PastBehavior
Fieldwork / 005Download PDF

PastBehavior Research

Trust as Distribution

What happened when companies tried to turn trusted audiences into durable growth

16 cases · 2010–2026 · Creator, media and consumer businesses

Ten representative cases are shown in full. All sixteen inform the pattern analysis. The complete case base and source list sit in the methodology.

Evidence labels: FACT (filing, press release, or on-record statement), REPORTED (credible third-party reporting), INFERENCE (conclusion drawn across cases, labeled as such).

The signal

On August 12, 2026, WTSL, the Silver Lake-backed firm run by Patrick Whitesell and Jason Lublin, took its first outside stake in Unwell at a reported $500 million pre-money valuation. Cooper’s statement accompanying the deal contained the claim this report exists to test.

$500M

Reported pre-money valuation for WTSL’s first outside stake in Unwell. Terms undisclosed; Unwell files nothing.

“Trust has become the ultimate distribution channel and seventy million women a month tune-into Unwell.”
Alex Cooper · August 12, 2026

FACT that the statement was made, per Axios, Variety, and The Hollywood Reporter. REPORTED for the valuation, since terms were undisclosed and Unwell files nothing. The audience figure is a company number with no independent verification.

The claim is worth taking seriously because Cooper has better standing to make it than almost anyone. She is the only person in this case set who has moved the same audience across three owners without losing it. The claim is also worth testing, because the history says the word “distribution” is doing work the evidence will not support.

Executive finding

Trusted audience access can become durable distribution. Most of the time it does not.

Across sixteen cases, trusted audiences were extremely effective at generating first demand. Into The Gloss produced a beauty brand that reached a $1.8 billion valuation with roughly 80% of customers arriving by peer referral. Rhode reached $212 million in net sales, direct-to-consumer only, with ten products. Feastables sold a million bars in 72 hours.

The durable businesses emerged when that demand was converted into something that could compound. Repeat behavior. Owned customer relationships. Recurring formats. Retail distribution. An institution that performs without the original personality.

Where the conversion did not happen, the asset behaved like a very good campaign. Prime Hydration reached roughly $1.2 billion of sales in its first full year and then fell 71% in UK turnover the following year. Penn Entertainment bought Barstool for around $551 million as customer acquisition for its sportsbook, reached blended US market share below 5%, and sold the business back to its founder for $1.00 while writing off up to $850 million.

Across the cases, durable transfer depended on four conditions.

01

Adjacency

The new offer sits close to why the audience trusts the source.

02

Control

The trusted party controls the delivered experience.

03

Continuity

Acquisition does not change how the audience consumes the property.

04

Institutionalization

Not yet demonstrated in full

Behavior eventually survives the person who first earned the trust.

The first three predict whether trust transfers at all. The fourth predicts whether what transfers is an asset or a contract.

No case in this set has yet demonstrated all four. The durable cases satisfy the first three and either began institutionally or have not been tested without the principal. Every failure violated at least one of the first three. That gap is worth sitting with, because the market is currently paying very large prices for an asset whose final transferability remains mostly untested.

01 / The distinction

Attention, audience, permission, distribution

Companies use “audience,” “trust,” and “distribution” as though they describe one thing at different sizes. The economics separate at a specific point.

  1. 01

    Attention

    Gets seen

  2. 02

    Audience

    Comes back

  3. 03

    Permission

    Moves a decision

  4. 04

    Distribution

    Does it repeatedly at improving economics

Attention is purchasable and has no memory. Audience is repeat attention with identity attached, and it is where a cost curve first bends, because the second impression costs less than the first. Permission is trust over a specific decision, which is narrower than a follower count implies and is the thing a company is actually buying when it hires a creator. Distribution is the capacity to move behavior repeatedly, toward more than one offering, at a cost per action that falls or holds.

Whether the addressing mechanism is owned matters inside all of this. An audience reachable only through a platform is leased, and the terms can change. That is a property of the audience stage rather than a stage of its own.

The second-thing test

An audience stops being media inventory and starts behaving like distribution at the point where the second and third things sell more easily than the first.

First thing
Second thing
Third thing

Cost per action falls or holds as each subsequent thing is introduced.

That is the diagnostic. It converts an unanswerable question about trust into an arithmetic one about launch economics.

Companies rarely disclose the launch-level economics needed to run it directly. Not one of the sixteen cases published cost per acquisition on its second, third, and fourth offering against its first. The cases in this report therefore rely on observable proxies: repeat purchase behavior, revenue from subsequent products, retention through a change of ownership or terms, and the ability to extend the relationship into new offerings at all.

