A Q&A on audience ownership, algorithm risk, and why the advice you give every client is the one you never took.

You tell every client to own their audience. Then you build yours on rented land, from a landlord who can raise the rent tonight, under a lease you never read. I have 28,000 followers on a platform I do not own, cannot export, and cannot reach without permission from a ranking system that changes without telling me. Below is the argument, in the form of the questions people actually ask me, with the receipts attached.

1. What do you actually mean by “rented land”?

1. Reach on a platform is revocable and repriceable without notice, and the repricing is retroactive. Facebook disclosed in 2012 that the average Page reached about 16% of its own fans. EdgeRank Checker measured that fall to 6.5% by March 2014. Social@Ogilvy found Pages above 500,000 likes down near 2%. Current estimates run 1% to 6%, with large Pages often under 2%.
2. So a Page with 100,000 followers reaches roughly 2,200 people per post. That is a 90% plus writedown on an asset you spent years building, applied by a counterparty who never asked your opinion. Followers are a relationship with a landlord. An email list is a line item.

2. Fine, but did an algorithm change ever kill real companies with real payroll?

1. Google shipped Panda on February 23, 2011. Google said it affected roughly 12% of US queries. Suite101 lost about 94% of its search traffic. Matt Cutts of Google was asked about it publicly and said he felt confident about the algorithm, which is a very expensive sentence to be on the wrong end of.
2. Demand Media, the eHow parent publishing a reported 7,000 pieces of content a day, took a $6.4 million quarterly loss and saw its share of Google downstream traffic fall roughly 40% between January and mid-April 2011. Mahalo, About.com, HubPages and EzineArticles all got hit.
3. Ignite Visibility estimates more than 80% of sites damaged by Panda never fully recovered. These were companies with staff, offices and revenue. They were tenants who thought they had a deed.

3. That was fifteen years ago. What is the 2026 version of the same story?

1. Pew Research Center tracked the real browsing behavior of 900 US adults across 68,879 Google searches and published the results on July 22, 2025. When an AI Overview appeared, users clicked a traditional search result 8% of the time. Without one, 15%. That is a 47% relative collapse in clicks.
2. Clicks on links inside the AI summary itself: 1%. Sessions ending after the page: 26% with an AI Overview versus 16% without. That traffic does not move to a competitor. It stops existing.
3. SparkToro and Datos put 58.5% of US Google searches ending with no click to any external site. Chartbeat, measuring more than 2,500 publisher sites from November 2024 to November 2025, found search traffic down 60% for small sites, 47% for mid-sized and 22% for large ones. Ahrefs measured position-one click-through rate on AI Overview keywords falling from 0.073 to 0.016.

4. Everyone cites Airbnb on this. What did Airbnb actually do?

1. In 2019 Airbnb spent $1.62 billion on sales and marketing, about 23.7% of revenue. In 2020 it cut combined brand and performance marketing by 58%, from $1.14 billion to $482 million. Performance marketing absorbed $541 million of the $662 million reduction, more than four times the cut to brand.
2. Traffic came back to roughly 95% of 2019 levels with performance spend at effectively zero. By Q4 2020, 91% of Airbnb traffic was direct or unpaid. In Q1 2021 Chesky told investors the figure was 90% unpaid or direct while marketing spend was down 50%.
3. Marketing fell to 14.2% of revenue in 2020 against 23.7% in 2019, and Airbnb went into its December 2020 IPO with a profitable quarter. Chesky reframed the job of marketing as “education” rather than buying customers, and named PR and word of mouth as the things that built the brand.
4. The transferable insight is unglamorous: if your direct and branded traffic share is already above 70%, a large slice of your paid budget is buying visits you would have gotten anyway.

5. Is the Airbnb story too clean? Be honest.

1. Yes, partly. Skift read every Airbnb sales and marketing disclosure from 2020 through 2026 and found quarterly brand and performance spend climbed from $119 million to $512 million while the famous 90% figure never budged.
2. The reason is boring and important: brand advertising produces visits that later get counted as direct. Airbnb 2026 filings credit “paid growth initiatives” for the increase. The 90% held steady for six years while the budget quadrupled underneath it, and no untrue statement was made.
3. “Direct” is a counting decision, and so is most of what sits in your attribution dashboard. An email address is a row in a file you can export, verify and hand to a buyer.

