response, or a font or script blocking the render.
INP replaced First Input Delay as a Core Web Vital in March 2024. It measures how long the page takes to visibly respond after a tap or click, across the whole visit rather than only the first interaction. Poor INP is nearly always heavy JavaScript occupying the main thread.
How much content jumps around while loading. Caused by images without dimensions, fonts swapping late, and banners injected above existing content. It is the metric users complain about without knowing its name, because it makes them tap the wrong thing.
Here is the part specific to this market. Field data reflects your real audience, on their real devices and their real connections. A Dubai site whose visitors are largely in the Gulf, on good mobile networks and recent handsets, will record different numbers from a site with the same code serving South Asia or East Africa on older devices. Two identical builds can score very differently, and the useful question is never the score in isolation but what the field data says about the people who actually visit you.
This also explains a frustration we hear often. A developer runs a laboratory test, sees a strong result, and cannot understand why Search Console reports a failing group of URLs. Laboratory tools simulate one device on one connection. Field data aggregates everyone. When the two disagree, the field data is the one describing your customers. Hosting location is part of this too: a server physically close to your audience, or a content delivery network with a presence in the region, removes latency that no amount of code optimisation will recover. Hosting location is the one technical SEO variable in Dubai that code changes cannot compensate for. It is worth checking where your site is actually hosted before optimising anything, and it is a standard part of keeping a site healthy rather than a one off task.
Machine readable
Structured data is a block of markup, normally JSON-LD, that states plainly what a page is about. That this is a professional service, located here, with these opening hours. That this is a product, at this availability. That this page answers these questions. Prose implies those things to a human reader; structured data states them unambiguously to a machine.
Two returns come from it. The first is appearance in search: review stars, FAQ expanders, breadcrumb trails and similar treatments are generated from structured data, and they change how much of the result page you occupy. The second matters more each year. Assistants and answer engines lean on structured data to decide what a page is and whether it answers a question, so the same markup that earns a rich result also makes a page easier to cite.
The constraint worth knowing before you invest is that Google’s guidelines require the markup to describe content actually visible on the page. Marking up reviews you do not display, or questions you do not answer, is a policy violation that can cost you rich results across the whole site. Markup is a description of the page, not an addition to it. Validating that description is a standing part of technical SEO for any Dubai site claiming rich results.
Answer engines
A growing share of buyers now ask an assistant before they ever open a results page. The technical requirements overlap with classic SEO but are not identical.
Whether an assistant can use your site comes down to access, structure and attribution, and all three are technical concerns.
Access. Answer engines use their own crawlers, with their own user agent strings, and they respect robots.txt. A site that permits Googlebot while silently blocking other agents, often through a firewall rule or a bot protection product rather than anything in robots.txt, is invisible to them. Whether you want to be crawled by them is a business decision and a legitimate one to answer either way. What you should not do is make that decision by accident, and most sites we audit have.
Structure. Assistants extract passages, not whole documents. A page that answers its own question in the first two sentences of a section, and whose headings say what the section contains, is far easier to quote correctly than one where the answer emerges gradually across several paragraphs. Self contained sections matter more than they do for classic ranking, because a passage lifted out of context has to still make sense.
Attribution. Claims backed by a named, linked primary source are more likely to be repeated with your name attached. The same structured data that supports rich results helps here, because it states the facts about your organisation in a form that needs no interpretation.
None of this is a separate product with a separate invoice. It is the same crawlability, the same structure and the same markup, checked against a second set of consumers. Technical SEO in Dubai now has to satisfy answer engines as well as Google, and the overlap between the two is most of the work. Where it connects to the writing rather than the code, it belongs with generative engine optimisation, and the two are usually scoped together.
Mobile and security
Google indexes the mobile version of your site. Not the desktop version with a mobile check afterwards: the mobile version is the version. Anything present on desktop and absent on mobile is, for indexing purposes, absent. Content collapsed behind tabs is fine and is treated normally. Content genuinely removed from the mobile template is not indexed at all, and we still find Dubai sites where the mobile build quietly drops half the body copy to make the design fit.
The corollary is that responsive design is the safe default. One URL, one HTML document, layout adapting to the viewport. Separate mobile subdomains create duplicate content, hreflang complications and two codebases to keep in step, and they are rarely worth it now.
Security is the other. HTTPS is a technical requirement in Search Essentials and browsers mark plain HTTP pages as not secure, which costs conversions regardless of search. The failures we see are rarely a missing certificate. They are mixed content, where an otherwise secure page loads an image or a script over HTTP, certificates that expire because nobody owns renewal, and redirect logic that sends visitors through an insecure hop on the way to the secure destination.
