How to tear down a competitor channel

On 19 August 2026, a direct HTTP pull of one competitor's homepage returned an 11-word server-side body while its sitemap listed roughly 60,521 URLs and its backlink profile showed 485 referring domains. This article shows how to run that same teardown on a competitor channel or site: what they publish, where their traffic comes from, and how their links were earned.

The short answer

A competitor teardown is three measurements, not an opinion. First, count what they actually publish: pull their sitemap and count URLs by type. Second, read where their traffic comes from: organic share, direct share, referral share, paid search share. Third, audit their links: referring domain count, dofollow percentage, and how many of those links look purchased. On 18 August 2026, one competitor in this category measured 161,345 monthly visits with 50.1 percent organic traffic and 485 referring domains, of which 430 were flagged as purchased in a 554-link sample. Those three numbers describe that business better than any review page.

Count what they publish, not what they claim

The fastest teardown step costs nothing: fetch the competitor's sitemap.xml directly over HTTP, without running JavaScript, and count the URLs. The result often contradicts the marketing.

One competitor in the short-video category, StoryShort, listed roughly 60,521 total URLs on 19 August 2026. Of those, 11,994 were category pages per language, across 5 languages, plus 411 pages, 119 blog posts and 21 workflow pages. The measurement also surfaced a hidden locale: ru, not listed in the sitemap, detected through its backlink targets. Their title template was uniform across the catalog: "{Topic} AI Videos - Create {Topic} Videos | StoryShort", applied to topics ranging from zodiac-killer to muffins.

Compare that to the server-side body of their homepage: 11 words, measured 19 August 2026 by pulling the HTML, stripping script and style blocks, and counting the remaining text without executing JavaScript. Four tested pages returned the same figure. A site can carry sixty thousand URLs while its own homepage ships eleven words of crawlable content. That gap is the finding.

Do the same arithmetic for yourself. ViewMade's own site measured 229 total URLs with 36 original pages and 16 tool pages on 22 August 2026, and its homepage SSR body carried 1274 words. Small catalogs are not a defect; uncounted catalogs are.

Read the traffic mix before reading the totals

Total visits flatter big sites and hide strategy. The mix tells you what actually works.

From Similarweb's public data layer, July 2026 snapshot, measured 18 August 2026:

Site	Monthly visits	Organic %	AI referral %
invideo.io	6,846,906	47.8	2.61
vidiq.com	6,306,935	17.3	2.61
opus.pro	5,934,254	30.7	1.62
revid.ai	498,593	43.3	0.22
crayo.ai	375,546	40.6	0.11
storyshort.ai	161,345	50.1	0.19
short.ai	121,613	51.3	0.11
hooked.so	110,170	63.0	0.0
tubegen.ai	83,024	32.0	0.0
fluxnote.io	80,752	33.0	3.92

Three patterns come out of this table. Hooked.so runs the highest organic share at 63.0 percent with only 189 referring domains, which means content is doing the work, not links. Fluxnote.io shows the highest AI referral share at 3.92 percent, so it is winning citations inside AI answers despite being among the smallest sites measured. Vidiq.com, second largest by visits, has the lowest organic share of the top three at 17.3 percent, meaning its traffic leans on brand and direct channels.

For your own channel, the actionable move is the same: find which single source dominates your strongest competitor and ask whether that source is available to you. If a small rival gets most of its traffic from organic search with few links, they are ranking pages, and page architecture is copyable.

Audit the link profile for purchased links

Backlink counts mislead until you split them. On 19 August 2026, a backlink scan of a 5618-record dataset produced these figures for StoryShort: 485 referring domains, 3611 total links, 25 percent dofollow, domain rating 46, and 430 purchased links out of a 554-link sample, or 78 percent. The two most-linked pages were /ru/make/ai-influencer-generator with 109 links and a German blog post reviewing Crayo with 103 links.

The category-wide pattern matters more than any single site. Sites with high purchased-link ratios were absent from the top 12 across commercial head keywords, while low-ratio sites held top positions. Measured 19 August 2026: inreels.ai at spam score 86 fell below the measurement threshold; storyshort.ai at 79 missed the top 12 in 14 of 15 tested commercial queries; autoshorts.ai at 56 was absent from its head keyword. Meanwhile medeo.app at spam score 4 ranked third for "brainrot video generator", hooked.so at 5 ranked first for "true crime ai video maker", and vicsee.com at 6 ranked first for "faceless video generator".