6. What does the Airbnb story look like from the host chair instead of the boardroom?

1. New York City began enforcing Local Law 18 on September 5, 2023. Airbnb listings under 30 days fell from 22,246 in August 2023 to roughly 4,000 by May 2024, a drop of 82%. Skift measured the one-year decline at 83%. It never bounced back: as of June 2026, 3,500 hosts held legal registrations, about 15% of the roughly 23,000 listings active before the law. Fines run $1,000 to $7,500 per violation.
2. Then the platform repriced. In late 2025 Airbnb moved most hosts onto a mandatory host-only fee of 15.5%, rising to 17.5% with a strict cancellation policy. Hosts did not get a vote. On a $1,000 booking that is $155 leaving the room.
3. And here is the part that stings: on Airbnb you cannot email a past guest or offer them a returning-guest rate. The platform owns that relationship and guards it, which is precisely the behavior of a company that read its own filings.
4. Hosts who noticed have numbers. Hostfully reports direct bookings averaging 20% of reservations, up from 16% in 2021, with operators running 50-plus listings hitting 25% to 40%. Red Lodge Reservations in Montana went from 3% to 14% direct in one year on the back of a single automated post-checkout email carrying a booking link and a discount code.
5. Lodgify 2026 data shows direct bookings producing 51.3% longer booking windows and 45.2% longer stays, and StayFi found 37.5% of operators reported more direct bookings in 2025 than the year before. The winning tactic is charmingly dumb: trade the WiFi password for an email address.

7. What is a list worth when somebody actually writes a check?

1. Morning Brew: Insider Inc. closed an all-cash deal in October 2020 for a majority stake at a valuation around $75 million. The flagship had 2.5 million subscribers, about 3 million across all titles, on roughly $20 million of revenue, off a $750,000 seed round. Axel Springer bought the rest outright in February 2025.
2. The Hustle: HubSpot acquired it in February 2021. Axios reported roughly $27 million; HubSpot SEC filings recorded $20.3 million including $17.2 million in net cash. Subscribers at the time: 1.5 million. That is somewhere between $13.50 and $18 per email address.
3. No one has ever bought a company for its follower count. Lists appear in a data room and survive an acquisition. Followers appear in a screenshot.

8. About that $36 return. Is the number real or is it marketing about marketing?

1. Litmus puts average email return at $36 for every $1 spent, rising to $45 in retail, ecommerce and consumer goods. Sending 5 to 8 emails a month averages the highest return at $48 per dollar. Lists above 10 million subscribers return $46 versus $39 for lists under 500,000.
2. The $42 figure you see everywhere is a 2019 UK DMA number denominated in pounds, so pairing it with the Litmus dollar figure is not like-for-like. Cite it carefully or a smarter person will notice.
3. The honest reading: a chunk of the ratio is an artifact of how cheap the denominator is. Email costs almost nothing to send, which is the entire pitch rather than a strike against it.
4. The leverage sits in automation. Klaviyo reports flows delivering about 41% of email revenue from 5.3% of sends. Omnisend found automated messages producing roughly 41% of email-driven orders off 2% of volume. Litmus also found personalization pushing returns to 43:1 versus 12:1 for those who rarely bother, and A/B testers landing at 42:1 versus 23:1.

9. So why is my list dead?

1. Because there was never a recurring reason to open it. A list is a promise to be worth opening. If you cannot name the recurring reason before you build the signup form, you are constructing an obligation and calling it an asset.
2. Lists also rot on their own. HubSpot, drawing on MarketingSherpa research, puts database decay at about 22.5% a year, roughly 2.1% a month. B2B lists run faster, 25% to 30%, because people change jobs. Median US employee tenure hit 3.9 years in January 2024, the lowest since 2002.
3. You need 2% to 3% new subscribers a month simply to stand still. Decay happens whether or not you send. Active senders learn it from bounce reports; dormant senders learn it the day they finally hit send.

10. How do I choose the recurring reason?

1. Pick one thing that repeats and put it on a predictable date. An index. A teardown. A ranked list. A number that moves. Something a reader could describe to a colleague in one sentence.
2. Cadence without a reason is a chore. A reason without cadence is a rumor. You need both, and you need them before the signup form, not after.

