RTINGS.com has two sitemaps. The one everybody looks at holds 6,628 URLs — the reviews, the test methodology pages, the explainers. The second one holds 22,487 automatically generated "X vs Y" comparison pages, and almost nobody writing about this site mentions it exists.
That ratio is the whole business. For every editorial page a human touched, there are roughly three and a half machine-built ones, and they are all in the index, all interlinked, all pointing back at the lab data that produced them.
Our own Anatolii Ulitovskyi published an analysis of this site on LinkedIn in December 2025, breaking their approach into seven tactics — interlinking, deliberately oversized meta tags, bottom-funnel content, AI-written technical specs, real lab visuals, and links that arrive without being asked for. He fact-checks this blog, and the seven-part structure below is his.
What this article adds is the other half: which tool does each of those jobs, what it costs, and what the answers look like when you point them at rtings.com today. Every figure marked as ours was measured on 13 August 2026 and is visible in a screenshot you can open at full size.
One finding got a section of its own, because Anatolii's piece predates the test that produced it: the playbook that puts RTINGS at Google position 1 leaves them almost invisible to an AI model answering from memory.
RTINGS runs ~29,100 indexed URLs, of which 22,487 are auto-generated comparison pages. They hold #1 on Google for "best tv for bright room", above Reddit, Forbes and CNET. Their link profile is 722,598 backlinks from 35,625 referring domains, earned mostly from wikis and forums, and it is currently shrinking. And when we asked AI assistants twelve buyer questions in their categories, RTINGS was named in 3 of them — 0 of 6 when the model answered from memory, 3 of 6 when it searched the web.
What RTINGS is, and why the model is worth copying
RTINGS buys consumer electronics at retail, tests them in a lab with the same instruments every time, and publishes the numbers. TVs, headphones, monitors, mice, keyboards, printers, blenders, running shoes. The tests are repeatable, the scoring is consistent across a category, and the raw measurements are published rather than summarised.
That last part is what makes the programmatic layer possible. Once every TV in your database has a measured contrast ratio, response time and peak brightness, a page comparing any two of them can be generated without a writer. The content is not spun — it is a rendering of data that already existed.
Here is what the top of their sitemap actually contains, counted on 13 August 2026:
What one of the 22,487 pages actually looks like
Before the tactics, it is worth seeing the unit of production. This is a single side-by-side page, pulled from the comparison sitemap.
Nothing here is an opinion. Two products, the same measured fields, rendered in the same order, so the reader's eye can run down a column and stop where the numbers diverge. The page answers a question a human would need twenty minutes and two browser tabs to answer, and it cost nothing to produce beyond the tests that were already done.
That is the entire trick, and it is why the ratio in the chart above is the right way to read this site. The editorial pages are not the product. They are the input to the product: every review adds one more row to the database, and every new row multiplies the number of comparison pages that can exist. Reviewing a thirtieth TV in a category does not add one page, it adds twenty-nine.
It also explains the shape of their internal linking, which is the first tactic on Anatolii's list.
1. Interlinking and a simple structure
Anatolii's first point is that RTINGS wins on structure before it wins on content: navigation menus, an internal search bar and sitemaps that make every page reachable in very few clicks. With 29,000 URLs that is not a nicety — an orphaned page in a set that size is simply a page that does not exist.
The thing worth copying is the direction of the linking. Each generated comparison page links back to both product reviews; each review links to the test methodology that produced its numbers; the methodology pages link out to every product they were used on. The link graph is a consequence of the data model, not something a person maintains in a spreadsheet.
We pointed our free Website Audit at their homepage to see what that discipline looks like from a crawler's side.
Two of those notices are worth pausing on. "Nofollow internal links" on a site this size is usually deliberate — it is how you stop crawl budget draining into filter and sort URLs. And the missing social preview tags are a genuine miss that costs them nothing in Google and something in every share.
The audit is free, three checks a day, no account. If you are running a programmatic build, the number to watch is not the score but which group loses points: technical at 65/67 means the structure holds at scale.
