Research Report
July 2026 | Mentioned Research
Reddit Holds Half of Google's Top 3, and 3% of Bing's
Second wave of a longitudinal study. We ran the same 1,300 keywords across 26 industries in January and July 2026. This time we also ran them on Bing, ChatGPT, Perplexity and Gemini, to test whether Reddit's position in search holds outside Google.
August 19 update: we re-ran this panel after ChatGPT's mid-August change. The ChatGPT interface went from citing Reddit in 6.0% of these answers to 0.4%, while the same queries through OpenAI's API and on Google were unchanged. Read the follow-up report.
Does Reddit Show Up in Search?
On Google, constantly, and near the top. On Bing, almost never. Same 1,300 keywords, same provider, same depth, same day:
69.3%
of result pages show Reddit, and in the top 3 on 50.0%
3.1%
of result pages show Reddit, and in the top 3 on 1.3%
A 22.6x gap, the widest of any platform we measured and more than double YouTube's 9.9x. Reddit ranking is a property of Google, not of search in general, so treat it as a bet on one engine.
Does Reddit Show Up in AI Search?
Partly, and it depends almost entirely on which assistant. We put the same 1,273 questions to all three and recorded whether a Reddit link appeared in the sources each one cited:
68.5%
of answers cite Reddit
16.7%
of answers cite Reddit
6.0%
of answers cite Reddit
There is no single number for "AI search". An eleven-fold spread separates the most and least Reddit-reliant assistant, wider than the gap between Google and Bing. Anyone quoting one figure for AI as a whole is averaging across products that behave nothing alike.
So Should You Invest in Reddit?
It depends on which surface you are buying, and the honest answer differs by channel.
Yes for Google and Perplexity, and they are one purchase. Reddit takes a top-3 Google slot on 50% of the searches we tested, twelve times YouTube's rate. Perplexity then cites Reddit on 95.8% of the queries where Google ranks it in the top 3, and 90.4% of the time it is the same thread. Ranking a thread on Google gets you both surfaces at once.
No for ChatGPT. It cites Reddit in 6.0% of answers and the rate does not move with Google rank: 5.0% when Reddit is ranked 1–3, 6.9% when Reddit is absent from Google entirely. Whatever drives ChatGPT's sourcing, Reddit ranking is not a lever on it. If ChatGPT visibility is the goal, this is the wrong channel.
Check your category first. Since January, the share of Reddit queries that stopped returning Reddit ranged from 2.2% in Software & Apps to 46.2% in Real Estate. Commercial-intent queries held up well overall, retaining Reddit 89.5% of the time, but the spread is wide enough that the average is a poor guide to any single category.
One caveat before any of this: Reddit's visibility is specific to Google. On Bing the same keywords return Reddit 3.1% of the time against Google's 69.3%. This is a bet on how one search engine ranks one platform, and on the assistants that read from it.
What Changed Since the Last Report?
The January 2026 report is preserved at its own dated URL, /research/2026-01, along with its raw dataset. This wave adds two sources and changes one, which affects how the numbers compare.
New: Bing as a control engine. January measured Google only. One engine cannot separate "people research on forums" from "Google ranks Reddit". We now run every keyword on Bing too, through the same provider, at the same depth, on the same day.
New: three AI assistants. ChatGPT, Perplexity and Gemini are each asked the same 1,300 questions, and we record whether Reddit appears in the sources each one cites.
Changed: the SERP data provider. January was collected through Serper. By July, Serper's Google endpoint returned only about 9 organic results per query regardless of the depth requested, while Google itself still served around 28. A like-for-like re-run was not possible through it, so this wave uses DataForSEO, which returns full depth. We measured the two providers against each other on the same queries the same day before switching: they agree within roughly 1 point at depth 10 and beyond, and on 88.9% of individual queries at top-9. Every January-to-July comparison on this page is depth-30 against depth-30.
A provider change mid-study can manufacture a trend on its own. This one would have shown a false 6-point decline if we had not checked. The matched-depth figure is larger, and at some cutoffs it moves the other way.
Abstract
Background: Reddit's rise in Google search is well documented [3], and is widely assumed to reflect a broad shift toward peer opinion in product research. That assumption has not been tested against a second search engine.
