The Google Mirage
Type “best blogging platform 2026” into Google. Or “best ecommerce platform.” Or “best CRM,” “best project management tool,” “best password manager.” The same thing happens every time. The first page is wall-to-wall articles with the same structure: a numbered list, a paragraph or two of bland praise per item, a “winner” at the top, and an affiliate link on every product name.
It looks like research. It is not research.
Almost every “best of” article you see at the top of Google for a software category is a paid placement. Not in the sense that there’s a banner ad next to it — in the sense that the rankings themselves are for sale. The order is decided by who paid the website owner, and how much. The site owner does not disclose this. The reader has no way to know. And the products that bootstrap founders actually need — the small, well-built, fairly priced ones — don’t exist on that page, because their founders didn’t cut a five-figure check to a media broker.
I know this because I run two products that have been on the receiving end of those pitches. BlogCloud, a publishing platform under our Prana Entertainment portfolio. ShopSnap, an ecommerce builder under the same roof. Both got the call. Both have the receipts.
Here’s what those receipts actually say.
Exhibit A: BlogCloud and the NeilPatel.com Pitch
In October 2022 I filled out the “Partnership Contact Form” on NeilPatel.com asking how to get BlogCloud reviewed and listed in his “best blogging platforms” content. NeilPatel.com is one of the top-ranking marketing properties on the planet. His personal-brand site has historically published guides that pull seven figures of monthly visitors. If you are a blogging tool and you can crack that list, you exist. If you can’t, you don’t.
The response did not come from Neil. It came from a man named Stephen Harrell, Senior Business Development Manager at a company called Stone Press. Here is the cordial, professionally written first line of his reply:
“I’d love to set up a call to learn more about BlogCloud and discuss how we can potentially work together. Let me know your availability and I will confirm.”
Nothing in that sentence sounds like the editorial team of an “unbiased” review site. Nothing in his signature does either. Look at what was attached to his name:
• Stephen Harrell — Senior Business Development Manager
• StonePress.com | QuickSprout.com | NeilPatel.com | CrazyEgg.com
That is not an editor. That is an ad sales rep with a portfolio of properties. The call he wanted to set up was a sales call. The “potential partnership” was a quote.
Stone Press, by its own description on LinkedIn, is “a fully remote digital media company working on building the largest and most respected portfolio of B2B websites.” On its own homepage it says the team has experience “as publishers, advertisers, marketers, and founders — from building major review sites and running affiliate programs to growing SaaS blogs from nothing.” Translation: they run the review pages and they sell placements on the review pages. Both sides of the transaction. Same company.
When I asked Stephen what “working together” actually cost, the number that came back was in the tens of thousands of dollars. Not for a banner. For consideration. For the privilege of being evaluated. For a shot at appearing somewhere on a list that millions of readers treat as if it had been written by a journalist.
I did not pay. BlogCloud was not added.
Take a second to absorb the shape of that. Somewhere on NeilPatel.com’s “best blogging platforms” list is a product that did pay. The list does not say that. The article does not say that. The reader walks away believing the platforms on that page won on merit. They won on procurement.
Exhibit B: ShopSnap and the $1,290 Sticker Price
Different product, different category, same script.
In March 2023 I reached back out to Ecommerce-Platforms.com, a site that describes itself in its own bio as “an unbiased review site that shows the good, great, bad, and ugly of online store building software.” Run by Catalin Zorzini, a London-based founder who has built his personal brand on it for over a decade. I had emailed them about reviewing ShopSnap a while earlier, never heard back, and tried again.
This time I got a response. Not from Catalin — from a writer named Raul, replying from a help desk address. Here, verbatim, is what he wrote:
“Hi there, Thanks for getting in touch! We can do a dedicated sponsored review for $1,290. We get 500,000+ uniques a month. Let me know if you’re interested and I’ll send over more details.”
Let’s sit with that for a moment, because it’s almost beautifully honest if you read it the right way.
In one sentence we learn that the publication’s pricing is $1,290 per review. We learn that the deliverable is a “dedicated sponsored review.” We learn the traffic stat that justifies the price. What we don’t learn — because Catalin’s readers don’t — is that those words also describe the editorial product on the rest of his site. The same kind of write-up, the same format, the same byline conventions. A reader landing on Ecommerce-Platforms.com from a Google search for “best Shopify alternatives” doesn’t see two categories of content labeled “sponsored” and “editorial.” They see one site. One voice. One “unbiased review site that shows the good, great, bad, and ugly.”
Behind the curtain, $1,290 buys you the good. Not paying buys you the ugly — or, more often, total invisibility.
How the Racket Actually Works
There are two main flavors of this game and the two emails above are textbook examples of each.
Flavor one is the broker model. A company like Stone Press acquires, operates, or partners with high-authority review sites and runs partnership sales across the whole portfolio. They have a bizdev team. They have a CRM. They have a menu. When a founder reaches out, they pitch a package across multiple properties — QuickSprout, NeilPatel, CrazyEgg, sometimes a half dozen others. The economic logic is that one negotiation buys you a syndicate of placements, which is exactly how display ad networks work, except these are inside the editorial.