2. Meta tags built long on purpose
Anatolii's second observation is that RTINGS writes meta tags far longer than the display limits, and that this is a feature rather than sloppiness — a longer description gives Google more raw material to construct a snippet from, and Google rewrites descriptions for the majority of queries anyway.
We measured this ourselves across a 20-page sample (14 reviews, 6 comparison pages), and the result sharpens the claim rather than repeating it.
So the copyable tactic is narrower than "write everything long". Descriptions are written as a paragraph of genuine summary — the longest we found ran to 697 characters and read like the opening of the review, because that is what it was. Titles stay conventional.
That distinction matters if you are generating tags at scale. A 259-character description is not padding; it is source material for a snippet engine that is going to rewrite you anyway. A 90-character title is just a truncated title.
If you want to build tags this way without hand-writing 29,000 of them, our Meta Tag Generator produces title and description sets with a SERP preview, and the Search Simulator shows you the truncation point before you publish.
3. Content across the funnel, weighted to the bottom
The third point in Anatolii's analysis is that RTINGS does not live on informational traffic. He puts informational keywords at 34.3% of their traffic; the rest sits on decision-stage terms — "best X for Y", head-to-head comparisons, the queries people type with a credit card already out.
The 22,487 comparison pages are that thesis rendered in URLs. "Samsung S90D vs LG C4" is not a topic anybody writes a blog post about; it is a question thousands of people ask in the last five minutes before buying.
We looked at one of their bread-and-butter terms with the free Keyword Research tool.
This is the part people get wrong when they copy programmatic SEO. They chase the 40,000-a-month head term and lose. RTINGS wins 810-a-month terms, thousands of times over, on pages that cost almost nothing to produce because the data was already in the database.
Look at the intent field in that screenshot too: Info. Even a term that reads commercially — someone deciding which television to buy for a sunny living room — is classified as informational, and that is the trap in reading intent labels literally. The person typing it is four minutes from a purchase. Anatolii puts informational keywords at 34.3% of RTINGS' traffic, which is the same observation from the other direction: the label on the keyword and the stage of the buyer are not the same thing.
The practical version for a build like this is to stop sorting by intent tag and start sorting by what the answer has to contain. If the honest answer is a ranked list of specific products, that page belongs in the programmatic layer. If it is an explanation, it belongs in the 265-page editorial set that feeds it.
Free tier is 3 searches a day and 10 keywords per list; a paid plan lifts that to 100 keywords, which is the difference between checking an idea and planning a build.
Grouping is the other half of the job. Ten keywords from four of their categories, dropped into the Keyword Cluster Tool, came back as four clusters with pillar topics attached — which is the shape a programmatic site plan actually needs.
4. AI writes the technical layer
Anatolii's fourth point is the one that makes people uncomfortable: analysing their top pages, he puts the mix at roughly 25% human-written and 75% AI-generated, with the AI handling the technical specification writing against structured data.
Read carefully, that is not "they publish AI slop". The generated layer is the part where a machine turns measured values into sentences — contrast ratio, response time, peak brightness in nits — and a machine is better at that than a tired writer, because it never mistypes the number.
The human 25% is where judgement lives: what to recommend, what the trade-off is, who should buy the cheaper one.
If you are building that hybrid, the pieces we make map onto it directly. The AI Content Writer drafts the structured layer, 3 articles a day free. The Article Rewriter reworks a template's output so 500 pages built from one skeleton do not read identically. The AI Humanizer takes the mechanical edge off a draft, and the AI Content Detector tells you where you still sound like a machine before a reader does.
The honest caveat: none of that manufactures the measurements. RTINGS' generated content is trustworthy because the numbers underneath it came from an instrument. Generate sentences about data you do not have and you have built 22,487 pages of nothing.
5. Real visuals are the moat, and no software sells you one
This is the section with no tool in it.