Methods: 1,300 keywords across 26 industries (20 consumer, 6 B2B), 50 per industry, split by search intent. Each keyword was queried on Google and Bing to a depth of 30 organic results via DataForSEO, capturing Reddit, YouTube, Facebook, X and LinkedIn placements from the same response. The same queries were then put to ChatGPT (gpt-5-search-api), Perplexity (sonar-pro) and Gemini (gemini-3.5-flash with Google Search grounding), recording whether Reddit appeared in each answer's cited sources. The January wave was re-analysed at matched depth for comparison.
Results: On the 1,273 keywords both engines returned, Reddit appeared on 69.3% of Google result pages, and in the top 3 on 50.0% of them. That is twelve times YouTube's top-3 rate of 4.2%, even though YouTube appears on more pages overall (91.1%). On Bing, Reddit appeared on just 3.1% of the same queries, a 22.6x gap, more than double YouTube's 9.9x and the widest of any platform we measured. Against January, Reddit lost 13.2 points of depth-30 presence (McNemar χ²=114.2), but the loss was concentrated by intent: commercial queries retained Reddit 89.5% of the time versus 68.3% for transactional (χ²=43.2, p<0.001), and presence at position 1 rose 2.4 points. The three AI assistants diverged sharply: Perplexity tracked Google almost exactly (68.5% vs 69.3%; φ=+0.679), citing the identical thread in 90.4% of overlapping cases; Gemini cited Reddit in 16.7% of answers with a weak Google gradient (φ=+0.116); and ChatGPT cited Reddit in 6.0% with no correlation to Google rank at all (φ=−0.024).
Conclusions: Reddit's visibility is specific to Google, and it comes from ranking high rather than from appearing often. Commercial-intent queries were largely unaffected by the six-month decline. The three AI assistants behave differently enough that they should be treated separately: Perplexity follows Google's ranking closely, Gemini only loosely, and ChatGPT not at all, with Reddit appearing in 68.5%, 16.7% and 6.0% of their answers.
Key Findings
50%
of all searches put a Reddit thread in Google’s top 3
22.6x
more Reddit on Google than on Bing, the widest gap of any platform
89%
of commercial-intent queries kept Reddit since January
68% / 17% / 6%
of Perplexity / Gemini / ChatGPT answers cite Reddit
Is Reddit Actually Visible, or Just Present?
Measured to a depth of 30, YouTube appears on more Google result pages than Reddit: 91.1% against 69.3%. The two platforms sit in very different places on those pages.
| Platform | On page | In top 3 | Avg position |
|---|---|---|---|
| 69.3% | 50% | 3.4 | |
| YouTube | 91.1% | 4.2% | 10.6 |
| 27.7% | 1.1% | 13.2 | |
| X (Twitter) | 14.1% | 2.2% | 12.6 |
| 3.5% | 0.5% | 11.9 |
YouTube averages position 10.6, which is page two. Reddit averages 3.4 and lands in the top three on half the queries we tested. Almost all of YouTube's coverage sits below the point where clicks drop off, so counting appearances alone gives a misleading read on both platforms.
Does This Hold Outside Google?
If Reddit ranks because people research on forums, a second search engine should show something similar. We ran the identical 1,273 keywords on Bing, through the same provider, at the same depth, on the same day.
22.6x
more Reddit on Google than Bing (69.3% vs 3.1%)
9.9x
more YouTube on Google than Bing (91.1% vs 9.2%)
5.3x
more Facebook on Google than Bing (27.7% vs 5.3%)
Every platform does better on Google, so part of this is just that the two engines index and rank differently. The size of the gap is what separates Reddit: 22.6x, against YouTube's 9.9x and LinkedIn's 2.1x. Bing carries plenty of user-generated content, with YouTube on 9.2% of queries, X on 6.4% and Facebook on 5.3%. Reddit lands at 3.1%, below every other major platform.
The subset relationship is equally stark: of 1,273 keywords, only five surfaced Reddit on Bing but not on Google. Bing's Reddit presence is very nearly a strict subset of Google's.
So Reddit's position in search is not a neutral reflection of what searchers prefer. It is specific to one engine, the one that signed a $60M/year content agreement with Reddit in 2024 [2]. We cannot prove causation from this data. But an explanation based on user preference still has to account for why the effect is ten times weaker on a competing engine answering the same queries.
Is the Opportunity Growing or Shrinking?
Running an identical keyword set twice isolates the change. Across 1,299 paired keywords at matched depth, Reddit's presence fell from 82.8% to 69.5%, a 13.2 point decline, with 214 keywords losing Reddit and 42 gaining it (McNemar χ²=114.2, p≪0.001). But the decline is not uniform across the page:
Presence at position 1 rose, from 23.6% to 26.0%, and the losses grow with depth. Breaking the change down, 90% of it comes from queries dropping Reddit altogether rather than from surviving queries carrying fewer threads. Among queries that still show Reddit, the average number of threads barely moved, from 1.30 to 1.26. Roughly one query in five stopped returning Reddit at all; the rest look much as they did in January.
