Key takeaways
- In August 2026, sites with domain rating 46-75 earned about 46% of ChatGPT citations. DR 91-100 sites averaged just 4.4%, below the DR 0-15 group at 6.3%.
- The DR 91-100 share fell from roughly 7% to about 3% in the middle of the month, on the same day Reddit's share of ChatGPT citations collapsed.
- This is share of citations, not share per domain or per page. Other studies show high-DR sites win on breadth (more pages cited), not per-page strength.
- You do not need a top-tier DR to get cited. You need relevant, reachable, extractable pages, plus brand mentions on other sites.
- Treat all of this as correlation from one month of data, and re-check it next month.
What the August 2026 numbers actually say
For years the working assumption in SEO was simple: more authority, more visibility. A Promptwatch data report on August 2026 ChatGPT citations puts a dent in that assumption for AI search.
The report, ChatGPT citation share by domain rank for August 2026, covers ChatGPT citations from Aug 1 to Aug 31, 2026. It uses Promptwatch's real-UI monitoring. It only counts citations from domains with a known rank, which is 89.65% of all ChatGPT citations in the period. Here is the monthly average share by DR group.
| Domain rating | Average share of ChatGPT citations |
|---|---|
| DR 0-15 | 6.3% |
| DR 0-30 (combined) | about 14% |
| DR 0-45 (combined) | almost a third |
| DR 46-60 | 22.6% |
| DR 61-75 | 23.4% (highest group) |
| DR 46-75 (combined) | 46% |
| DR 76-90 | 18.9% |
| DR 91-100 | 4.4% |
Share climbs with DR up to the 61-75 band and then drops off. The most striking line is the last one. The most authoritative sites on the web get a smaller slice of ChatGPT's citations than sites with almost no authority at all.
Claudia Serrano summed it up in a LinkedIn post: a solid mid-authority site with relevant content is enough to show up.
The mid-August drop for DR 91-100
The monthly average hides a story inside the month. According to Promptwatch Data, DR 91-100 sites averaged about 7% in the first week of August. They fell to about 5% on Aug 8, then to about 3% on Aug 14, and stayed there. That is a loss of more than half their share.
Most of that lost share went to the middle tiers. DR 61-75 gained 1.6 points and DR 46-60 gained 0.9 points.
Two dates matter here, and both line up with other things ChatGPT did.
- Aug 8: ChatGPT began using the site: operator at scale. Promptwatch's site: operator report shows site: queries jumping from about 0.4% to about 17% of fan-outs. Searchable's separate analysis found a similar change, from 0.2% to 18%, with pages consulted per answer rising from 11.6 to 22.5.
- Aug 14: reddit.com's share of ChatGPT citations fell from about 4% to 0.5%, per the Reddit citations report.
I would not claim these events caused the DR shift. The report doesn't either. But the timing is hard to ignore. When ChatGPT starts aiming searches at named domains and consulting twice as many pages, the answer pool gets wider, and a wider pool tends to include more mid-sized sites.
Why mid-authority sites might win
Nobody can prove the mechanism from one month of data. These are the explanations that fit the evidence best.
More pages are consulted per answer
Searchable's study of 5,000+ tracked brands found that pages consulted per ChatGPT answer went from 11.6 to 22.5 overnight on Aug 8. Promptwatch separately found almost half of ChatGPT responses cite 10 or more unique domains (Sep 28 to Oct 4, 2026). With that many slots per answer, a handful of giant sites cannot fill them. Specialist and mid-sized sites get room.
Domains with high DR are few
Here is the caveat that matters most. The report measures share of citations, not share of domains, and it does not normalise for how many domains sit in each DR band. Very few domains have DR 91-100, while far more sit in the middle. Part of the DR 46-75 advantage could simply be that there are many more of them.
So the right reading is not that mid-DR sites are favored per site. It is that mid-DR sites collectively supply nearly half of what ChatGPT cites. For a marketer that is still useful, because it tells you where the competition for citations lives.
Authority buys breadth, not per-page strength
A Green Flag Digital study of 999 finance domains found a similar pattern. DR 90+ sites took half of all citations because they had more pages cited (median 524 versus 188 for sites under DR 50). Their citations per page were no higher. Inside the top 500 pages, Domain Rating had no detectable effect.
The study's authors call DR a door opener: it helps a page get into the cited set, but it does not make that page win once inside. They also warn that the sample is finance only, US only, and that R-squared sat between 0.09 and 0.18, so more than 80% of variation in citations is unexplained by authority, links and rankings.
Relevance beats prestige for specific questions
The same study found a WSJ page earning 231 citations with zero ranking keywords. A page that answers the exact question on a site ChatGPT already draws from can win with almost no SEO footprint. Mid-authority sites that go deep on one topic fit that pattern well.
Content type matters more than the logo
Promptwatch's content type report found product pages are the most cited type at 22.4% of ChatGPT citations (Sep 28 to Oct 4, 2026), followed by landing pages and listicles. Mid-sized brands often publish exactly these pages about their own products. Huge publishers mostly publish editorial.