Anatolii's argument here is the strongest thing in his piece: "You can't easily fake a lab test." RTINGS photographs the actual unit on the actual bench, films the test, and publishes it. AI search engines, as he puts it, struggle with physical reality.
Every other tactic on this list is reproducible by anyone with a budget and a weekend. This one is not. It is why their generated pages rank while a competitor's identical-looking generated pages do not: the data is theirs, measured, and visibly so.
If you are in a category where you can measure something first-hand — response times, delivery windows, failure rates, prices you actually paid — that is your version of the lab. Publish the raw numbers, not the conclusions, and let the programmatic layer sit on top of them.
If you cannot, be honest with yourself that you are building the copyable 80% of this playbook and none of the defensible 20%.
6. At this scale, links arrive on their own
Anatolii's sixth point is that link-building stops being the job once a site is genuinely useful — the links come from wikis, forums and community threads without anyone pitching for them.
We ran rtings.com through our free Backlink Analyzer to see what that looks like as a profile.
Three things in that screenshot are worth more than the headline number.
The best links are nofollow. Wikipedia's "Contrast Ratio of 2015 TVs" citation and Apple's discussion threads pass no PageRank at all. They are there because someone needed a number and RTINGS had it. That is what earned links look like in the wild, and it is the opposite of what a link-building campaign produces.
36,960 pages on the domain have links pointing at them. Not the homepage — the long tail. The programmatic layer is not just ranking, it is accumulating citations.
The profile is shrinking. The link growth panel runs 906,485 down to 695,846 between 15 May and 13 August 2026. Anatolii cited a mid-2024 peak of over 54,000 referring domains; we measure 35,625 today. Different tools count links differently, so those two numbers are not strictly comparable — but the direction inside our own chart is unambiguous, and it is a useful corrective to the idea that a site like this only ever compounds.
The free tier returns 50 backlink rows per check, three checks a day; paid returns 100. The KPI block above is free either way.
7. The part the playbook does not cover: does AI name them?
Everything above is a Google strategy, and by that measure it works. We checked their position on one of their own category terms with the free Rank Tracker.
Now the same brand, through our free AI Visibility Checker, which asks AI assistants the questions a buyer asks and reports whether the brand gets named.
That contrast is the most useful thing in this entire case study.
722,598 backlinks, Domain Trust 86, position 1 on Google — and a language model asked "best monitor for office work" with no web access does not think of them. Not once in six answers. Everything RTINGS has built is a retrieval asset: it works brilliantly when something goes and looks, and does almost nothing when the answer comes from a model's unaided memory.
Their topic bars show where the ceiling is: TVs 1 of 4, headphones 1 of 4, monitors 1 of 2, speakers 0 of 2. A category they publish in extensively, and the AI never reached for them.
One honesty note, because it cuts against the number we just published. We ran this check twice on the same day. The first run returned 1 of 12 and a citation list that did not include rtings.com at all; the run in the screenshot returned 3 of 12 with rtings.com at the top of the cited sites. Language models are not deterministic and a searching model gets whatever the live web hands it that minute. Treat any single run as a sample, and treat "0 of 6 from memory" as the finding that repeated.
The tool for each job
Every job in this playbook, the tool that does it, and what you actually get without paying. Limits below are the ones enforced on our server, not the ones on a marketing page.