And the losses avoided commercial intent
Of the queries that showed Reddit in January, 10.5% of commercial ones lost it, against 24.6% of informational and 31.7% of transactional (χ²=43.2, df=2, p<0.001). "Best X", reviews and comparison queries, the ones a buyer runs before choosing, were left largely intact. The losses fell on generic how-to content and on "buy X" queries, where Google favours retailers anyway.
The headline number therefore overstates the commercial impact. Reddit lost thirteen points of raw presence while keeping most of the queries that lead to a purchase.
Does Reddit's Own Reporting Agree?
Reddit reported Q2 2026 on 30 July, one day before we finished this wave. The business is growing quickly, and its filings describe the same softening in Google referrals that our measurement picked up.
$805M
Q2 revenue, up 61% year over year
130.3M
Daily active uniques, up 18%
$6.18
Global ARPU, up 36%
53.2M
US daily uniques, down from 53.5M in Q1
Revenue grew 61% to $805 million, an eighth consecutive quarter above 60%, with net income of $253 million against $89 million a year earlier [5]. Weekly active uniques passed 500 million. By any commercial measure the platform is expanding.
The exception is the number closest to this study. US daily active uniques fell sequentially, from 53.5 million in Q1 to 53.2 million in Q2, and management described Google search referrals as "choppy in the quarter" and more volatile towards the end of it [6]. Analysts on the call pressed on AI Overviews intercepting logged-out visitors before they reach the site.
That is independent corroboration of what we measured. Our SERP data found Reddit losing 13.2 points of depth-30 Google presence between January and July; Reddit's own filings report choppy Google referrals and a US user decline over the same window. Two different methods, same direction, same period.
One forward risk worth pricing in. The Google data licensing agreement, reported at roughly $60 million a year, is set to expire in the first half of 2027, and Reddit has given no timeline for renewal [7]. The arrangement that coincides with Reddit's ranking advantage on Google, and does not exist on Bing, is not permanent. Anyone treating Reddit visibility as a durable channel should treat that renewal as a live variable rather than a settled one.
Which AI Assistants Actually Cite Reddit?
We asked ChatGPT, Perplexity and Gemini the same 1,273 questions and recorded whether Reddit appeared in the sources each one cited. The three assistants disagree with each other more than the two search engines do, so they are worth measuring separately.
Perplexity closely follows Google. Its Reddit rate of 68.5% is near Google's 69.3%, the two correlate at φ=+0.679, and where both surface a Reddit thread it is the same thread 90.4% of the time, down to the post ID. Google's ranking predicts it closely:
Rank a thread in Google's top 3 and Perplexity cites Reddit 95.8% of the time. Push it past rank 10 and that falls to 42.5%. Gemini shows the same gradient far more weakly, at 21.1% for rank 1–3 against 10.2% when Reddit is absent from Google (φ=+0.116). ChatGPT's line is flat: 5.0% at rank 1–3, 6.9% when Reddit is absent altogether (φ=−0.024).
Ranking on Google therefore gets you Perplexity almost outright, a modest lift in Gemini, and nothing measurable in ChatGPT.
The levels matter as much as the correlations. Reddit appeared in 68.5% of Perplexity answers, 16.7% of Gemini answers and 6.0% of ChatGPT answers. The claim that Reddit is the primary feedstock of AI answers holds for Perplexity, partially for Gemini, and not at all for ChatGPT on the queries we tested.
Gemini moved the most since January, rising from 10.6% to 16.7%. It also changed behaviour: it now answers without searching at all on 22.8% of queries, against 0.1% in January. Measured only across the answers where it did search, its Reddit rate is 21.7%. Both bases agree on the direction; we report both because we cannot fully separate a genuine model change from the different routing used this wave.
For planning purposes these are three separate budgets. Perplexity visibility comes with Google ranking, Gemini is loosely coupled to it, and ChatGPT is a separate problem with no known lever. ChatGPT and Perplexity agree with each other less (φ=−0.034) than either agrees with a search engine.
Which Industries Should Care?
Technical and B2B categories, where Reddit both ranks highest and held up best since January:
All six lost under 11% of their Reddit queries since January, against 40% in Legal Services and 46% in Real Estate. If you sell developer tools, software or anything B2B, this is the strongest version of the channel and it did not degrade.
