Domain rating tiers and ChatGPT citations: which PR targets the August 2026 data favors
In August 2026, mid-rank domains earned almost half of ChatGPT's rated citations while the very top tier earned 4.4%; a tier-by-tier comparison for PR and guest content planning.
Quick verdict
The biggest sites are not where most of ChatGPT's citations went in August 2026. Promptwatch's report on citation share by domain rank (August 1 to 31, 2026, ChatGPT) found that domains rated 46 to 75 earned 46% of citations, almost half. The 91 to 100 tier, the household-name sites, averaged 4.4% for the month.
For a PR or guest-content plan, the report suggests aiming at the 61 to 90 range and benchmarking against the 46 to 75 range, rather than waiting for a placement on the very largest outlets. That is a reasonable reading. It is not a promise that a mid-tier placement gets you cited.
What the report measured
The report covers the 89.65% of ChatGPT citations that came from domains with a known rank. The other 10.35% are excluded, so every share below is a share of the rated part, not of everything. It is one vendor's dataset, for one month, in ChatGPT. It uses domain rank, abbreviated DR, in tiers as the unit and says nothing about individual outlets or topics.
The tier comparison
| DR tier | Share of citations (monthly average) | Note from the report |
|---|---|---|
| 0 to 15 | 6.3% | DR 0 to 30 together is about 14% |
| 46 to 60 | 22.6% | Gained 0.9 points over the month |
| 61 to 75 | 23.4% | Gained 1.6 points, the largest gain |
| 76 to 90 | 18.9% | DR 61 to 90 held about 42% all month |
| 91 to 100 | 4.4% | Averaged 7% in week one, about 3% from August 14 |
Source: Promptwatch data report, August 2026, ChatGPT. The 16 to 45 band is not itemised in our source; the report states that DR 0 to 45 as a whole is almost a third of citations. Tiers 46 to 60 and 61 to 75 together make the 46% figure.
Two details deserve attention. First, the very top tier fell during the month: about 7% in the first week, about 5% on August 8, then about 3% from August 14. The report points out that August 14 is also the day Reddit's share collapsed. That timing is a coincidence we can observe, not a cause we can assert, and the report does not claim one.
Second, the groups that gained were in the middle, with DR 61 to 75 up 1.6 points and DR 46 to 60 up 0.9.
What this does and does not mean
A domain rank describes a site, not a page, and ChatGPT cites pages. A strong placement on a DR 70 site is a different thing from a weak one. The report does not say that raising your own DR raises your citations, and it does not rank outlets within a tier.
It also cannot say whether mid-tier sites are cited because they match prompts better, because they publish more specific material, or because of how ChatGPT's sources shifted in August. We would be inventing the explanation if we picked one. One month of data in one engine is also too short to call a trend. Check it again with a fresh window before you reallocate a PR budget.
What to do without any tool
- Do not wait for a top-tier placement. The report shows DR 91 to 100 as a small slice of citations. Spread effort across more outlets rather than betting on one.
- Aim PR and guest content at the 61 to 90 range. Look for publications and industry sites in that range that cover your category. Pitch something specific: original data, a named expert, a worked example.
- Benchmark against the 46 to 75 range. Compare your own domain and your competitors' with that band, and compare the outlets you pitch with it too.
- Check relevance before rank. A DR 70 site on an unrelated topic is not a good target. The engines answer prompts, and a page has to match the prompt.
- Record what you do. Keep a sheet of each placement, its date and the prompts you expect it to help with. Without it, you cannot connect work to results later.
You can look up DR for a list of target outlets in any SEO suite that reports it. Our source does not say which provider the report used, so tier boundaries in your tool may differ slightly.
Measuring where your mentions show up
The measurement step is knowing which third-party pages are cited for your prompts and whether your brand appears in them. In our catalog, Promptwatch documents domain-level and page-level citation analytics, offsite mentions (your brand named in third-party pages without a link) and a Page Tracker for specific URLs. Citation Trends classifies each citation by source type and reports first cited, ramp-up days, peak rate and decay half-life per page, which tells you how quickly a placement is picked up and how long it lasts.
For PR work specifically, the offsite mentions view is the one that closes the loop. It shows pages that name your brand without linking to you, which is a ready-made list of outlets to thank, correct or build on. Action Board turns gaps into a prioritized to-do list, with a Discovery agent suggesting up to five evidence-backed actions per weekly run.
Plan notes: Explore is free for ChatGPT only. Essential is $95 a month, Professional is $245 and Business is $579. Crawler logs start on Professional, which matters if you want to confirm that the AI bots can read the placement's host site and your own pages. A 7-day trial exists on monthly billing for Essential and Professional.
For comparison, our entries for AthenaHQ and Scrunch AI describe visibility tracking and audits, but neither records a domain-rank view of citations. Ahrefs Brand Radar integrates with Ahrefs' link and content tools, but our entry does not say it breaks AI citations out by domain rank tier. We would not assume either way.
Who should pick which
- A brand with a small PR budget: pick ten mid-tier outlets in your niche before chasing one large one, and track each pitch.
- A brand already in top-tier press: keep it for reputation, but stop expecting it to carry your ChatGPT visibility on its own.
- An agency running outreach for clients: you need citation data per client to defend the strategy. Promptwatch's agency self-serve plans start at $199 a month for Kick-off and include crawler logs, with unlimited projects and prompts.
- A team without a tool: run your prompts monthly, list the cited domains, look up their ranks and build the target list by hand.
For related reading, see monitoring earned media impact on AI search, tools that measure PR impact on AI visibility and tracking the third-party sites AI uses.
Bottom line
In August 2026 the middle of the rank range carried more of ChatGPT's rated citations than the very top, and the very top lost share during the month. That is enough to justify widening a PR plan, not enough to promise a result. Test on your own prompts, and recheck the data when the next window lands.