Domain rating tiers in ChatGPT citations: a buyer's checklist for GEO software
In August 2026, mid-authority sites earned almost half of ChatGPT's citations while the biggest domains slipped. What that means for PR and content budgets.
Our verdict first: if you are paying for placements on the biggest websites in the hope of ChatGPT citations, August 2026 is a reason to check that spend. Promptwatch's ChatGPT citation share by domain rank report found that sites in the DR 46-75 range earned 46% of ChatGPT's citations over August 1 to 31, 2026, "almost half". Sites rated DR 91-100 averaged 4.4% of daily citations for the month, less than the lowest band, DR 0-15.
That doesn't mean big sites are worthless or that small sites win by default. It means authority alone is a weak predictor, and that your GEO software should show you the domain-rating profile of what gets cited for your prompts, not the market's.
How the report measures it
The report groups every cited domain into rating bands from DR 0-15 up to DR 91-100 and draws one stacked bar per day. Each bar adds up to 100% of citations whose domain has a known rank. That covers 89.65% of ChatGPT citations; the other 10.35% had no known rank and are left out. The page doesn't say which domain-rating provider it uses, so don't compare these bands directly with a number from your own SEO tool without that caveat. All figures here are for August 1 to 31, 2026, from the domain rank report.
The numbers by band
Monthly shares, in order of size, from the same August 2026 report, covering August 1 to 31, 2026:
- DR 61-75: 23.4%, the largest band
- DR 46-60: 22.6%
- DR 76-90: 18.9%
- DR 0-15: 6.3%
- DR 91-100: 4.4% (the average daily share for the month)
The page also gives grouped totals: DR 46-75 at 46%, DR 61-90 at "around 42% of citations all month", DR 0-30 at "about 14%", and DR 0-45 at "almost a third". The two biggest gainers were DR 61-75, up 1.6 points, and DR 46-60, up 0.9 points (Aug 1 to 31, 2026, domain rank report).
Promptwatch's "ChatGPT citation share by domain rank in August 2026" chart (Aug 1 to 31, 2026), captured November 14, 2026.
The August 14 break
The most striking part of the report is the top band's slide within the month. DR 91-100 sites averaged 7% of citations in the first week of August, fell to about 5% on August 8, and dropped to about 3% on August 14, where they stayed for the rest of the month. The page says the second drop happened "the same day Reddit's share of ChatGPT citations collapsed", and that the top band "lost more than half their share" (domain rank report, Aug 1 to 31, 2026).
The page links the timing to Reddit; it doesn't claim to know what ChatGPT changed. Our read is that some of the top band's share was Reddit itself and other giant platforms, so a change in how ChatGPT used those platforms would show up as a drop in DR 91-100. That is a hypothesis. For the Reddit side of the story, see our post on Reddit citations dropping in ChatGPT.
Promptwatch's "ChatGPT citation share per domain rank group" chart (Aug 1 to 31, 2026), captured November 14, 2026.
What this means for PR budgets
Here is how we'd translate the data into spending decisions. This is our interpretation, not the report's.
A placement on a mid-authority trade publication or a specialist review site isn't a consolation prize. In August 2026 the DR 46-75 band carried almost half of ChatGPT's citations. If your PR agency prices placements by domain rating and steers you toward the top of the scale, ask them to show which of those outlets ChatGPT actually cites for your category's questions.
Mega-sites still matter for brand awareness, classic search, and referral traffic. The point is narrower: for ChatGPT citations specifically, the top band shrank to about 3% of citations from August 14, 2026, per the domain rank report. A budget split that assumes the biggest names are the surest path into AI answers was not supported by August's data.
Keep a selection effect in mind too. The report doesn't say how many domains sit in each band, but the middle bands almost certainly hold many more sites than the top one, so part of the DR 46-75 share may simply reflect how many of those sites exist. What a buyer wants to know is which specific outlets in your niche get cited, not which band wins on average.
What this means for your own content
If your own domain sits in the middle bands, the report is encouraging: a mid-authority site can compete for ChatGPT citations. If it sits low, DR 0-30 still held "about 14%" over August 1 to 31, 2026 (domain rank report). Authority helps, but the content on the page decides whether it gets quoted. Our read is that topical fit and a clear answer near the top of the page probably matter more than a few rating points, though the report doesn't test that.
What GEO software should show you
The public report is a market average. Your decisions need the same view for your prompts. Here's what to look for.
- Domain rating attached to every cited domain, so you can see the authority profile of the sources behind your category's answers.
- That profile over time, because August shows it can change inside a month.
- Source-type classification next to the rating, so a DR 70 review site and a DR 70 news site aren't lumped together.
- Competitor domains highlighted in citation search, so you see which outlets cite rivals and not you.
- Offsite mention discovery, ranked as opportunities, to build a PR target list from pages that already name brands like yours.
- URL-level tracking for each placement you pay for, so you know whether that article ever gets cited.
Questions for the vendor
- "Can you show me the domain rating of each domain cited for my prompts, and how that mix changed last month?"
- "Which rating provider do you use, and what happens to domains with no rating?"
- "Can I filter citations by source type and by domain rating at the same time?"
- "If I land a guest article next week, how do I track whether ChatGPT cites that URL?"
- "Can you list third-party pages that mention my competitors but not me?"
How Promptwatch handles it
Promptwatch is our pick for this job. Its Citation Trends tab classifies citations into 22 content types and 20 source types and shows citation rate by type and domain rating over time, which is the report's view applied to your own monitors. It also gives per-page timing: first cited, ramp-up days, peak rate, and decay half-life.
Promptwatch's changelog entry for Citation Trends (the illustration is example data), captured November 14, 2026.
For PR planning, citation search highlights competitor domains, and offsite mentions list pages that name a brand without linking, ranked as opportunities. Page Tracker follows any URL, so a paid placement can be checked daily against the prompts you care about. Citation rank shows where your own domain lands in an answer's source list, and self-citation rate shows how much of the citing is done by your own site versus third parties.
Pricing starts at $95 a month on Essential (50 prompts, daily tracking). Professional at $245 adds crawler logs, custom reports, and the Looker Studio connector, which helps if your PR team reports in Looker.
Alternatives, on catalog facts: Ahrefs Brand Radar sits inside Ahrefs next to its link and content tools, with the AI Visibility Index from $199 a month on top of an Ahrefs plan; it is strong for backlink context, but our catalog describes its prompts as modeled from keyword data. AthenaHQ starts at $295 a month on Starter, and Scrunch AI Core at $250 a month. Neither entry lists domain rating over time inside citation analytics.
The takeaway
In August 2026, ChatGPT cited mid-authority sites far more than the very biggest ones: DR 46-75 at 46%, DR 91-100 at 4.4% on average and about 3% from August 14 (Promptwatch report, Aug 1 to 31). Before your next PR budget review, find out which outlets ChatGPT cites for your own questions. Promptwatch's 7-day free trial on monthly billing is enough to build that list for one market.