Prime is the clearest case consistent with failing the test. The launch was among the most efficient consumer product launches on record. UK turnover then fell from £112.2 million to £32.8 million in a single year, with net profit down 91.6% to £312,393 (FACT Prime Hydration UK Ltd filed accounts). Nothing about the audience changed. What did not happen is a second offering that sold more easily than the first.

The test also explains why two companies with identical audience sizes are worth very different amounts. Grantland and The Ringer had comparable editorial quality and overlapping staff. One was shut down five months after its founder left. The other has survived six years, a change of parent strategy, and repeated speculation about the same founder.

02 / What history shows

The four conditions, the evidence, and the test

The four conditions, with the evidence behind each and the test a company can run before committing money.

Condition 01

Adjacency

The new offer sits close to why the audience trusts the source.

Transfer worked where the new offer sat inside the domain the audience already trusted the source about. Into The Gloss readers trusted Weiss about what to put on their face, and Glossier sold things to put on your face. Rhode’s audience trusted Bieber about a specific aesthetic, and Rhode sold that aesthetic in ten products.

Transfer failed where the offer sat outside it. Barstool’s audience trusted Barstool about sports culture and did not accept Barstool’s authority over which regulated sportsbook to fund.

Adjacency is defined by the reason for the trust rather than by category. Feastables is adjacent to MrBeast’s content because the golden-ticket mechanic makes the product part of the show. MrBeast Burger was not adjacent, despite both being food.

Test / Adjacency

Name the specific claim the audience accepts from this source without verifying it. If your offer requires a different claim, you are outside the boundary.

Condition 02

Control

The trusted party controls the delivered experience.

Where the trusted party does not control the delivered experience, trust is spent on the first purchase and then destroyed at a rate proportional to the quality gap.

MrBeast Burger was licensed to a ghost-kitchen operator across roughly 1,700 third-party kitchens and produced litigation in which the creator’s own filing described the food as “low quality” and at times “inedible” (FACT complaint filed SDNY, July 31, 2023; the allegations are contested by Virtual Dining Concepts, which countersued for $100 million).

Test / Control

Who is responsible if the product disappoints, and who will the audience blame. If those are different parties, do not proceed.

Condition 03

Continuity

Acquisition does not change how the audience consumes the property.

Every acquisition in this set that preserved the audience preserved the terms of consumption. Every one that changed them lost audience. (INFERENCE supported by five cases.)

Spotify made Gimlet and Parcast shows platform-exclusive. The studios’ unions stated the change caused listener declines as high as three quarters for some titles (REPORTED joint union statement, October 2022, from an interested party during a labor dispute; Spotify has not published contradicting figures). Spotify reversed exclusivity across its major properties by early 2024.

The properties that grew after acquisition were all acquired with an explicit non-interference commitment. HubSpot’s formulation is the most direct: “adding resources without adding interference” (FACT HubSpot blog, February 24, 2026). Plaid structured This Week in Fintech as an independent subsidiary retaining editorial control. Insider left Morning Brew’s brand, office, org structure, product, and headcount intact.

Test / Continuity

List what you would change in the first year. If anything on that list is visible to the audience, price the deal lower.

Condition 04

Institutionalization

Behavior eventually survives the person who first earned the trust.

The mechanisms that appear in properties that outlived, or credibly could outlive, the founder:

Recurring named formats carried by more than one voice. The Ringer runs 30-plus shows with franchises identified by format rather than by host. Morning Brew built five newsletters under a house voice. Grantland had a masthead and one reason to visit.

An owned addressing mechanism. Morning Brew, The Hustle, This Week in Fintech, and Starter Story are all email-substantial, which lets the company reach the audience without asking a platform. This is necessary and not sufficient, because an email list produces attention.

A commercial model that does not require constant selling into the relationship. Where a property is under pressure to attribute revenue immediately, the pressure shows up in the product.

Test / Institutionalization

Name one format or product in the property that performs when the principal is absent. If you cannot, the property is personal and you are buying a contract.

The honest limit on this condition

No case in this set has completed the transition. The Ringer’s founder is still there six years later. Morning Brew is the closest working example and it was built with a house voice from the start, which means it never had to convert. Institutionalization is the condition with the strongest logic and the thinnest proof.

03 / The cases

Ten representative cases

Each states what the asset was, what was attempted, what happened, and the mechanism that transfers. Two of the ten are told as paired comparisons in Section 04.

Call Her Daddy at Barstool

2018–2021

Condition violated

Institutionalization

Barstool acquired the show a month after launch and owned the IP. In the 2020 contract standoff, Portnoy offered the hosts the IP in exchange for one additional year. Cooper took the deal, stayed twelve months, and left for Spotify in 2021 for a reported $60 million over three years. Portnoy claimed Barstool was losing $100,000 per missed episode during the standoff (REPORTED New York Times, May 2020).

