Websites Getting ChatGPT Visitors Don’t Have Stronger Backlinks Than Similar-Size Sites
Websites with ChatGPT among their top referral sources looked stronger on links at first glance. Compared with sites that get similar traffic, the gap almost disappears.
If you've been told you need a pile of new backlinks before ChatGPT will send you customers, Diib's data doesn't back that up. Among sites with similar traffic, the ones already getting ChatGPT visitors had about the same link strength as everyone else. Having real, steady traffic tracked more closely with ChatGPT referrals than links did.
OpenAI just announced visual ads that appear next to images people generate in ChatGPT, starting in the US later this month. The businesses most likely to care are the ones whose customers already find them through ChatGPT. So what do those websites have in common? A common guess is backlinks: links from other websites, the same signal Google has rewarded for years. Diib’s data doesn’t support that guess.
What the data shows
Diib looked at websites with Google Analytics connected and checked their top five referral sources between August 1 and October 5, 2026. 1,600 of about 55,600 websites (2.9%) had ChatGPT among them.
- At first glance, those sites looked stronger. Their median Domain Authority (DA, Moz’s 1–100 estimate of link strength) was 18, versus 15 for other sites, and they had about 260 linking websites versus about 160.
- But they were also bigger. The typical ChatGPT-referred site had about 7,700 visits in August and September, versus about 2,350 for the rest.
- Compared with sites of similar traffic, the link gap nearly disappears:
| Visits, Aug–Sep 2026 | Median DA: ChatGPT referrals | Median DA: no ChatGPT referrals |
|---|---|---|
| Under 1,000 | 12 | 10 |
| 1,000–5,000 | 14 | 15 |
| 5,000–20,000 | 19 | 19 |
| 20,000+ | 24 | 25 |
Traffic mattered more. About 1% of sites with under 1,000 visits had ChatGPT as a top referrer, compared with roughly 5% of sites with 5,000 or more.
What this does and doesn’t tell you
This is aggregated Diib data from websites with connected analytics. It only counts ChatGPT when it ranked among a site’s top five referral sources, so smaller amounts of ChatGPT traffic are missed. It shows a pattern, not a cause. Diib can’t see what people asked ChatGPT before they clicked.
Source: View the original source