We checked who owns Google for 20 buyer queries. Reddit is number one for twelve. An actual product, for three.
When someone Googles "best X," you assume the top result is the best product. It is not. We classified the top ten for 20 buyer queries. The number one spot belongs to Reddit far more often than to any vendor, and when it is not Reddit it is usually a review site or an analyst, not a product either.
Original study · measured 2026-08-10
Most SEO strategy starts from a wrong picture of the enemy. You imagine the top of Google for "best CRM" or "best VPN" is a ranked list of the products, and your job is to climb it past your competitors. So we measured who actually sits there. We ran serp_profile across 20 "best X" buyer queries and classified every result in the top ten by what it is: a user platform (Reddit, YouTube), an authority (.gov, .edu, Wikipedia), or a commercial page.
The competitor you are fighting for the top slot is not a competitor.
Reddit owns the top of Google for buyer queries
Reddit sits in the top three for 19 of the 20 queries, 95%, and is the number one result for 12 of them. Not the top ten, the number one position, ahead of every product in the category. It is number one for best CRM for startups, best VPN, best web scraping api, stripe alternative, notion alternative, best postgres hosting, best password manager, best vector database, and more. A ten-year-old forum thread outranks the companies that sell the thing.
YouTube is in the top ten for 15 of 20. Between them, the two platforms take a slot on nearly every buyer query on the board.
An actual product is number one only three times
Here is the part that should reset your strategy. Of the 20 number-one results, only three are an actual product in the category: swagger.io for API documentation, proprofsdesk.com for customer support, and gumloop.com for marketing automation (and gumloop is an AI automation tool that ranks oddly for that term). The rest of the non-Reddit winners are not products either:
| query | number one result | what it actually is |
|---|---|---|
| best data observability tool | gartner.com | analyst, not a vendor |
| best project management tool | icagile.com | an Agile certification body |
| best transactional email api | emailtooltester.com | a review aggregator |
| best web analytics tool | wpmailsmtp.com | a WordPress plugin, off topic |
| sam.gov api | open.gsa.gov | government docs |
| twelve others | reddit.com | a forum thread |
You are not losing the number one spot to a better competitor. You are losing it to Reddit, to review sites, and to analysts. The product pages, all of them, are fighting for what is left below.
The type mix confirms it
Across all 184 classified results, 67% are commercial, 30% are user platforms, and just 3% are authority pages. That looks like a commercial-dominated SERP until you weight by position. Platforms hold number one on 12 of 20 queries; commercial pages hold it on 7; authority on 1. The commercial pages are the bulk of the page and the bottom of it. The visible, clicked, top of the page is Reddit.
There is no authority wall, which is the good news
Only 1 of 20 queries has any authority page (.gov, .edu, Wikipedia) in the top ten, and that is sam.gov api, where the government owns its own term. For ordinary commercial buyer queries there is no unbeatable institution sitting on the head. That matters: an authority-locked SERP is a wall you cannot climb, but a Reddit-locked one is not a wall, it is a different door. You do not out-rank the Reddit thread. You get recommended inside it, and you win the citation layer underneath it.
This is the same story as the citation layer
The citation_domains study found that when AI cites sources for these same buyer queries, the overlap with Google is mostly Reddit and YouTube. Put the two together: Reddit is eating the top of Google search and a large share of AI citations at the same time. The single highest-leverage move for a buyer query is often not a page on your own site at all. It is being present, by name, in the Reddit thread and the YouTube roundup that both channels keep pointing to.
What this means
serp_profile is built to stop you from misreading the SERP. Counting unique domains makes a Reddit-owned page look "fragmented and open" when it is neither. Classifying by type, and printing the names, tells you which of three situations you are in:
- Authority-locked (a .gov or .edu owns the head): rare for commercial terms, and genuinely hard. Compete on the citation layer instead of the ranking layer.
- Platform-locked (Reddit or YouTube owns the head): the common case, 12 of 20 here. The play is presence inside the thread, not out-ranking it.
- Contestable (real products hold the head): the case where classic on-page SEO is worth the spend, and it was only 3 of 20.
Run serp_profile(queries=["best X for Y"]) on your own buyer queries and read the number-one owner before you write a single page. It will usually tell you the page you were about to write is aimed at a slot Reddit already owns.
Method
20 "best X" and category buyer queries, one Grounder SERP call each, top ten organic results, on 2026-08-10. Each domain is classified by type: authority (.gov, .mil, .edu, Wikipedia), platform (Reddit, YouTube, GitHub, Medium, and similar), or commercial (everything else). Classification is by domain type, not backlink strength, so gartner.com and a small vendor both read "commercial," which is why the study reads the actual names. The "what it actually is" column for the number-one results is a manual reading of named domains, shown so you can verify each one. serp_profile is deterministic given the same SERP: no model, no score.
A point-in-time measurement. Search results and AI answers change, and grounded models vary between runs, so your own numbers will differ. Verithia measures. The interpretation is yours.