Cooper’s own account of why she fought for it: “I’m so happy that I trusted myself and fought for this IP” (FACT The Hollywood Reporter, October 2024).

Mechanism

Barstool held the brand, the archive, and the contract, and could not hold the audience, because the audience’s relationship was with a person who was legally free in twelve months. Owning the IP of a personality-led property is not owning the asset.

Penn Entertainment and Barstool

2020–2023

Condition violated

Adjacency

Penn paid roughly $551 million in total for full ownership, explicitly as customer acquisition. Its February 2023 release described Barstool as “a proven, powerful media brand with an authentic voice and vast, loyal audience that provides us with a strong top of funnel for new customer acquisition” (FACT).

Blended US sports betting share ended below 5%. Barstool posted a $16 million loss in the first half of 2023. Several states hesitated on licensing because of the brand’s content history, which Portnoy acknowledged directly: “We got denied gambling licenses because of me.”

In August 2023 Penn sold 100% of Barstool back to Portnoy for $1.00 plus non-compete covenants and 50% of any future sale, recording a pre-tax non-cash loss of $800 million to $850 million (FACT Penn 10-Q, August 9, 2023). Penn’s CEO on the analyst call: “There’s probably only one long-term owner of Barstool Sports, and that’s Dave Portnoy and Barstool Sports.”

Mechanism

The most expensive available demonstration that reach and permission are different assets. The audience paid attention to Barstool and did not accept Barstool’s standing over a regulated financial decision. Separately, a brand whose value derives from a person who cannot pass a suitability review is not an asset a regulated operator can hold.

Gimlet and Parcast at Spotify

2019–2023

Condition violated

Continuity

Acquired for roughly $189 million and $54 million plus retention incentives. Spotify made most shows platform-exclusive on the thesis that exclusivity drives differentiation. The unions put resulting listener declines as high as 75% for some shows. Spotify cancelled 11 shows in October 2022, merged the studios in June 2023, and reversed exclusivity. Ek described podcasting on the January 2023 call as “a drag to our gross margin profile” (FACT). Roughly $1 billion of cumulative podcast investment against a business of approximately €200 million in 2022 revenue.

Mechanism

Audiences attach to a habit rather than to an owner. Moving a trusted property behind a wall does not move the audience with it, and marketing does not recover the loss.

Morning Brew at Insider and Axel Springer

2020–present

Conditions satisfied

Adjacency, control, continuity. Institutional by construction.

Acquired October 2020 at a valuation around $75 million against roughly $20 million of expected revenue and $6 million of expected profit. The founders retained a minority stake with an earn-out. The brand, office, org structure, products, and all sixty employees stayed (FACT Axios, October 29, 2020). Professional brands alone generated $25 million of revenue in 2024, with the business on pace to exceed $70 million in total annual revenue in 2025 (REPORTED secondary, not filed).

Mechanism

The strongest available case of an originator-founded property becoming institutional. Multiple newsletters, a house voice rather than a personal byline, verticals with their own audiences. Neither founder was the reason to open the email, which is why both could step back without a visible break.

Glossier and Into The Gloss

2010–present

Condition eventually violated

The ceiling on trust-only distribution

A beauty blog reaching roughly 10 million monthly page views before any product existed. Glossier launched in 2014 into that audience, reached a $1.2 billion valuation in 2019 and $1.8 billion in 2021, with roughly 80% of customers arriving by peer referral (REPORTED Recode/Vox interview with Weiss, 2018).

In January 2022 the company laid off more than 80 people, about a third of corporate staff. Weiss to staff: “Over the past two years, we prioritized certain strategic projects that distracted us from the laser-focus we needed to have on our core business. We also got ahead of ourselves on hiring. These missteps are on me” (FACT internal email obtained by Retail Dive). She stepped down as CEO in May 2022. In February 2023, after eight years of refusing wholesale, Glossier launched in Sephora, reportedly contributing around $100 million of retail sales in year one.

Mechanism

Trust built the company and could not scale it. The recovery came from adopting the conventional retail distribution the original thesis was constructed against.

Rhode and e.l.f. Beauty

2022–present

Conditions satisfied

Adjacency, control, continuity. Institutionalization untested, with the creator retained.

Rhode reached $212 million in net sales in the twelve months to March 31, 2025, direct-to-consumer only, ten products (FACT e.l.f. and CNBC, May 28, 2025). e.l.f. acquired it for $600 million cash plus $200 million of stock, with up to $200 million of earn-out over three years. Bieber became chief creative officer and head of innovation.