The risky moment
Every serious traffic collapse we have been called in to diagnose traces back to a change nobody expected to matter.
A migration is any change to how URLs resolve: a new domain, a redesign with a new URL pattern, a platform move, HTTP to HTTPS, a restructure of categories, or a merger of two sites. The common factor is that addresses Google has learned stop behaving the way they did.
Done carefully, a migration costs a temporary dip while the index catches up. Done carelessly, it costs traffic that does not return, because the equity attached to the old addresses was never passed to the new ones. The difference is almost entirely preparation, and the preparation is dull.
It starts with a complete inventory of existing URLs, gathered from your sitemap, your analytics, Search Console and a crawl, since no single source has all of them. Every URL in that inventory needs a decision: it moves to a specific new address, or it goes and returns 410, or it stays. What must never happen is the shortcut of redirecting everything to the homepage, which Google treats as a soft 404 and which discards the value of every page you redirected.
Then the mapping is tested on staging, before launch, page by page for the templates and by sample for the long tail. On launch day the redirects go live at the same moment as the new site, never after. Afterwards you watch Search Console daily for a fortnight, because that is where a mistake shows up first, and monthly for a quarter.
If you are planning a platform change or a rebuild, bring this conversation forward rather than treating it as a launch task. The URL structure of the new site is a decision made early and expensive to revisit, and it is the one decision in a rebuild that most affects whether you keep the traffic you already have. Bringing technical SEO into a Dubai rebuild at the wireframe stage costs a fraction of repairing it afterwards.
Monitoring
Third party tools estimate. Search Console reports what Google actually did with your site.
Splits your URLs into indexed and not indexed, and gives a reason for each exclusion. Crawled but currently not indexed, discovered but not crawled, duplicate without a user selected canonical, blocked by robots.txt. Each of those names a different problem with a different fix, and reading them properly is most of a technical audit.
Takes a single address and shows how Google sees it: whether it is indexed, which canonical Google chose as against the one you declared, and the rendered HTML after JavaScript has run. This is the fastest way to confirm a rendering problem, and the answer takes a minute.
Groups URLs by template and reports field data against the thresholds. Because it groups, a fix to one template shows as a shift across every page using it, which makes it a reasonable way to prioritise engineering work by reach rather than by page.
Two reports that are empty on a healthy site and worth checking anyway. A manual action means a human reviewer has penalised the site. A security issue usually means a compromise. Both are the kind of thing an owner finds out about far too late without a monthly glance.
One caveat on all of it. Search Console reports lag by two to three days, and both traffic and indexing move for reasons that have nothing to do with your site, including seasonality that in the UAE follows Ramadan, Eid and the summer months rather than a Western calendar. Judging a change over a week is guesswork. Judging it over a quarter, against the same period last year, is analysis. Technical SEO reporting in Dubai should be read on that cadence, not weekly. Reading these reports alongside your analytics is what turns them into decisions.
What you get
The deliverable is a written document, not a tool export. Automated crawlers produce hundreds of flagged items with no sense of which three matter, and handing that to a development team is a reliable way to get nothing fixed. Our audits are ordered by expected impact against implementation cost, so the first section is the short list worth doing this month.
Every finding states the problem, the evidence, the specific URLs affected, the fix in terms a developer can act on, and the source in Google’s documentation. Where a finding is a judgement call rather than a rule, it says so, because your developers are entitled to disagree with us and should be able to see the reasoning.
If your own team implements the fixes, that is often the most direct route and we are glad to work that way. If you would rather we did, that is scoped separately once the audit has shown what is involved, and you get a fixed written quote for it. Either way the technical SEO findings for your Dubai site belong to you and stay useful if you later change agency. We do not require a retainer to run an audit, and SEO is available as an optional monthly retainer if you want the work continued.
Status codes
Every request gets a code. Crawlers act on it literally, and a polite lie causes more damage than an honest error.
The failure we find most often is the soft 404: a page that no longer exists, tells the visitor so in words, and returns 200 anyway. To a person it reads as a missing page. To a crawler it reads as a perfectly good page whose content happens to be the sentence “sorry, this page could not be found”. Google detects many of these and reports them, but it spends crawl allowance discovering them first, and the ones it does not detect sit in your index representing nothing.