Flavor two is the solo operator model. One person built a property with content, ranked it on Google, and now monetizes by selling reviews directly. Catalin Zorzini at Ecommerce-Platforms.com is the textbook version: a personal-brand site, an affiliate disclosure buried in the footer, and a sales page no one reads that quietly admits the model. The reader, of course, never sees the sales page. They see the listicle.
Both models share three structural traits that make them functionally indistinguishable from advertising:
• Ranking is correlated with payment, not quality. The same product can be #2 on one site and absent from another, not because of objective measurement, but because the founder paid one and ignored the other.
• Negative information about paying advertisers is filtered out. A paid review is not going to be a hit piece. Nobody pays $1,290 to be told their product is mediocre.
• Non-paying products are excluded entirely. This is the cleanest form of bias because it doesn’t require lying. It just requires omitting. The reader can’t critique products they’ve never heard of.
If you put those three traits on a piece of paper and removed the word “review,” no reasonable person would call it journalism. They would call it a media kit.
What’s Missing: The Words “Paid Placement”
This is the part that should make you angry, and the part where the law has slowly started to catch up.
Every site I just named has some version of a disclosure page. QuickSprout’s About page, for example, very politely says: “I keep the business running by generating revenue through affiliate links in the content. For some of the products that are recommended, I may generate a commission. That said, my editorial guidelines are strict to maintain honest and unbiased reviews.”
Read that carefully. The disclosure is about affiliate commissions. Affiliate commissions are not the same thing as the pay-to-be-considered model Stephen Harrell was selling. Affiliate is downstream — you earn a cut when a reader clicks through and buys. Paid placement is upstream — you pay to appear on the list at all, regardless of whether anyone ever buys anything. The disclosure language obscures the second by talking only about the first. Even if a reader notices the affiliate disclosure (almost none do), it doesn’t tell them that the products listed paid to be there. It strongly implies the opposite — that the editorial was “strict,” “honest,” and “unbiased.”
This is the entire scam in one paragraph. Disclose the small thing in legalese, hide the big thing entirely, then attach the word “unbiased” to the same sentence.
The FTC Caught Up. Sort Of.
In 2024 the Federal Trade Commission finalized its long-discussed rule against fake reviews and testimonials. Most of the press coverage framed it as being about Amazon-style five-star manipulation — bot reviews, paid Yelp reviews, employees pretending to be customers. That’s in there. But buried in the rule is something more interesting: it explicitly targets, in the FTC’s words, review sites controlled by a single company but presented as independent sources, and reviews where material connections aren’t clearly disclosed.
Reread that and look at the Stone Press business description again. “We’ve been on all sides of performance marketing… as publishers, advertisers, marketers, and founders.” That sentence is a confession. The same company is producing the review and selling the placement. The new FTC rule says that’s a deceptive practice when not clearly and prominently disclosed.
Will it be enforced against the marketing-industrial complex of “best software” review sites? Unclear. The FTC is not exactly overstaffed, and these sites are not exactly Amazon. But the law now agrees, in writing, with what every founder who has ever gotten one of these emails already knew: this is not editorial. It is advertising in a costume.
Why This Is Worse Than It Looks
It is tempting to file this under “internet has always been like this” and move on. Don’t. The damage of pay-to-rank review sites is bigger than “some lists are biased.”
It distorts what gets built. When the only way to be discovered is to write a five-figure check, you select for founders who have raised venture capital and have a budget line called “distribution.” You select against bootstrap operators, indie hackers, and small teams shipping focused, well-priced products. Over a decade of this, the SaaS landscape skews toward overfunded, over-marketed, mediocre tools because those are the tools that can afford the placement fees. The thing people complain about in B2B software — that everything is bloated, overpriced, and indistinguishable — is partly caused by this exact selection pressure.
It corrupts buyer decisions. Imagine you’re a small business owner. You search “best CRM for solopreneurs.” You read the top three articles. You pick the platform that appears in two of them. Twelve months later you’re locked into a subscription you hate, with data you can’t easily export, paying double what you should because the actually-best tool for your case was a $19/month indie product nobody paid to be listed. The pay-to-rank model didn’t just take a fee from a vendor — it transferred wealth from you to the bigger, badder option. Quietly. Without your knowledge or consent.
It poisons the training data for AI. Every general-purpose large language model was trained on the open web. That web is saturated with these listicles. When you ask an LLM “what’s the best blogging platform,” it has read a thousand articles that say the same six names in different orders, all written under financial pressure. Without intervention, the model just regurgitates the consensus of an SEO landscape that was bought and paid for. The bias compounds.
It debases the word “review.” Once enough people figure out that “best of” lists are paid, they stop trusting any review they read online. The legitimate ones — Wirecutter, Consumer Reports, Ars Technica, RTINGS — get tarred with the same brush. That distrust is a tax on every honest publisher who didn’t take the money.
The AI Counterattack
Here is the part where I tell you what to do instead, because complaining without a remedy is content marketing for cynicism.