Post-close, Rhode contributed $128 million to Q3 FY2026 net sales, approximately 36 of e.l.f.’s 38 percentage points of growth, while organic net sales grew approximately 2% (FACT earnings call, February 4, 2026). Full-year Rhode net sales reached $390 million, up 80%, and it became the number one beauty brand at Sephora North America (FACT Q4 FY2026 release, May 20, 2026). e.l.f. recorded a $57.6 million increase in contingent consideration for FY2026 because Rhode beat its earn-out thresholds, and $16.1 million more in Q1 FY2027 (FACT 8-K filings).

Mechanism

Note which side supplied what. Bieber supplied demand. e.l.f. supplied Sephora, Mecca, supply chain, and seven global facilities. The creator stayed, contractually and operationally, and the acquirer supplied distribution the brand did not have.

Prime Hydration

2022–2025

Condition violated

The second-thing test

Launched January 2022 with scarcity mechanics and two very large creator audiences. Roughly $1.2 billion of sales in 2023 (REPORTED Bloomberg). UK turnover fell from £112.2 million to £32.8 million in 2024 with net profit down to £312,393 (FACT filed accounts). US sales down roughly 40% in H1 2024 per Numerator and 42% year over year as of July 2025 per Circana (REPORTED). Supplier litigation followed over contracted volume shortfalls.

Mechanism

The creator audience produced a demand event rather than a demand curve. When scarcity ended and distribution widened, the thing that had been selling stopped selling, because it had been selling participation.

Unwell Network and Alix Earle

2023–2025

Condition tested

Whether trust transfers creator to creator

Earle was announced in August 2023 as the network’s flagship signing, with Cooper saying she felt “honored to be at a place in my career where I can pass along knowledge and advice for a new generation of creators” (FACT Variety, August 2023).

In February 2025 the network dropped Hot Mess, SiriusXM stopped selling its ads, and Unwell renounced all rights to the show (REPORTED Variety exclusive, February 25, 2025). Earle later described the period as involving lawyers and confusion. Cooper stated publicly that “Unwell gave her everything back.”

Mechanism

The network model requires trust to flow from the founder to signed talent. In the first significant test, what flowed the other way were the rights, back to the talent. This is the same shape as the 2020 Barstool case with the roles reversed, five years later. Single data point, and the weakest evidence base in the report.

Grantland and The Ringer

Told as a paired comparison in Section 04. Go to Section 04

MrBeast Burger and Feastables

Told as a paired comparison in Section 04. Go to Section 04

04 / The paired comparisons

Same audience. Different operating model.

Four comparisons where the audience, founder, or brand stays substantially constant while the operating model changes. They do not isolate a single causal variable. Each one changes several things at once, including time, capital, personnel, and market conditions. They are useful because holding the audience constant makes the mechanism easier to see than a standalone case does.

Grantland ↔ The Ringer

Same founder. Same category. Opposite outcome.

Grantland (2011–2015)

Structure
One masthead, one reason to visit
Owner behavior
ESPN declined to renew Simmons, said publicly it remained committed to the site
Outcome
Suspended October 30, 2015, five months after Simmons left

The Ringer (2016–present)

Structure
30-plus named shows, multiple hosts
Owner behavior
Spotify kept shows non-exclusive, kept The Ringer structurally separate in the 2023 restructuring
Outcome
Six years, through a full reversal of the acquirer’s podcast strategy

Spotify’s consideration for The Ringer was partly deferred and contingent on performance and on Simmons and other executives remaining (FACT SEC filing, February 2020). Ek at the time: “What we really did with The Ringer, I think, is we bought the next ESPN.”

The important change

Recurring formats carried by voices other than the founder. Grantland was institutionally structured and personally trusted. That combination does not survive the person.

MrBeast Burger ↔ Feastables

Same creator. Same audience. Same year range. Opposite outcome.

MrBeast Burger (2020–2023)

Model
Licensed to a ghost-kitchen operator, roughly 1,700 third-party kitchens
Distribution
Delivery apps
Outcome
Litigation, creator’s own filing calling the food inedible

Feastables (2022–present)

Model
Owned product, manufactured to specification
Distribution
Walmart at launch, later 30,000-plus retail locations
Outcome
Roughly $250 million of 2024 sales and about $20 million of profit, exceeding the media business for the first time (REPORTED Bloomberg-obtained investor documents, March 2025, unaudited)

The important change

Control of the delivered experience, plus conventional retail distribution. This pair makes the Control condition more visible than any other comparison in the case set.

Glossier DTC ↔ Glossier + Sephora

Same brand. Same community. Distribution added.

Glossier DTC · 2014–2022

Eight years of direct-to-consumer with roughly 80% referral-driven acquisition produced a $1.8 billion valuation and then a layoff of a third of corporate staff.

Glossier + Sephora · 2023–

Adding Sephora in 2023 reportedly produced around $100 million of retail sales in the first year.