The categories with the least to work with are Legal Services (18.4% top 3), Healthcare/Medical (26.0%) and Beauty & Skincare (30.0%). Full table below, sorted by top-3 rate. Each industry is 50 keywords, so a single figure carries a 95% confidence interval of roughly ±14 points. Read the ordering and the broad tiers, not small differences between neighbours.
| Industry | In top 3 | On page | Bing | Lost since Jan |
|---|---|---|---|---|
| Developer ToolsB2B | 76% | 94% | 8% | 4.5% |
| Software & Apps | 72% | 94% | 4% | 2.2% |
| Career & Education | 68% | 90% | 2% | 4.5% |
| Technology/Electronics | 68% | 86% | 4% | 11.1% |
| B2B SaaS/SoftwareB2B | 64% | 90% | 6% | 4.4% |
| CybersecurityB2B | 64% | 78% | 4% | 10% |
| Gaming | 60% | 76% | 4% | 22.4% |
| Camping & Outdoor Gear | 58% | 76% | 6% | 21.7% |
| Photography | 58% | 70% | 4% | 23.9% |
| Mental Health | 56% | 84% | 6% | 7.3% |
| Fashion | 54% | 68% | 0% | 19% |
| Music & Audio | 54% | 80% | 2% | 15.6% |
| Parenting | 54% | 68% | 0% | 25.6% |
| Travel | 54% | 84% | 4% | 11.1% |
| Food & Cooking | 48% | 66% | 0% | 29.5% |
| Home Improvement & DIY | 48% | 60% | 4% | 31.8% |
| Automotive | 46% | 54% | 0% | 27% |
| Home & Appliances | 44% | 62% | 10% | 35.4% |
| Sports & Outdoors | 44% | 68% | 2% | 17.1% |
| Pets | 42% | 66% | 2% | 23.3% |
| Health & Fitness | 32% | 64% | 0% | 11.8% |
| Personal Finance | 32% | 54% | 0% | 31.6% |
| Real EstateB2B | 32% | 42% | 0% | 46.2% |
| Beauty & Skincare | 30% | 50% | 0% | 34.2% |
| Healthcare/MedicalB2B | 26% | 50% | 6% | 17.4% |
| Legal ServicesB2B | 18.4% | 32.7% | 0% | 40% |
Technical and B2B categories both rank highest and held up best: Developer Tools and Software & Apps sit above 70% top-3 with loss rates under 5%. The categories that gave up the most ground were Real Estate (46.2% lost), Legal Services (40.0%), Home & Appliances (35.4%) and Beauty & Skincare (34.2%), where listing portals, local packs and publisher content compete hardest.
What Should I Do About It?
- 1.Half of your category's searches show a Reddit thread in the top three. That is where clicks concentrate. If the thread does not mention you, either a competitor is mentioned or nobody is.
- 2.Check whether your queries still return Reddit. One in five stopped since January, and the pattern varies by category. It is measurable per keyword, and worth checking before committing budget to a category that has lost the channel.
- 3.Commercial-intent queries held. In software, developer tools and B2B categories the loss rate was under 5% and top-3 placement is above 60%. The channel did not degrade there.
- 4.Treat AI visibility as three separate things. Perplexity comes with Google ranking, usually citing the same thread. Gemini gives a modest lift, 21% at rank 1–3 against 10% when Reddit is absent. ChatGPT gives none, citing Reddit in 6% of answers however you rank.
What Did the First Wave Find?
The first wave is preserved as its own page, exactly as published, so this update can be checked against it rather than replacing it: read the January 2026 report. It covered Google only, across the same 1,300 keywords and 26 industries, to a depth of 30 organic results via Serper. Its headline figures were:
82.8%
Reddit present in the top 30
4.2
Mean position (median 2)
97.5%
Commercial-query presence
54.3%
In the top 3
By intent, January read: commercial 97.5%, informational 77.4%, transactional 68.3%. The strongest industries were Gaming (98%), Home & Appliances (96%) and Software & Apps (92%); the weakest were Health & Fitness (68%), Legal Services (51%) and Healthcare/Medical (46%).
The January headline number was correct. Recomputed from its raw data it gives 82.8% at depth 30, matching the 82.7% published at the time. What changed since is the context around it: a second engine to compare against, three AI assistants, and a data provider that still returns full depth.
