The Business That Rented Its Customers From Google

What happened

A small team spent four years building a product-review site, buying and testing products and publishing long comparisons, with search delivering nearly all the readers. Google's September 2023 helpful content update rolled out over two weeks, and independent review sites began reporting sharp declines. In March 2024 Google announced a core update and said plainly it expected to cut low-quality, unoriginal content in results, publishing three new spam policies alongside it. The update finished rolling out on 19 April 2024, and Google later reported the result had come in above target at 45% less low-quality content.

The case, start to finish

Nothing about the site changed. The ranking function did.

A documented pattern, told through a composite publisher

The publisher here is an anonymized composite. The events are documented: Google's September 2023 helpful content update and its March 2024 core update, both announced by Google itself. The traffic figures belong to one real casualty, the independent air-purifier review site HouseFresh, which published on its own site that it went from about 4,000 daily visitors from Google Search to about 200.

The model is a good one on paper. Buy the products. Test them. Publish long, genuinely useful comparisons. Earn a commission when a reader buys. The publisher owns the writing, the test bench and the reputation, and for four years search delivered effectively all of the demand, so effectively all of the work went into ranking.

Two things were being rented, and neither appeared on any statement. The distribution belonged to a search engine, with no contract, no account manager and no appeal. The commission rate belonged to the retailer: Amazon's published Associates schedule pays 3.00% on Home Improvement, 4.00% on most unlisted categories, and 1.00% on Grocery and on Health and Personal Care. A few pennies on a dollar of somebody else's revenue, from traffic somebody else controls.

The update that announced itself in advance

The September 2023 helpful content update rolled out between the 14th and the 28th, per Google's own status dashboard, and independent review sites began reporting sharp declines. On 5 March 2024 Google announced the core update and said plainly what it was for, expecting the combination of that update and its previous work to reduce low-quality, unoriginal content in search results by 40%. Three new spam policies arrived alongside it, including expired domain abuse and site reputation abuse. The rollout finished on 19 April, and Google later reported the outcome above target, at 45%.

From inside a publisher, that number is the whole problem. Whatever you think of where the line fell, 40% of a category of results is somebody's traffic, and there was no way to know in advance which side of it you were on. Nothing about the site had changed. Same articles, same products, same test bench, and the demand simply stopped arriving.

Operating leverage, running backwards

What made the collapse unrecoverable rather than merely painful was the cost structure. Almost everything in this model is fixed and already spent. The articles exist. The products were bought and tested. The writers were paid. There is no cost-of-goods line that falls when traffic falls, so there is nothing to trim into survival.

That is operating leverage running in reverse. The same structure that makes 4,000 daily visitors extraordinarily profitable makes 200 daily visitors a full cost base with no revenue under it. And the ground kept moving beneath even the traffic that remained: the Pew Research Center, analyzing 68,879 Google searches by 900 U.S. adults during March 2025, found users clicked a result on 8% of visits where an AI summary appeared, against 15% of visits where none did.

The rebuild worked by giving up on the channel. Instead of re-engineering the site around guesses about the new ranking signals, the composite put its effort into demand it owns: a newsletter, video, a community. HouseFresh's published response had the same shape. Revenue came back as a fraction of the peak, which is the honest cost of four years spent building on rented ground, but it is a fraction nobody else can switch off.

Rented reach, owned relationships

The general form of this has nothing to do with search in particular. Marketplaces, app stores and social feeds all offer the same trade: enormous reach now, on terms the other party sets and can change without telling you. That is an excellent way to find customers and a dangerous way to hold them. Platform risk is not the risk of doing something wrong, it is the risk of the rules changing while you are doing everything right.

The practical measure is a question you can answer this afternoon. If the largest channel went to zero tomorrow, how many customers could you still reach directly, by name? Whatever that number is, it is the real size of the business. Everything above it is borrowed.

So spend part of every good quarter converting reach into a relationship: an email list, a customer file, a community, some reason to come back that does not route through a search box. It always feels less efficient than pouring the same effort back into the channel that is currently working, which is exactly why so few people do it while the channel is still working.

Timeline

• 2019–2023 A small team builds a product-review site: buy the products, test them, publish long comparison articles, earn a commission when a reader buys. Search delivers effectively all of the demand, so all of the work goes into ranking.
• 14–28 Sept 2023 Google's September 2023 helpful content update rolls out, per Google's own Search Status Dashboard. Independent review sites start reporting sharp declines.
• 5 Mar 2024 Google announces the March 2024 core update and states plainly what it is for: it expects the combination of this update and its previous work to "reduce low-quality, unoriginal content in search results by 40%."
• Mar 2024 Alongside the update Google publishes three new spam policies: scaled content abuse, expired domain abuse, and site reputation abuse, the last enforced from 5 May 2024.
• 19 Apr 2024 The core update finishes rolling out. Google later reports the result came in above target, at 45% less low-quality, unoriginal content.
• 2024–2025 The remaining clicks get scarcer. Pew Research Center, analysing 68,879 Google searches by 900 U.S. adults during March 2025, finds users clicked a search result on 8% of visits where an AI summary appeared, versus 15% of visits where none did.
• Rebuild The composite stops trying to win the channel back and starts building demand it owns: a newsletter, video, and a community. Revenue recovers to a fraction of the peak, but it is a fraction nobody else controls.