The important change

The addition of shelf space to an audience the company already owned. The community was real and had a ceiling.

Rhode before e.l.f. ↔ Rhode after e.l.f.

Same brand, before and after an operator was added.

Before · twelve months to March 31, 2025

$212M

DTC-only net sales in the year before acquisition.

After · first full fiscal year

$390M

Net sales in the first full fiscal year after, up 80% — number one at Sephora North America, while the acquirer’s own core business grew approximately 2%.

The important change

Retail distribution and supply chain, supplied by the buyer. This is the pair that most directly contradicts the framing in the opening signal.

What the four comparisons establish together

Read as a set, the comparisons point at one conclusion.

Across the durable cases, trust was paired with something conventional that could compound it: shelf space, recurring formats, owned customer relationships, or supply-chain control.

(INFERENCE four comparisons plus the wider case set.)

That is the substantive answer to the claim in the opening signal.

Trust creates demand. Distribution begins when that demand compounds.

05 / What sets the price

Priced by the highest-value use of the relationship

Trusted attention used to be priced by advertising economics, meaning some multiple of what advertisers would pay to reach the audience, bounded by the advertiser’s margin and by auction competition. There is now good evidence it is being priced by the highest-value use of the relationship.

HubSpot / Starter Story

Feb 2026
Audience under old model
Creator media, sponsorships
What the buyer can do with it
Reduce paid CAC into a CRM funnel among pre-seed to Series A founders
Ad revenue post-deal
Retained, secondary

Plaid / This Week in Fintech

Mar 2026
Audience under old model
Newsletter and event sponsorship
What the buyer can do with it
Cross-sell an expanded product suite to the decision-makers who read it
Ad revenue post-deal
Retained, independent

OpenAI / TBPN

Apr 2026
Audience under old model
Roughly $30M run-rate advertising
What the buyer can do with it
Direct channel to founders, developers, and investors ahead of a reported listing
Ad revenue post-deal
Shut down

WTSL / Unwell

Aug 2026
Audience under old model
Podcast advertising, products, production
What the buyer can do with it
Platform for roll-up acquisitions in a demographic the buyer cannot otherwise reach
Ad revenue post-deal
Retained

The OpenAI transaction is the clearest evidence of the shift. TBPN ran roughly $5 million of advertising revenue in 2025 and was reportedly on track to exceed $30 million in 2026 (REPORTED Wall Street Journal via CNBC, April 2, 2026). OpenAI is winding that business down as part of the deal and placed the show inside its Strategy organization reporting to its chief global affairs officer, rather than in a media division (FACT on the reporting line, per OpenAI; REPORTED TheWrap, April 6, 2026, on the advertising wind-down).

A buyer that deletes the entire revenue line is not paying an advertising multiple.

Three conditions appear to be driving this, and this is INFERENCE across the case set rather than a documented claim by any participant. Paid acquisition costs have risen enough that owned media compares favorably on a multi-year view. Platform distribution has become less dependable, which raises the value of an addressable relationship. And buyers have appeared whose value per relationship is extraordinarily high, so an AI lab facing a public listing can rationally value narrative among 58,000 of the right subscribers at a number no advertiser would pay.

The counter-evidence

The counter-evidence is substantial and belongs here rather than in a footnote. Acquisition price is not proof of strategic value. This case set contains a $551 million purchase disposed of for $1.00 within three years, with a write-off exceeding the purchase price. Spotify spent roughly $1 billion on podcast acquisitions and reversed its core strategic assumption within four years. In both cases the buyer had a coherent theory about the highest-value use of the relationship, and the theory was wrong.

For an established company on the buy side, the consequence is uncomfortable. You are no longer bidding against advertisers. You are bidding against parties whose economics per relationship may be an order of magnitude above yours.

06 / Why Gen Z matters

A cohort effect, not a new law of physics

Gen Z does change the equation, and not for the reason most companies assume.

Historical trust data suggests today’s gap combines a cohort effect with a normal age effect. Younger adults have consistently become more trusting as they age, while successive generations have also entered adulthood from lower baselines. (Robinson and Jackson, 2001; Schwadel and Stout, 2012, on GSS data 1972 to 2010; Twenge, Campbell and Carter, 2014. Detail in the source notes.)

Two consequences follow.

Companies should expect trusted intermediaries to remain unusually valuable, without assuming today’s exact Gen Z behavior is permanent. A strategy priced on this cohort being permanently unreachable by institutions will over-invest, because part of the gap will close on its own as the cohort ages.

More important, low institutional trust does not automatically mean high creator trust. It means high trust in sources with legible incentives. Reviews, peers, and employers hold the same permission. The 2026 Edelman Trust Barometer found the most trusted institution globally is “my employer” at 78% among employees (FACT January 18, 2026). A February 2026 survey of 2,000 Gen Z Americans found customer reviews the most trusted source of brand credibility at 72%, with influencer content at 55%, below several independent sources (REPORTED Walr for We Are Talker, published methodology, commissioned study).

