What 118 Marketing Notes Keep Saying

A man filled ordinary tap water into wine bottles, added a fancy label, and told people it came from a glacier. Rich people believed it so much, by one account, that they said it was worth $500 to $1,000 a bottle. In the end, that “luxury water” sold for $80,000. The water never had a source. The story did.

That note — issue #92 — is one of 118 in this repository’s backlog.

Where the notes came from

This repository is where marketing notes have lived since 2023: a reel here, a link there, an essay half-finished, each saved as a GitHub issue. As of October 2026, that is 118 issues, and the tap-water note is their best advertisement.

For this post, every issue was researched one by one:

The full per-issue pass — 118 entries in ten clusters — is in the repository at .research/issues-research.md. What follows is what those entries say when they are lined up.

1. People don’t buy the thing

The first cluster is one observation repeated at different price points.

Issue #11 is a clip about a lemonade stand. Why would anyone buy from a stranger’s stand when a vending machine sells the same lemonade? From the machine transcript:

So when we see a kid with a lemonade stand, it’s different than when we see a vending machine selling lemonade, even if it’s exactly the same product, because the story around it is what people are paying for.

Issue #65 is a man comparing a local shop to a luxury house — the transcription hears the house as “Lora Piano” — and the point survives the mis-hearing:

Because you’re not buying the cashmere, you’re buying the brand, you’re buying the story, you’re buying the association, that’s why brands have value, because it’s irrational, and humans are irrational.

— Verbatim machine transcription, archived in this repository

The tap-water note is the version with verified numbers. The food writer’s account of the stunt (cfood.org):

He filled normal tap water into wine bottles, added fancy labels, and told people it was from Kazakhstan glaciers! Rich people believed it so much… they said it’s worth $500-$1000 a [bottle]. In the end, that ‘luxury water’ sold for $80,000!

The reel behind the note says the water came from “some secret spring in Japan or Antarctica.” The fetched article says Kazakhstan glaciers. This post follows the fetched article.

Starbucks is the version that lasted. The company opened in Seattle on March 30, 1971 (Wikipedia). The siren dates to the same year — Terry Heckler’s design, inspired by Seattle’s proximity to Puget Sound and the coffee industry’s seafaring roots (about.starbucks.com). In the transcript of the #64 reel, someone explains the layers in one sentence:

A logo gets you recognized. A visual identity gets you remembered. And a brand identity gets you chosen.

The counterweight is a #109 transcript (Steve Jobs, per the reel caption — the original interview is not named):

The answer is because customers don’t form their opinions on quality from marketing. … They form their opinions on quality from their own experience with the products or the services.

— Verbatim machine transcription, archived in this repository

And the author’s own counterpoint, in #58: “In business I do think people are needed to get things done.” Not every sale is a story. Some are a product that is good enough that the story is optional.

The section holds together in one line: the story buys the first look, and the experience has to keep the promise.

2. Distribution is the product

Seventeen notes make the same claim in different clothes.

David Sacks named it in a 2014 Draper University talk (the reel is #43):

what I would call the law of distribution arbitrage, which is that successful distribution techniques are copied until they are no longer unusually effective

— Verbatim machine transcription, archived in this repository; speaker and talk confirmed by startuparchive.org

Reid Hoffman, in Blitzscaling, as quoted in a Startup Archive post (the reel is #69):

a good product with great distribution will almost always beat a great product with poor distribution

Alex Schultz, Facebook VP of Growth, at Y Combinator in 2014 (the lecture, annotated transcript, repo transcript):

Retention is the single most important thing for growth.

And the demand side of it, from a #70 reel (Elon Musk, per the caption):

if you just like something, it’s okay, you’re not going to care that much, but if you love something… if you love it, you’re going to talk and then that will generate word of mouth

— Verbatim machine transcription, archived in this repository

The cases that carry the claim:

The last notes in the cluster are about the tooling: Postiz, an open-source social scheduler (37,051 stars, AGPL-3.0, pushed 2026-10-12), and OpenReply, an open-source tool that replies to your comments with a DM (2,609 stars, MIT, pushed 2026-10-05). Star counts measured via gh api in October 2026. Distribution now ships as software you can host yourself.

The rule: the product gets you noticed once. Distribution gets you noticed next. Sacks’ law is the warning — every channel that works gets copied until it stops working, so the job is not finding one channel, it is keeping finding channels.

3. Price is perception, and the numbers prove it

Ten notes. The cleanest statement of the cluster is Kotler’s customer-perceived-value model: value is total customer benefit minus total customer cost (mbamanagementmodels.com).

The cases:

The cluster says price objections are almost never about the number. They are about the ratio — what they get against what it costs them in time, risk, effort, and belief. Change the ratio, and the number stops being the problem.

4. Sell the destination, not the plane

A Brian Tracy talk was saved to the backlog twice — once under #35, once under #40 — the same recording, two uploaders. So it is quoted once here and credited to both:

15% of buying behavior is determined by the past. 85% of buying behavior is determined by the anticipated future.