11. Should I trade a PDF for the email address?

1. Trade a PDF and you get an address belonging to someone who wanted the PDF. Trade the recurring thing and you get an address belonging to someone who wants you. The second one still opens in eighteen months.
2. Litmus found single opt-in programs showing returns about 80% higher than double opt-in. Unless regulation forces your hand, the double opt-in orthodoxy is worth retesting.

12. Can I actually leave with my list? Have you checked?

1. Every email platform is also a platform. The difference is that this one lets you walk out with the file. Almost no one ever tests whether they can, which is the same energy as never testing a backup.
2. Export monthly and keep the CSV somewhere you control. Keep the sending domain authenticated too: since February 2024 Gmail and Yahoo require SPF, DKIM and DMARC from bulk senders, one-click unsubscribe, and spam complaints under 0.3%. Ownership without deliverability is a filing cabinet.

13. How should the thing actually sound?

1. Write to one person. The best newsletters read like a letter from someone slightly too candid. The worst read like a company that recently discovered the word “insights.” You already know which one gets opened at 6:40 in the morning in a parking lot.
2. Speed beats polish. GetResponse found 44.14% of clicks arrive within 60 minutes of send. Newsletters average a 3.84% click-through rate against 3.25% overall; triggered messages hit 5.02%.

14. What separates a newsletter that sells from one that just reports?

1. Reply to the replies. A newsletter that produces conversations is a sales channel. A newsletter that produces open rates is a report you write to yourself. The distinction lives entirely in whether the sender treats an answer as an interruption.
2. Open rates are also increasingly fiction. Apple Mail Privacy Protection inflates them, which is part of why quoted averages range from 31% to 42% depending on whose panel you read. Count replies, clicks and revenue.

15. Bottom line, in one paragraph.

1. Platforms are a distribution channel and a good one. Airbnb is excellent at putting a property in front of someone who has never heard of it, and social is excellent at putting you in front of someone who has never heard of you. Use both for discovery.
2. Then build the thing they cannot take from you, cannot reprice at 15.5%, and cannot bury under an AI Overview. I have 28,000 followers on land I rent. I have given this advice to every client I have ever had. It took me an embarrassingly long time to take it myself, and the only reason I am telling you is so it takes you less.

Sources

Pew Research Center, AI Overviews and clicks, July 2025: https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/

Search Engine Journal, Google Panda: https://www.searchenginejournal.com/google-algorithm-history/panda-update/

Search Engine Land, Panda fallout, 2011: https://searchengineland.com/your-sites-traffic-has-plummeted-since-googles-farmerpanda-update-now-what-66769

HubSpot, decline of organic Facebook reach: https://blog.hubspot.com/marketing/facebook-organic-reach-declining

Campaign, Airbnb 90% unpaid or direct: https://www.campaignlive.co.uk/article/airbnb-halves-performance-marketing-spend-100m-cut-q1/1716094

Campaign, Airbnb 58% marketing cut: https://www.campaignlive.co.uk/article/airbnb-slashes-spend-permanent-shift-performance-marketing-brand/1708621

Skift, Airbnb rebuilt the engine it cut: https://skift.com/2026/08/17/airbnb-has-quietly-rebuilt-the-marketing-engine-it-was-famous-for-cutting/

Skift, NYC near-ban on Airbnb: https://skift.com/2024/12/27/takeaways-from-nycs-near-ban-on-airbnb/

Lodgify, Local Law 18 one year on: https://www.lodgify.com/blog/local-law-18-one-year-report/

CRE Daily, NYC rentals still down 85%: https://www.credaily.com/briefs/nyc-short-term-rentals-still-down-85-after-2023-law/

Hostfully, direct booking vs. Airbnb, 2025: https://www.hostfully.com/blog/airbnb-vs-direct-booking-website/

Axios, Insider buys Morning Brew: https://www.axios.com/2020/10/29/insider-inc-buys-majority-stake-morning-brew

TechCrunch, HubSpot acquires The Hustle: https://techcrunch.com/2021/02/04/hubspot-acquires-media-startup-the-hustle/

Litmus, ROI of email marketing: https://www.litmus.com/resources/email-marketing-roi

HubSpot, database decay, 2.1% monthly: https://www.hubspot.com/database-decay