72%

Customer reviews — the most trusted source of brand credibility

55%

Influencer content — below several independent sources

Survey of 2,000 Gen Z Americans, February 2026. One commissioned study (Walr for We Are Talker, published methodology) — a single data point, not a universal market statistic.

The commercial evidence for Gen Z creator influence is also weaker than its reputation. Published 2026 compilations variously report that 32%, 47%, 56%, and 87% of Gen Z buy on creator recommendation. A figure that ranges from 32% to 87% across vendor-commissioned studies is not a measurement, and it should not be presented to a board as one.

07 / The failure modes

Five documented ways this fails

Five, each attached to a documented case.

  1. 01

    Confusing reach with permission

    Penn Entertainment / Barstool

    Penn acquired Barstool as “a strong top of funnel for new customer acquisition” and reached blended US sports betting share below 5% before writing off up to $850 million. The audience paid attention. It did not accept direction over a regulated financial decision.

  2. 02

    Assuming the trust belongs to the parent

    ESPN / Grantland

    ESPN stated publicly that it remained committed to Grantland after declining to renew Simmons, and closed the site five months later.

  3. 03

    Extending past what the trusted party controls

    MrBeast Burger

    MrBeast Burger, fulfilled by roughly 1,700 third-party kitchens, produced litigation in which the creator’s own filing described the food as inedible.

  4. 04

    Changing the terms of consumption after acquisition

    Spotify / Gimlet and Parcast

    Spotify’s exclusivity strategy on Gimlet and Parcast was followed by listener declines the unions put as high as three quarters for some titles, and Spotify reversed exclusivity across its major properties within eighteen months.

  5. 05

    Selling scarcity rather than a product

    Prime Hydration

    Prime’s UK turnover fell 71% and profit 92% in a single year once distribution widened and scarcity ended.

A sixth mode — thin evidence

One further mode is worth naming despite thin evidence. Companies that rent creator access for years and hold no addressable relationship at the end produce no public artifacts and no postmortems, so the record is silent. Treat it as plausible and undocumented.

08 / Rent, borrow, build, buy, institutionalize

The five routes to permission

A

Rent access

When it works
Infrequent decisions, low strategic importance, situations where a first purchase is the whole objective.
Cost
Ongoing and rising. No accumulation.
Horizon
Immediate.
Primary risk
Mistaking a good campaign for a durable asset. Prime is that mistake at scale.
Evidence required first
A repeat-purchase read from a prior comparable campaign. Without one, you are buying trial.
Economic consequence
Cost per acquisition stays flat and drifts upward as competition for the same creators increases.
B

Borrow permission

When it works
A trusted third party already holds standing over the exact decision you want to influence, and the relationship can run long enough to compound. Evidence favors sustained single-voice relationships over campaign bursts.
Cost
Higher per unit than rental, well below building.
Horizon
Twelve to thirty-six months before compounding is observable.
Primary risk
The permission is narrower than the audience.
Evidence required first
Documented evidence that this audience previously acted on this source’s recommendation in an adjacent category.
Economic consequence
Better conversion on borrowed occasions. Nothing that survives the end of the relationship.
C

Build a trusted property

When it works
Repeated decisions in a category where you hold genuine domain knowledge, with tolerance for a multi-year payback. HubSpot’s media portfolio is the working example, though the company has never published the CAC comparison that would prove the economics.
Cost
Low in cash, high in patience and editorial discipline.
Horizon
Three to five years before the property generates its own demand.
Primary risk
Internal pressure to attribute revenue immediately, which turns the property into a funnel and removes the reason anyone consumed it.
Evidence required first
Written board agreement on the measurement period, before the first hire.
Economic consequence
Permanently lower dependence on paid media, held as an asset that appears nowhere on the balance sheet.
D

Buy existing permission

The label matters. What is for sale is permission over a set of decisions. It earns the word distribution only after it passes the second-thing test under your ownership.

When it works
The permission is scarce, adjacent to your commercial decision, and worth more under your economics than under its current model. That last condition is what makes an acquisition rational rather than expensive.
Cost
A multiple set by whoever has the highest-value use, who may not be you.
Horizon
Two to four quarters if the asset was already commercial.
Primary risk
Changing the terms of consumption after closing.
Evidence required first
Confirmation of who the audience actually follows, established by observing what happened when that person was previously absent.
Economic consequence
Rhode delivered 80% growth post-close while the acquirer’s core brand grew 2%. Penn wrote off up to $850 million. Both are available in this route.
E

Institutionalize

When it works
You can add recurring formats, additional voices, and an owned addressing mechanism without changing what the audience came for.
Cost
Moderate cash, high judgment.
Horizon
Five years and up, and possibly never completed.
Primary risk
The conversion is visible to the audience. Introducing a second voice to a personal property is the moment trust gets tested.
Evidence required first
At least one existing format that already performs without the principal present. If none exists, the correct next step is to create one and measure it.
Economic consequence
The only route that converts a contract into an asset.