| The job | Tool | Free | What paid adds |
|---|---|---|---|
| Check the structure holds at scale | Website Audit | 3 checks a day, no account | — |
| Find the low-difficulty buyer terms | Keyword Research | 3 searches a day, 10 keywords per list | 100 keywords per list |
| Group them into templates and hubs | Keyword Cluster Tool | 2 runs a day | — |
| Confirm you actually rank | Rank Tracker | 3 checks a day, 3 keywords | 10 keywords, competitor authority scores |
| See who links to you and why | Backlink Analyzer | 3 checks a day, 50 link rows | 100 link rows |
| Find out if AI names you | AI Visibility Checker | 3 checks a day, 6 questions | — |
| Draft the generated layer | AI Content Writer | 3 articles a day | — |
| Stop templated pages reading alike | Article Rewriter | 3 rewrites a day, 10,000 characters | — |
| Take the machine edge off a draft | AI Humanizer | 3 runs a day, 10,000 characters | — |
| Write the page against a live score | Content Optimizer | Unlimited scoring, 3 SERP analyses a day | — |
| Generate title and description sets | Meta Tag Generator | 2 runs a day | — |
| Mark up reviews and FAQs | Schema Markup Generator | Article and FAQPage | Product, HowTo, LocalBusiness and 3 more |
Two things that table deliberately does not do. It does not put a price on the paid column, because this category reprices constantly and a number I verify today is a liability in three months. And it leaves out our app-hosted tools — the internal link checker, the toxic backlink checker, the topical map generator — because their limits live in a different system and I have not verified them for this article.
Where this playbook does not transfer
You probably do not have a lab. That is the load-bearing wall. Without first-party measurements, the generated layer has nothing true to say, and generated pages about data you did not collect are the exact thing Google's helpful content work is aimed at.
You do not have twenty years of citations. Wikipedia links to RTINGS because RTINGS was the only place with a contrast ratio for a 2015 television. That is a decade of compounding, not a campaign.
Scale without a data model is just thin pages. The 22,487 comparison pages work because each one renders two complete measured records. Spin 22,487 pages off a thesaurus and you have built a liability.
And it does not buy you AI visibility. That is the finding above, and it is the one most likely to matter over the next two years. If the model has to search to find you, you are dependent on being findable at retrieval time — which is a different discipline from ranking.
There is a subtler cost too, and it shows up in their own numbers. A database-driven site can only publish what its schema holds. RTINGS measures speakers and publishes speaker reviews, but the AI check returned 0 of 2 on speaker questions — the category exists in the template and not in the model's answer. When your content is generated from fields, the fields become the ceiling on what you can ever be known for. An editorial site can chase a new angle in an afternoon; a programmatic one has to add a column, re-test a product line, and regenerate.
The last thing worth saying out loud: this playbook takes years. The link profile in our screenshot is a decade of citations arriving one wiki edit at a time, and it is currently shrinking rather than growing. Nobody builds this in a quarter, and any agency that tells you otherwise is selling the 22,487 pages without the lab.
Run this teardown on a competitor in an afternoon
Everything in this article was produced with free tiers, a browser and about two hours. Here is the order.
- Read their sitemaps first. Fetch
/robots.txt, follow everySitemap:line, and count the URLs in each. A second sitemap you did not expect is usually where the strategy is hiding. - Sample twenty pages and measure the meta tags. Titles and descriptions, character counts, against the 60 and 160 display limits. Patterns show up fast, and they tell you whether tags are hand-written or generated.
- Audit one page. Not for the score — for which group loses points. Technical failures at scale mean the architecture is not holding.
- Pull the link profile. Look at the top referring pages, not the total. If they are forums and wikis, the links were earned; if they are guest posts and directories, they were bought.
- Check a category term. One "best X for Y" query tells you whether the model is working today, and who else is on the page.
- Ask the AI. Run the brand through an AI visibility check and compare it against the Google position. Where those two diverge is the opportunity.
Crawl a page, pull the link profile, check a ranking, and find out whether AI names you — the same four tools that produced the screenshots above, on your site instead of theirs.
- Website Audit — 3 checks a day, no signup
- Backlink Analyzer — 50 live link rows per check
- Rank Tracker — your real position in Google's top 10
- AI Visibility Checker — 6 buyer questions, answers shown in full
Frequently asked questions
What is programmatic SEO?
Generating large numbers of pages from a structured data set rather than writing each one. RTINGS is the clearest example we found: 22,487 of their indexed URLs are automatically built "X vs Y" comparison pages, each rendering two complete sets of lab measurements. It works when the underlying data is real and yours; it produces thin pages when it is not.