You're in the owner's chair

Six months after the update your search traffic is down roughly 95%. Your costs are already spent: the articles are written, the products are bought, the test bench is built. A well-reviewed consultancy offers to re-optimize the site for the new signals. A competitor has just bought an aged domain with existing authority to start over on. What do you do?

• Rebuild demand you own — newsletter, video, community
• Hire the consultancy and re-optimize for the new signals
• Buy an aged domain with existing authority and rebuild the content there

The same articles, the same tests, a different ranking function

• October 2023: 4,000 daily visitors from Google Search
• After the March 2024 core update: 200 daily visitors from Google Search

HouseFresh's own published account of its own traffic: “Since October 2023, we’ve gone from welcoming 4,000 people from Google Search each day to just receiving 200.” Nothing about the site changed between those two bars.

Business model

Publish genuinely useful buying advice, rank for the search someone types before a purchase, and take a commission on the sale. The publisher owns the content, the testing, and the reputation. It does not own the one thing the model runs on: the distribution.

Revenue model

Affiliate commissions on a rate card the publisher does not set. Amazon's published Associates schedule pays 3.00% on Home Improvement, 4.00% on most unlisted categories, 4.50% on physical books, 1.00% on Grocery and on Health & Personal Care. So the business earns a few pennies per dollar of somebody else's revenue, from traffic it also does not control.

Cost structure

Almost entirely fixed and already spent. The articles exist. The products were bought and tested. The writers were paid. There is no cost-of-goods line that falls when traffic falls, which is why this model prints money at scale and cannot be trimmed into survival. That is operating leverage running in reverse: the same structure that made 4,000 daily visitors extraordinarily profitable makes 200 daily visitors unrecoverable.

Strategic challenge

The publisher had one customer acquisition channel, on zero notice, with no contract, no account manager, and no appeal. Nothing about the site changed in September 2023 or March 2024. The ranking function did, and Google published its intent in advance, in a number: 40% less low-quality, unoriginal content. Whatever you think of where the line fell, every point of that reduction was somebody's traffic.

Key decision

Stop trying to win back a channel that had just demonstrated it could zero the business, and start building demand the business owns. HouseFresh's published response was the same shape: expand onto YouTube, Reddit, social platforms and a newsletter rather than re-engineer the site around whatever the new signals were guessed to be.

What worked

Owned channels, and the one asset the collapse proved was real: the testing. A publisher that actually buys and measures the products has something an algorithm change cannot delete: readers who came for the tests will follow them to an inbox. The newsletter recovered slowly and imperfectly, but the list is the only part of the funnel that no third party can reprice.

What failed

Everything that treated search as a durable asset: hiring against projected traffic, buying more product to test on the strength of last quarter, and, most expensively, measuring the business in sessions rather than in relationships. There was never a version of the site defensive enough to survive a change in what the ranking system was optimizing for.

Risk factors

Single-channel demand with no contractual relationship; an affiliate rate card set unilaterally by the retailer; a cost base that is fixed and sunk; AI summaries answering the query on the results page; and the slower structural risk of building a brand nobody can name, because every visitor arrived through a search box and left through a link.

Lesson summary

If one company can switch off your customers, that company is your business and you are its supplier. Search, marketplaces and social feeds are rented distribution: excellent for reach, catastrophic as a foundation. Measure what share of demand you could still reach tomorrow if the channel went to zero, and spend some of every good quarter converting rented reach into an owned relationship.

Key data

• 40% less low-quality, unoriginal content (later reported at 45%) Google's stated goal, March 2024 core update
• 14–28 September 2023 September 2023 helpful content update rollout
• 5 March – 19 April 2024 March 2024 core update rollout
• ~4,000 → ~200 daily visitors from Google Search HouseFresh's published traffic
• 8% of visits, vs 15% without (Pew, March 2025 data) Clicks when an AI summary appears
• 3.00% Amazon Associates rate, Home Improvement

Sources & basis

The business in this story is a stand-in, not a company you can look up. This case is an illustrative composite: the operator, the people and most of the dollar figures represent a pattern rather than reporting one firm's history. What the list below cites is the other half, the documented industry data and public reporting the composite was assembled from, including any real company whose published figures the case draws on by name. The mechanism and the arithmetic are real even where the business is not.

1. Google, "New updates to address spam and low-quality results" (5 March 2024) — the 40% target, the 45% reported outcome, and the three new spam policies including the 5 May 2024 site-reputation-abuse enforcement date View source ↗
2. Google Search Status Dashboard — September 2023 helpful content update rollout dates View source ↗
3. HouseFresh, “HouseFresh disappeared from Google Search results. Now what?” — the site's own published traffic figures and response View source ↗
4. Pew Research Center, "Google users are less likely to click on links when an AI summary appears in the results" (22 July 2025) — 68,879 searches, 900 U.S. adults, March 2025 browsing data View source ↗
5. Amazon Associates standard commission income rates View source ↗