09 / Decision map

Start with the decision, not the audience

Start with the decision. Every failure in this set began with a company that had identified an audience it wanted and had not specified the behavior it wanted from them.

  1. 01

    What specific decision are you trying to influence?

    Name it as a behavior with a timeframe. “Reach Gen Z” is not a decision. “Get a 22-year-old to open a brokerage account within 30 days” is.

  2. 02

    Who already holds permission over that exact decision?

    Not who has the audience. Who has standing on that question. These are frequently different parties, and the gap between them is where the money goes.

  3. 03

    How adjacent is your offer to the reason they are trusted?

    Skincare recommendations to skincare products is adjacent. Sports commentary to a regulated betting account is not, and establishing that cost Penn up to $850 million.

  4. 04

    Who controls the delivered experience?

    If a third party controls fulfillment and the audience will blame the trusted party, the trust will be consumed at the rate of the quality gap.

  5. 05

    Is the relationship personal or institutional?

    Name a format or product in that property that performs when the principal is absent. If you cannot, you are buying a contract, and it should be priced and structured as one.

  6. 06

    How often do you need this access, and will the second thing be cheaper to sell than the first?

    Once or twice, rent. Continuously in an adjacent category, build. Frequently in a category you cannot credibly enter, borrow with a single voice over a long term. If you only have one product to sell, you do not need distribution. You need a campaign.

  7. 07

    What must stay independent for the trust to survive, and can your organization leave it alone?

    Every preserved audience in this set was preserved by an owner that committed publicly to non-interference and honored it.

10 / The rule

Trust becomes distribution when the next thing gets easier to sell.

Until that happens, the company has demand. The question is who owns the relationship producing it.

Adjacency, control, continuity, and institutionalization are the conditions that predict whether the next thing will get easier to sell. They are not part of the definition. A person can personally make several products in sequence and still hold genuine distribution, provided the launch economics improve each time. The test is the economics.

Applied to the opening signal

Unwell’s $500 million valuation rests on the claim that trust is distribution. On this evidence, Unwell has demonstrated demand at a very high level and has one documented test of transfer to a third party, which went the way this history predicts. The company is now buying. The useful thing to watch is not the next acquisition. It is whether anything Unwell owns continues to perform when Alex Cooper is not in it.

What this means for established companies

The research answers the question about the asset. The question underneath it is what an established company should do about younger consumers, and the history points at a change in where that conversation starts.

Established companies should stop starting with “How do we reach Gen Z?”

Reach was the constraint when distribution was scarce and attention was concentrated. Neither is true now. Reach is purchasable at scale by anyone with a budget, and the companies in this case set that failed did not fail on reach. Penn had extraordinary reach into young men. Prime had extraordinary reach into teenagers.

The better question is:

Which decisions do we need younger customers to make, and who already has permission over those decisions?

That question produces different work. It requires naming the decision, identifying who currently holds standing over it, testing whether that standing is personal or institutional, and then selecting a route based on how often you need the access and whether your economics justify owning the relationship. The answer is sometimes that nobody holds the permission you need and the right move is to build it, slowly, in a category adjacent to what you already do.

It also produces a different failure profile. A company that starts with reach can spend a great deal of money and never learn whether it owns anything. A company that starts with the decision will know within two launches.

Methodology and source notes

Case selection
Sixteen cases spanning 2010 to 2026, selected to include clear successes, clear failures, partial successes, cases where the audience remained valuable while the strategic owner failed, cases where the creator was the asset, and cases where the franchise became the asset. Selection was on availability of dated public evidence about what happened after the transaction, not on transaction size.
Evidence labels
FACT — filing, press release, or on-record statement. REPORTED — credible third-party reporting. INFERENCE — conclusion drawn across cases, labeled as such.
Source hierarchy
SEC filings and company earnings releases first, including e.l.f. Beauty’s FY2026 releases and 8-Ks, Penn Entertainment’s August 2023 10-Q, Spotify’s 2020 6-K and 2019 to 2020 filings, and UK filed accounts for Prime Hydration UK Ltd. Then company announcements and on-record executive statements. Then contemporaneous reporting by named outlets. Academic sources are used only for the age and cohort analysis.
The full sixteen cases +

Call Her Daddy at Barstool; Barstool and Penn Entertainment; Grantland and ESPN; The Ringer and Spotify; Gimlet and Parcast at Spotify; Morning Brew at Insider and Axel Springer; The Hustle and Starter Story at HubSpot; This Week in Fintech and Plaid; TBPN and OpenAI; Into The Gloss and Glossier; Rhode and e.l.f. Beauty; Prime Hydration; MrBeast Burger and Virtual Dining Concepts; Feastables and Beast Industries; the Unwell Network and Alix Earle; Call Her Daddy across Spotify and SiriusXM.