The same applies to a deleted product redirected to its category. If a genuinely equivalent page exists, redirect to it. If it does not, let the address return 404, or 410 when you know the removal is permanent. In technical SEO terms a Dubai site with no 404s at all is usually a site hiding them. A 404 is not a fault to be eliminated; it is the correct answer to a request for something that is gone.
Server response time underpins all of it. Google’s own guidance treats a slow server as a limit on how much of a site gets crawled, and it is also the first component of Largest Contentful Paint, so it is charged to you twice. On UAE shared hosting the usual causes are an overloaded neighbour, an uncached content management system rebuilding every page per request, and a database that has never been indexed properly. Page caching, object caching and a content delivery network with regional presence address most of it, and they cost far less than the redesign people often reach for instead.
The page exists and has content. Verify that your templates are not returning 200 for empty category pages or expired listings, which is the most common source of thin pages in the index on Dubai property and recruitment sites.
A 301 says the move is permanent and passes signals to the destination. A 302 says the original will return. Using 302 for a permanent move is common, and it delays consolidation on the new address for no reason.
Repeated server errors cause Google to back off and crawl less. An occasional 500 is survivable; a pattern during your busiest hours is a crawling problem as well as a customer problem, and log files will show it before analytics does.
How it runs
Four stages, with a decision point at the end of each so you are never committed further than the value shown.
A short call and a look at the public site to establish page count, rendering approach, languages and platform. That determines the quote, which is fixed and written, and arrives within 45 minutes during business hours.
Crawl, render tests on every template, Search Console and field data review, log file analysis where logs are available, hreflang and structured data validation. Output is the prioritised written document.
A session with you and your developers to go through the findings, argue about the debatable ones, and agree who implements what. You leave with an ordered list, not a document to interpret.
Once fixes are deployed, whoever did them, we verify each finding and confirm in writing what is resolved and what is not. Search Console is then watched over the following weeks for the indexing response.
What we need from you is read access to Search Console and your analytics, the name of the platform and who maintains it, and a route to your developers. Server log access is valuable and not always possible; we work without it and say so where a finding would have been firmer with it. You own everything produced, and nothing in the audit depends on us continuing to work with you afterwards.
It is worth saying plainly, because it is not in an agency’s interest to. If your site is a recently built, server rendered brochure of thirty pages, on decent hosting, with a valid certificate and a sitemap, and Search Console shows your pages indexed and your Core Web Vitals passing, then technical SEO is not your constraint. Buying an audit will produce a tidy document confirming that things are fine, and your money is better spent on the content and the visibility work that actually creates demand.
The sites where this discipline pays are the ones with a real symptom: pages published and not indexed, a JavaScript front end, a catalogue large enough for crawl budget to bite, more than one language, a migration completed or planned, or a Core Web Vitals group failing in the field. If you recognise your site in that list, the audit is the quickest way to find out how much of your problem is engineering before you spend anything on a full search engagement.
If you are not sure which of those describes you, say so when you get in touch and we will tell you honestly, including when the answer is that you do not need us yet. Technical SEO in Dubai is worth buying when it removes a measured obstacle, and not before. Our team works with businesses across Dubai and the wider UAE, in Arabic and English, and a fixed written proposal follows within 45 minutes during business hours with no obligation.
Straight answers
Technical SEO covers only whether a search engine can reach, render, understand and index your pages, and how fast they are for real users. It does not cover keyword research, writing or link acquisition. A full SEO engagement includes those; technical SEO is the foundation underneath them, and it is often bought on its own as a one off audit.
The clearest signal is a gap between pages you have published and pages Google has indexed, which the Pages report in Search Console shows directly. Other tells are pages appearing in search with the wrong title, content missing from the cached version, or the site working in a browser but rendering blank to a crawler.
For the audit, no. It runs from the public site plus read access to Search Console and analytics. For the fixes, either your developers implement our written findings, or we need staging access and a deployment route. Which of those applies is agreed before work starts.
It depends almost entirely on page count and whether the site is server rendered or JavaScript rendered. A brochure site is a different exercise from a catalogue with faceted navigation. You get a fixed written quote scoped to your site within 45 minutes during business hours.
It removes obstacles rather than creating demand. If pages were not being indexed, or were slow enough that users left, fixing that can change results quickly. If the pages were already indexed and fast, technical work alone will not move positions, and any agency saying otherwise is selling you something.
Sources
Fixed price, in writing
Got it. Your quote is being written now.
In business hours you will have it within 45 minutes. Check your inbox for the confirmation.