How many pages does RTINGS.com have?
Counting both sitemaps on 13 August 2026: 6,628 URLs in the main sitemap (5,934 reviews, 265 explainers, 241 test-methodology pages) and 22,487 in a separate comparison sitemap, for roughly 29,100 in total. The biggest editorial categories are TVs at 1,235 and headphones at 1,066.
Does RTINGS use AI to write its content?
Anatolii Ulitovskyi's analysis of their top eleven pages put the mix at about 25% human-written and 75% AI-generated, with the machine handling technical specification writing against structured data. We did not re-run that classification. What is checkable is that the generated layer describes measurements taken in their own lab, which is what makes it defensible.
Should meta descriptions be longer than 160 characters?
RTINGS clearly thinks so. Across our 20-page sample their descriptions averaged 259 characters, 19 of 20 ran past 160, and the longest was 697. The logic is that Google rewrites most snippets anyway, so a longer description is more raw material rather than wasted text. Note the asymmetry though: their titles averaged 53 characters, comfortably inside the normal limit.
How strong is the RTINGS backlink profile?
On 13 August 2026 our Backlink Analyzer returned 722,598 backlinks from 35,625 referring domains, Domain Trust 86/100, with 530,785 dofollow against 191,813 nofollow, and links pointing at 36,960 separate pages on the domain. The profile is currently contracting — the growth panel fell from 906,485 to 695,846 across three months.
Why does a site that ranks #1 on Google get ignored by AI?
Because they measure different things. In our check, ChatGPT answering from its own training memory named RTINGS in 0 of 6 answers, while the same model with web search on named them in 3 of 6 and cited three of their review pages. Rankings and backlinks make you findable when something searches; they do not put you in a model's unaided recall.
Can I copy this strategy without a testing lab?
Partly, and you should be clear-eyed about which part. The structure, the interlinking, the long descriptions, the bottom-funnel keyword targeting and the templated build all transfer. The reason those pages are trusted does not. If you can measure something first-hand in your own category — real prices, real delivery times, real failure rates — that is your substitute. If you cannot, you are copying the reproducible 80% and none of the defensible 20%.
How was this case study measured?
Sitemap counts and meta-tag lengths were measured directly against rtings.com on 13 August 2026; the meta-tag figures come from a 20-page sample of 14 reviews and 6 comparison pages. The audit, keyword, cluster, rank, backlink and AI visibility figures are single free-tier runs of our own tools on the same day, and each is shown in a screenshot in this article. Figures attributed to Anatolii Ulitovskyi are from his December 2025 analysis and were not re-verified.
The verdict
RTINGS is the best argument in public for programmatic SEO done properly, and the least copyable one. The 22,487 generated pages are not the achievement — the lab that makes them true is. Strip the measurements out and you are left with a page template that Google has been getting better at ignoring since 2022.
What does transfer is the discipline underneath: a data model that produces its own internal links, descriptions written as summaries rather than keyword strings, thousands of small decision-stage queries instead of a handful of vanity terms, and a machine doing the mechanical writing so people can do the judgement.
And then there is the gap. A site with three quarters of a million backlinks and position 1 on Google was named by an AI model answering from memory in none of six attempts. Whatever you take from the Google half of this playbook, that number is the one worth checking on your own brand — because it is the half nobody has finished building yet.
Start where we started: crawl a page, pull the link profile, and ask the AI whether it has heard of you. All three are free, and you will know more about your competitor in an afternoon than most audits tell you in a fortnight.
Measured on 13 August 2026 against rtings.com using free tiers of our own tools, plus direct sitemap and meta-tag sampling. Tool results are single runs and language models are not deterministic — the AI visibility check returned 1 of 12 on an earlier run the same day and 3 of 12 on the one shown. Figures credited to Anatolii Ulitovskyi come from his December 2025 LinkedIn analysis and were not independently re-verified. Third-party metrics change; re-run anything here before you rely on it.