On the age and cohort research +

Age-period-cohort analysis of US survey data finds a consistent life-cycle pattern in which young adults begin adulthood with low interpersonal trust, become more trusting into middle age, and hold at the higher level. The pattern appears in raw age means, in age means controlling for survey year and birth cohort, and in full regression estimates (Robinson and Jackson, Social Science Research, 2001, and the 2013 replication). Separately, Schwadel and Stout, analyzing GSS data from 1972 to 2010, found trust declined across both periods and cohorts, with each cohort born since the 1950s less trusting than the previous one, excepting a slight increase around the 1980s cohorts. Twenge, Campbell and Carter (Psychological Science, 2014) reached compatible conclusions using GSS data alongside Monitoring the Future data on high school seniors. The Marquette Law School Poll published July 7, 2026 found a steady decline by birth cohort in trust in other people across nearly every category of person, alongside only modest cohort differences in trust in the federal government, which suggests the phenomenon is generalized rather than political.

Known weaknesses

  • Unwell, Beast Industries, Feastables, and Barstool are private and file nothing in the US. All financial figures for them are secondary. The Feastables numbers derive from leaked investor documents obtained by Bloomberg and The Verge in March 2025 and are directional only.
  • The Spotify exclusivity elasticity, the single most useful measurement in the report, comes from a joint union statement made during a labor dispute. Spotify has not published contradicting figures and has not confirmed it.
  • The Gen Z consumer-behavior literature is close to unusable. Most published figures come from vendors selling influencer services.
  • Deal prices for three of the four 2026 transactions are undisclosed. Where a number appears it is labeled as reported.
  • Survivorship runs against the conclusions in one direction specifically. Failed owned-media programs inside large companies produce no public record, which means Route C looks better in this evidence base than it may be in practice.
  • The most important limitation is recency. Three of the sixteen cases are less than six months old. The OpenAI, Plaid, and WTSL transactions have outcomes that cannot be known yet, and they carry most of the weight in Section 05. If any of the three reverses, that section weakens considerably. This report should be re-run in twelve months against those three outcomes.

Appendix

Hypotheses tested

Ten hypotheses, their status, and the conclusion the evidence supports.

Large audiences are not necessarily distribution assets.

Supported

Barstool had one of the largest young male audiences in America and could not convert it into sportsbook share above 5%.

Trust becomes distribution when it can repeatedly move behavior.

Supported

With the refinement that the movement must be across more than one offering at non-rising cost.

The highest-value owner increasingly sets the price.

Supported

This is a claim about price formation and not about whether the buyers are right.

Personal trust is easier to acquire than to institutionalize.

Strongly supported

Nine acquisitions in this set and zero completed founder transitions with the property intact.

Trust transfers best across adjacent decisions.

Supported

With adjacency defined by the reason for the trust rather than by category.

Immediate performance optimization can reduce the long-term value of an audience property.

Supported, weak evidence

The Spotify exclusivity case is monetization-driven rather than attribution-driven, and no clean attribution-pressure case exists in the public record.

A direct audience relationship matters more as platform distribution becomes less dependable.

Supported

By revealed preference rather than by measurement. Four buyers in eighteen months, and none published CAC data.

Gen Z’s lower incumbent-brand trust increases the value of intermediaries who already hold permission.

Partially supported

Generally over-claimed. Part of the gap is a life-cycle effect that will attenuate, and Gen Z rates customer reviews above influencer content on brand credibility.

Owning a younger audience and knowing how to operate it are separate capabilities.

Strongly supported

Penn owned Barstool outright for six months and concluded there was only one viable operator. Rhode’s post-close growth came from an operator supplying capability the founder did not have.

The durable asset is the franchise only when behavior survives the original personality.

Largely untested

Supported and largely untested in the affirmative. Grantland is the negative proof. The Ringer and Morning Brew are consistent with the hypothesis and neither has fully run the experiment.

What does past behavior say about your next move?

The useful part of precedent isn’t knowing what happened. It’s identifying which conditions made it work, whether those conditions exist in your business, and what changes if you choose the same route.

If you’re deciding whether to rent, borrow, build, or buy access to an audience, we can map the historical cases against your situation.

Talk through a decision