The good news is that the same general-purpose LLMs that have been trained on the polluted listicle web have also been trained on the rest of the web — the documentation, the GitHub repos, the developer forums, the unhappy customer threads on Reddit, the trade press, the engineering blogs, the actual reviews on G2 and Capterra (with a grain of salt; those are also gamed, but less totally), and the product home pages themselves. You can use them to triangulate around the SEO swamp instead of through it.
Here is a workflow I use when I need to choose software in a category I don’t already know.
Step 1. Define your actual constraints, in writing. Not “best ecommerce platform.” That’s the question Google is built to monetize. Instead: “I sell three physical products, I ship from one warehouse in Nevada, my margins are thin so transaction fees matter, I want to own my customer data, and I won’t pay more than $50/month.” Now you have a query no listicle can answer.
Step 2. Ask an LLM for a wide map, not a winner. Prompt: “Give me 15 ecommerce platforms that fit these constraints, including small or indie ones that don’t typically appear in ‘best of’ articles. For each, give me the company size, the pricing model, the transaction fee, the data export story, and one specific complaint from real users.” The trick is the explicit instruction to include products that don’t rank. That alone gets you off the SEO treadmill.
Step 3. Force the model to disagree with itself. Take its top three picks and run a second prompt: “Steelman the case against each of these. What does a frustrated former customer say? What is the specific failure mode in production?” LLMs are bad at picking winners. They are surprisingly good at criticism when you ask them for it directly. Use that.
Step 4. Verify against primary sources. For the two or three that survive, go to the actual product’s site, their changelog, their status page, their GitHub if they’re open source, their subreddit, and ideally a long-form customer thread on Hacker News. You’re looking for signal that the LLM hallucinated nothing critical — the pricing, the limits, the data ownership terms.
Step 5. Skip listicles entirely. If a search result has the structure “Top 10 X for 2026,” “Best X Compared,” “X vs Y vs Z (We Tested Them All),” close the tab. Those are pay-to-rank artifacts roughly 90% of the time, and the other 10% are content marketing by the vendors themselves.
This is not a perfect workflow. LLMs hallucinate prices. They miss recent product changes. They will occasionally invent a feature that doesn’t exist. But the error mode of an LLM trained on the open web is closer to “genuinely confused” than the error mode of a $1,290 sponsored review, which is “deliberately misleading.” I will take confusion over deception every day of the week.
The deeper point is this: AI breaks the economic model of pay-to-rank review sites the same way it’s breaking the economic model of every SEO-driven publication. People are not going to keep clicking on listicles when an assistant can answer the actual question. The traffic that funds the Stone Press portfolio is the traffic that ChatGPT, Claude, Perplexity, and Google’s own AI Overviews are quietly siphoning off. The five-figure placement fee gets harder to justify every quarter, because the eyeballs are evaporating.
The racket is not going to die because the FTC writes a strongly worded rule. It is going to die because the underlying business no longer cash-flows.
What an Honest Review Site Actually Looks Like
There are still review properties that operate on something close to journalistic standards, and they are worth pointing to so the comparison is concrete.
Wirecutter, owned by The New York Times, buys its own products, lab-tests them, publishes its methodology, and discloses affiliate revenue prominently. Consumer Reports refuses advertiser money entirely and is funded by subscriptions. RTINGS posts raw test data for every TV they review. Ars Technica labels sponsored content with a giant orange banner. None of these outlets are perfect, but they all share one trait that the pay-to-rank crowd lacks: the article you read is not for sale.
If a software “review site” doesn’t do at least the following, treat its rankings the same way you’d treat a Yelp review from the restaurant’s owner:
• Names the reviewer, with credentials, on every article.
• Publishes the methodology — what was tested, how, on what hardware, with what data.
• Discloses sponsorships and paid placements above the fold, not in a footer link.
• Includes products that didn’t pay, identified by name.
• Has a public correction policy and uses it.
Hold the standard. The fact that almost no “best software” site clears that bar tells you exactly how much of the SEO industry has been content-marketing in disguise the entire time.
Closing the Loop
BlogCloud is still running. So is ShopSnap. So are the rest of the properties under Prana Entertainment. We never paid Stone Press, we never paid Ecommerce-Platforms.com, and we are not on those lists. The customers who find us — and they do find us — are people who searched intelligently, asked an AI assistant, talked to a friend, or stumbled in through a side door we built ourselves. That’s a smaller funnel than buying placement on NeilPatel.com would be. It is also a more honest one.
If you’re a founder reading this and one of those partnership emails is sitting in your inbox right now, you already know what to do with it. If you’re a buyer reading this, you know what to do with the next listicle that lands at the top of your Google results. And if you’re running one of these sites, I’d gently note that the FTC has your business model on paper and your traffic curve is going the wrong direction.
The internet sold you a search result and called it a recommendation. It was an ad. It always was an ad. The receipts are right there, in the email I just walked you through, with a name and a price tag and a polite signature.
Stop trusting the list. Start asking better questions. Use the new tools — they were built to be useful in exactly the way the old tools were built to be sold.
Gal Ratner is Chief Architect at Prana Entertainment and a 30-year veteran of the Microsoft/.NET stack. He writes about software, AI, and the tech industry at https://www.blogcloud.io/b/Gal-Ratner