Vicsee's profile explains why. Of 627 clean links, 593 pointed at the homepage, 94.6 percent, with 18 percent brand anchors and zero advertising budget, and the result was the first position for "faceless video generator". When you tear down a competitor, compute their purchased-link ratio first. If it is high, their rankings are rented and vulnerable; if it is low and anchored to the homepage, their position is durable and worth studying instead of attacking.

Check SERP reality against the sitemap ambition

Publishing volume means nothing if the pages do not rank. Test the competitor's own brand query and a set of commercial queries in your niche, from one location, and record the results.

On 19 August 2026, a US-region SERP scan tested 15 commercial queries. StoryShort appeared outside the top 12 in 14 of them; the sole exception was "text to brainrot" at position 4. More striking: for the query "storyshort alternative", the first eight results were fluxnote.io, tubegen.ai, fluxnote.io again, eliro.pro, autofaceless.ai, subclip.app, seektool.ai and topai.tools. The brand itself was absent from its own comparison query's first page.

That is the teardown's sharpest finding: a site can hold 60,521 URLs, 485 referring domains and 161,345 monthly visits, and still lose every commercial query that matters, including queries about itself. Before you imitate a competitor's publishing volume, run their brand-plus-alternative query. If third parties own it, their moat is thinner than the visit count suggests.

What to do with this on your own channel

Run the four measurements on yourself before anyone else. Pull your own sitemap and count URLs by type. Check your traffic mix for one month. Sample your backlinks and estimate the purchased ratio. Search your brand name plus "alternative" and record who ranks.

Then pick one gap. If your organic share sits below 40 percent while a smaller competitor holds above 50, their page architecture is the thing to study, not their posting frequency. If your purchased-link ratio exceeds half your sample, stop buying and point new outreach at the homepage, following the vicsee.com pattern of 94.6 percent homepage links. If a third party owns your brand-alternative query, build one honest comparison page naming what the other tool does well; a page that admits the competitor wins somewhere is the page that gets cited.

This is the method ViewMade applies when it measures its own category: direct HTTP pulls, JavaScript off, every figure stamped with its measurement date, and internal cost figures kept out of public pages so customer pricing stands on its own.

What this does not tell you

Every number here comes from public surfaces: sitemaps, HTTP pulls without JavaScript execution, Similarweb's open data layer, and a backlink scanner sample. Similarweb figures for smaller sites carry wide error margins; the 161,345-visit figure is an estimate, not a server log. The backlink purchased-link ratio comes from a 554-link sample, not a full census. SERP scans reflect one location at one point in time; rankings shift weekly. None of this measures watch time, retention, revenue per view, or audience quality, because none of those appear on public pages. We have not measured upload cadence against subscriber growth, and we have not measured whether AI referral traffic converts. This article answers what a channel's public footprint looks like; it does not answer whether copying that footprint will grow your channel.

<!-- faq -->

Frequently asked questions

How many URLs should a competitor site have before I worry?

There is no threshold. Vicsee.com held 1634 URLs and ranked first for "faceless video generator" as of 19 August 2026, while a site with roughly 60,521 URLs missed the top 12 in 14 of 15 commercial queries. Page efficiency varies enormously: tubegen.ai drew 1153 monthly visits per page while storyshort.ai drew 2.7 in the same measurement window. Count the competitor's URLs, then divide their visits by that count. Low output per page signals a catalog problem you may be able to exploit.

Can I trust free traffic estimates for teardowns?

Treat them as directional. The figures in this article come from Similarweb's public data layer with an explicit caveat: estimates for small sites deviate more than for large ones. They are reliable for ratios, such as organic share at 50.1 versus 17.3 percent, and unreliable as absolute counts. Never build a business case on a single estimated number; build it on the pattern across several competitors measured the same way on the same date.

How do I tell if a competitor bought their backlinks?

Sample their links and look for three markers: anchor text heavy on commercial keywords rather than brand, links pointing at deep template pages instead of the homepage, and high spam scores on linking domains. In the 19 August 2026 dataset, sites scoring above 50 on spam metrics were absent from head-keyword top 12s, while sites under 20 held top-three positions. A homepage-anchored, brand-heavy profile like vicsee.com's, 94.6 percent homepage links at 18 percent brand anchors, reads as earned.

Does AI referral traffic matter yet?

It is measurable but small almost everywhere. Across the category measured 18 August 2026, AI referral shares ranged from 0.0 percent at hooked.so and faceless.so to 3.92 percent at fluxnote.io, with invideo.io and vidiq.com tied at 2.61 percent. Fluxnote.io proves a small site, 80,752 monthly visits, can lead the category on this metric. If you want AI citations, the working lever appears to be honest comparison pages that name competitors' strengths, since that is the content AI systems quote.