— Verbatim machine transcription, archived in this repository (the #35 reel; the #40 reel carries the same talk — its transcript — with the cleaner caption)

So, as we say, if you are selling a tour to the Canary Islands, spend 90% of your time talking about the destination and only 10% talking about the plane that you fly to the Canary Islands in.

— Verbatim machine transcription, archived in this repository (the #40 reel; its cut is the longer one, running to the travel-agent story). The 90/10 split is the speaker’s heuristic from the talk, not a measured statistic.

A comment on #98 (posted 2026-03-23) tells the airplane-engine story, then the fix:

I just met with a company yesterday. They clean airplane engines. They brought me their pitch deck and the first 10 pages of the pitch deck were all about their company values.

The customer is the hero of the story of your company and you are the guide. You never want to play the hero.

— Issue comment transcript, posted 2026-03-23 (the uploader is the StoryBrand account; the company is the speaker’s anonymized client anecdote)

The author’s own note, #18:

You are only relevant if you’re able to make your audience understand that your are [sic] helping them get somewhere. … Marketing is showing them that you are helping them win.

And the anti-sales note, from a comment on #116 (posted 2026-07-26):

I don’t want people to feel sold to. If you only feel sold to if it’s bad. … Good selling, feels good. … So you should sell well.

— Issue comment transcript, posted 2026-07-26

The move: the pitch is about the customer’s destination. The company is the guide. The plane gets ten percent of the time.

5. The mechanics that still work

The notes with numbers that survived the check:

  1. Speed. A 2011 HBR study (1.25 million leads, 42 US companies): a lead contacted within an hour was about 7× more likely to qualify, and the odds fell more than 60× after 24 hours (as summarized at stylograph.ai).
  2. The small yes. Freedman and Fraser (Stanford, 1966): 76% of people who first agreed to a small request then complied with the large one, versus 17% of controls (Wikipedia, foot-in-the-door technique).
  3. The magic numbers. Facebook: 7 friends in 10 days. Slack: 2,000 messages. Twitter: 30 follows in week 1 — not day one, as #107 has it (koji.so, jamesmorrison.substack.com).
  4. Consistency. Buffer’s study of 100,000-plus accounts over 26 weeks: the most consistent posters got 5× the engagement per post, and the curve plateaus around 21 weeks (Buffer).
  5. The follow-up. Belkins: 58.6% of cold-email replies come from follow-up steps 2–6. Most of the conversation happens after the first message (stylograph.ai).
  6. The offer. A lead magnet that solves one small thing: the average landing-page conversion is 23% when the offer is specific and targeted (Unbounce 2023, as cited by Rework).
  7. Test volume. For Meta ads, 1,000-plus impressions per variant is a directional baseline, not a verdict (freako.io, citing AdStellar). The AI draft in the backlog said 3,000–5,000 views. No source for that number was found.
  8. The 30-character limit. Apple’s App Store caps app names at 30 characters (Apple). The name as written in #4 — “CaughtUp Fun Social Scheduler” — measures 31. The no-space variant that was actually submitted fits at exactly 30. Measured with pwsh.
  9. The exit door. OpenAI’s own Sora 2 page now reads, “As of April 26, 2026, the Sora product is no longer available” (archived copy). The platform you build on can be discontinued.

The consistency note fits here too. A Grant Cardone comment transcript on #68 (posted 2025-11-19):

The ultimate way to build trust is through consistency with viewership.

— Issue comment transcript, posted 2025-11-19. The survey he cites for the “50% of the audience” figure is unnamed, and this pass could not find it.

6. What the notes don’t say

The same pass that verified the list above produced this one:

The honest version of this post is that list, sitting under the verified version above it.

7. The move

118 notes, one argument, in order:

  1. People buy the story, not the thing. (The lemonade stand, the cashmere, the tap water.)
  2. Distribution carries the story — and every channel that works decays as it gets copied. (Sacks, Hoffman, Schultz, PayPal, Dropbox, BMW.)
  3. Price is the ratio, not the number. (Kotler, the decoy, Domino’s.)
  4. Sell the destination. The customer is the hero. (Tracy, StoryBrand, #18.)
  5. Run the mechanics: speed, the small yes, the magic numbers, consistency, the follow-up, test volume. (HBR, Freedman & Fraser, Belkins, Buffer.)
  6. Write down what you could not verify.

Two more posts from this pass are already written on unmerged branches in this repository — a Snow White’s Wishing Apple post from #113 and a winner-take-all market framework post from #122. They will link in here once they land.

The notes keep saying the same thing, in different orders: people buy the story, and they buy it through distribution.

What is the story your distribution is actually telling?

Sources

The full per-issue source list, including every UNVERIFIED item, is in .research/issues-research.md.

Notes

Journal