Its main limitation, clear only in hindsight, is that one search engine cannot separate a claim about searchers from a claim about Google. The Bing comparison added this wave answers that, and it changes the interpretation.
Read the full January 2026 report →Download the January 2026 dataset (CSV)
How Did We Measure This?
1,300 keywords across 26 industries (20 consumer, 6 B2B), 50 per industry, categorised by intent: informational (~50%), commercial (~34%) and transactional (~16%). The identical set was collected in January and July 2026.
July data was collected via DataForSEO to a depth of 30 organic results, US location, English, on both Google and Bing. A single request returns every organic URL, so Reddit, YouTube, Facebook, X and LinkedIn placements all come from the same response rather than from separate studies at different depths. Median depth returned was 28 for Google and 20 for Bing.
AI answers were collected by putting each keyword to ChatGPT (gpt-5-search-api), Perplexity (sonar-pro) and Gemini (gemini-3.5-flash) and inspecting the structured citation payload each returns. "Cites Reddit" means a reddit.com URL in those citations, not a mention of the word Reddit in prose. For ChatGPT the two nearly always coincide, with only one answer in the sample naming Reddit without a corresponding citation. Gemini returns grounding redirects rather than destination URLs, so its sources are identified by the domain Google places in the citation title; we verified against live redirect resolution that the two agree, and we resolve the Reddit citations themselves to recover the underlying thread.
Provider change between waves. January was collected via Serper at depth 30. By July, Serper's Google endpoint returned only ~9 organic results regardless of the depth requested, while Google itself still served ~28, so we switched to DataForSEO for the July wave. To bound the effect of that switch we ran both providers on the same queries the same day: they agree within roughly 1 point at depth 10 and above, with 88.9% per-query agreement at top-9. All January-to-July comparisons here are depth-30 to depth-30.
What This Study Cannot Tell You
- Cross-provider comparison. The two waves used different SERP providers, calibrated at ~1 point of difference. A 13.2 point change is far outside that, but smaller differences in this report should not be over-read.
- Industry figures are noisy. At 50 keywords each, a single industry number carries a 95% CI of about ±14 points, and a change between waves about ±20. We report industry-level loss rates (which condition on a larger base) and broad tiers rather than small per-industry movements.
- Organic results only. We measure the organic list. Neither engine's AI overview, video carousel, discussion module or ad slots are counted, and Bing may place social content in modules our data does not capture. The Google-versus-Bing comparison is a statement about organic placement, not about everything a user sees.
- Gemini was collected through a different route. Google Search grounding on our direct API key returned quota errors for every request, so the Gemini wave was collected via OpenRouter, which passes through to Google's own grounding. The citations come back as
vertexaisearch.cloud.google.comgrounding redirects, the same shape the direct API returned in January. Gemini also declined to search on 22.8% of queries versus 0.1% in January, so we report its rate on both bases: 16.7% across all answers, 21.7% across answers where it actually searched. - US, English, desktop, one point in time. Results may differ by region and device, and Google's ranking changes continuously.
- Correlation, not causation. The Google-versus-Bing gap is consistent with the Google–Reddit agreement but does not establish it as the cause.
Sources and Raw Data
[1] DataForSEO. (2026). SERP API — Google and Bing organic results.dataforseo.com
[2] Reuters. (2024). Google and Reddit — $60M/year content licensing agreement.Reuters
[3] Sistrix. (2026). Reddit domain analysis — peaked at #2 most visible US domain in August 2025, since around #4.Sistrix
[4] Amsive. (2026). Google March 2026 core update analysis.Amsive
[5] Reddit, Inc. (2026). Q2 2026 results, reported 30 July 2026: revenue $805M (+61% YoY), net income $253M, DAUq 130.3M (+18%), WAUq 514.6M (+24%), global ARPU $6.18 (+36%).CNBC
[6] Reddit, Inc. (2026). Q2 2026 earnings call and shareholder letter — Google search referrals described as "choppy in the quarter"; US DAUq 53.2M against 53.5M in Q1.Earnings call transcript
[7] TechCrunch. (2026). Reddit reports a solid quarter but shows signs of AI's impact — Google data licensing agreement expiring in the first half of 2027.TechCrunch
About This Research
Mentioned runs this study to understand where product discovery happens. We publish the full datasets, including the results that complicate our own pitch. The Bing comparison and the ChatGPT numbers both cut against a simpler story.
Published: January 5, 2026
Last updated: July 31, 2026 (second wave — Bing, ChatGPT and Perplexity added; all figures recollected at depth 30)