What the research says about links, mentions and freshness
If DR is not the lever, what is? A few studies point the same way, with caveats.
| Signal | What the data says | Source |
|---|---|---|
| Brand web mentions | Correlate about 3x more strongly with AI visibility than backlinks (0.664 vs 0.218) | Ahrefs study of 75,000 brands, via Digital Applied's meta-analysis |
| Accessibility and search rank | Scored highest of 23 factors (9.5/10 and 9.4/10) | Zyppy meta-analysis of 54 studies |
| Backlinks and ranking keywords | 10x more of either goes with about 10% more ChatGPT citations | Green Flag Digital, finance pages |
| Domain Rating | No detectable effect among top-cited pages | Green Flag Digital |
| Freshness | AI-cited content about 25.7% fresher than organic top 10; the popular "4.3x" claim is unsourced | Digital Applied meta-analysis |
| Own-site citation | Brands recommended 44% more often when any of their pages is cited | 180 Marketing, 10,000 ChatGPT responses |
All of these are correlations. The authors of the underlying studies say so, and I would repeat it. Still, the direction is consistent: reachable pages, relevant content, and being talked about elsewhere on the web matter more than a number from a link database.
One more finding worth knowing
Searchable's analysis of the Aug 8 shift found that brand-owned pages gained citation share by 80% (1.80x) while barely changing in reading share (0.95x). Regulators and standards bodies like fda.gov and nist.gov gained heavily, while affiliate and review media such as forbes.com and techradar.com lost ground. If your own site is the best source on your own product, the trend is moving your way.
What mid-authority brands should do
Promptwatch's own recommendations from the report are sensible, and I would add context to each.
Do not wait for a better DR
A solid mid-authority site with relevant content is enough. Delaying AI search work until your DR climbs to 80 is a poor plan, since the data shows DR 76-90 sites already earn less share than DR 61-75 sites.
Aim PR and guest content at DR 61-90 sites
DR 61-90 sites held about 42% of citations all month. They are the most reliable earned-coverage targets. Placements on the very biggest sites may look impressive but, in August at least, those sites drew a small share. Since brand mentions correlate with AI visibility, getting named on the sites ChatGPT actually cites matters more than getting a trophy placement.
Benchmark against DR 46-75
That range is where most of the competition sits. Compare your citation share against rivals in that band rather than against Wikipedia-scale sites you will never match.
Do not write off low-DR sites
DR 0-45 sites still earned almost a third of citations. A newer site with a sharp, specific page can show up.
Fix access before anything else
Zyppy ranks URL accessibility first. If AI crawlers hit errors or blocked paths, nothing else matters. Check robots.txt, paywalls and rendering. Promptwatch's AI crawler logs show which bots actually read which pages, and the free Agentic Readiness Check covers the basics.
Publish pages ChatGPT likes to cite
Product pages, landing pages, documentation and clear comparisons last longer than news. In Promptwatch's content half-life data, landing pages and docs stayed cited longest, while news lost half its pages within 36 days. Green Flag also found half of the most-cited finance pages from October 2025 get zero citations today, so plan to refresh.
Cover the cluster, not one page
Because ChatGPT fans out into many sub-queries, a site with several tightly related pages has more chances to be pulled in. The free query fan-out generator shows what sub-queries a prompt might trigger.
Tools for tracking this yourself
You can see your own position without guessing. Tools like Promptwatch track which of your pages ChatGPT cites, which competitors are cited beside you, and where you land in the source list. It also pairs that with crawler logs and visitor analytics so you can tell an access problem from a content problem. Start with the free AI visibility report or the 7-day trial.
Other options exist for monitoring alone.
Otterly.AI
Profound
For a broader list, the GEO software directory at bestgeosoftware.com compares more of them. For DR itself, you can still check it in Ahrefs, which is how most DR figures in these studies are computed, though the Promptwatch report does not state its source of ranks.
How to read this data without fooling yourself
- Share of citations is not share per domain. See above.
- One month is a short window, and August had two ChatGPT behaviour changes inside it.
- DR is a third-party metric. It is a proxy, and it did not predict per-page citation counts in the Green Flag data.
- Different studies use different prompt sets and verticals. Finance, healthcare and travel citation patterns differ. Promptwatch's healthcare and software reports show government sites and documentation leading respectively.
- Citations also shift with platform policy. ChatGPT's Reddit share fell from roughly 4% to 0.5% in one day in August. Expect more of this.
The practical move is to re-run the same view monthly for your own topics, rather than trust any single headline number.
Bottom line
If you run a DR 50 or DR 65 site, the August 2026 data is good news, with honest limits. You are in the band that supplied almost half of ChatGPT's citations. The sites with the loftiest authority numbers took 4.4%. Focus on being reachable, relevant and mentioned, publish the page types that get cited, and benchmark against the DR 46-75 crowd. Then check the numbers again next month.