PastBehavior
Fieldwork / 003Download PDF

PastBehavior Research

The Fixed Capacity Playbook

Historical evidence on how businesses use partnerships, contracted demand, and distribution to make existing capacity more valuable.

Businesses with fixed capacity often treat underutilization as an acquisition problem. The historical record suggests something more specific. More customers only matter when they improve contribution against assets and costs that already exist.

We studied businesses across mobility, fitness, rental, real estate, government, and other capacity-constrained categories to understand when alternative distribution actually changed the economics — and when enormous customer bases still weren’t enough.

24 cases · 8 channel types · Historical evidence through 2026

Executive finding

Distribution is not the strategy. Improving the economics of existing capacity is the strategy.

Across the cases, partnerships worked when they did at least one of three things:

  1. 01Filled capacity during periods it would otherwise sit unused.
  2. 02Removed a fixed cost the operator would otherwise carry.
  3. 03Delivered contracted, recurring demand at a predictable location.

They failed when they merely produced more customers, discounted already-valuable inventory, or introduced usage whose marginal cost exceeded its contribution.

The distinction sounds small. Historically, it has separated durable distribution models from businesses that accumulated enormous audiences and still disappeared.

01 / The pattern

The same mistake appears repeatedly

The familiar sequence

  1. Unused capacity
  2. Acquire more customers
  3. More usage
  4. But not necessarily better economics

The sequence that held up

  1. Fixed capacity
  2. Find the trough
  3. Aggregate demand that fits it
  4. Reduce or share fixed cost
  5. Measure contribution per asset

The relevant question is not: “Can this channel acquire customers?”

It is: does this channel improve contribution from capacity we’re already paying for?

02 / What history shows

Four patterns kept appearing

01

Contracted demand beats audience size

The strongest models did not simply reach large groups of consumers. They contracted with institutions that could repeatedly assign demand.

Enterprise built its rental model around insurers, body shops, businesses, and other contracted demand. In 2023 it said contracted business represented the majority of its rental days. Avis has similarly reported that roughly half of its rental transactions originate through corporate contracts or affiliated organizations.

Lesson
Owning access to the organization assigning demand can matter more than owning access to the consumer.

02

The best demand often arrives during the trough

Wellhub routes employer-sponsored users into fitness facilities whose fixed costs exist whether the facility is full or empty. Its network grew to nearly 40,000 corporate clients and more than 5 million employee subscribers, while check-ins increased from 100 million in 2022 to 1 billion in 2025.

The important mechanism isn’t simply B2B2C distribution. It is matching aggregated demand with capacity that would otherwise go unused.

Caveat: operator outcome statistics associated with Wellhub come from a company-commissioned survey of its own partner facilities.

Lesson
A customer at the wrong time may be worth dramatically less than a customer who fills the trough.

03

Sometimes the partner should supply the expensive input

Residential shared-vehicle programs reveal a different mechanism. Large property owners have placed shared vehicles directly inside apartment portfolios. The property supplies concentrated customers and, critically, the parking location. The partnership therefore attacks acquisition and fixed cost simultaneously.

Similar structures appear elsewhere whenever a distributor supplies real estate, infrastructure, inventory, capital, or another input the operator would otherwise purchase.

Lesson
The best distribution partner may reduce cost before it creates a single customer.

04

Huge audiences can still lose

Several mobility businesses accumulated enormous customer bases and still disappeared. Share Now reported more than 800,000 North American customers before exiting the continent. GIG Car Share had the distribution potential of AAA behind it and still shut down. Getaround reported millions of guests and tens of thousands of active vehicles before closing its US operations.

These businesses had demand. Their underlying asset economics still failed.

Lesson
Aggregation cannot rescue structurally bad contribution per asset.

03 / The evidence

The cases

Ten of the twenty-four reviewed cases, selected for evidentiary strength. Expand any row for the underlying figures, caveats, and citations.

Enterprise

FACT
Constraint
No airport distribution or incumbent brand advantage
Distribution move
Insurer, body shop, and corporate referral relationships
What happened
Contracted business became the majority of rental days
What it teaches
Acquire the entity assigning demand, not the consumer
View evidence +

Enterprise Holdings reported $35bn in FY2023 revenue and stated that contracted business, which it defines as including B2B accounts and insurance replacement, accounted for the majority of its rental days that year. Replacement demand is uncorrelated with travel demand, which smoothed utilization through the post-2001 travel collapse. The channel was defensible enough that a competitor bought into it: Avis Europe acquired credit hire firm 3 Arrows for £40m in 1998 to enter the replacement segment.

  • Auto Rental News, October 13, 2023
  • Autobody News, January 13, 2024
  • Fleet News, December 18, 1998

Avis

FACT
Constraint
Large fixed fleet carried against uncertain retail demand
Distribution move
Corporate contracts and affiliation partners
What happened
Roughly half of transactions originate through contracted or affiliated channels
What it teaches
Institutional distribution operates at enormous scale
View evidence +

Approximately 50% of rental transactions at Avis locations in 2025 originated from travelers renting under corporate contracts or through affiliations with partner organizations.

  • Avis FY2025 Form 10-K, filed February 19, 2026 (SEC EDGAR)

Wellhub (formerly Gympass)

FACT
Constraint
Fixed facility cost, low average utilization
Distribution move
Employer-sponsored distribution into partner facilities
What happened
5M+ employee subscribers; check-ins from 100M in 2022 to 1B in 2025
What it teaches
Aggregate contracted demand into the hours capacity sits unused
View evidence +

Wellhub reports nearly 40,000 corporate clients, more than 5 million employee subscribers, and 100,000 partner facilities globally including 25,000 in the US. Gross payouts to US fitness operators more than doubled year over year. Reported revenue of $319M in 2025 comes from a secondary source and is flagged as such. Operator-side outcomes — 89% reporting higher retention, 73% reporting increased profitability — come from Wellhub's own survey of 600+ of its partner operators, a vendor-commissioned study of that vendor's partners. Directionally useful, not independent.

  • Wellhub press release, March 10, 2026
  • Health & Fitness Association, September 10, 2025 (vendor-commissioned operator survey)
  • Latka, November 2025 (secondary source, flagged)

ClassPass

FACT
Constraint
Perishable class inventory; average studio at 37% capacity
Distribution move
Marketplace listing only excess capacity, priced dynamically
What happened
$3.1bn cumulative revenue reported as generated for partners
What it teaches
Yield channels work best where marginal cost is near zero
View evidence +

ClassPass reports $3.1bn cumulative partner revenue globally; partners generating more than $1M annually rose 28% between 2024 and 2025; 94% of bookings come from users new to that studio. Classes over 80% full with direct members earn a 45% higher payout on average than classes under 50% full, because dynamic pricing prices scarcity. The transfer caveat is decisive: a class seat has near-zero marginal cost. Assets with real per-use cost — fuel, wear, cleaning, claims — can lose contribution on the same discounting behavior.

  • ClassPass Industry Impact Report, reported March 27, 2026
  • ClassPass partner blog, March 6, 2026

Washington DC fleet share

FACT
Constraint
Municipal vehicles used roughly 30% of the time
Distribution move
Shared-access technology deployed onto city-owned vehicles
What happened
360 vehicles replaced by 58; roughly $1M annual saving reported
What it teaches
Sometimes the opportunity is removing capacity, not filling it
View evidence +

The pilot began October 2008 with 29 vehicles at four sites, expanding to 58 vehicles at eight sites. DC replaced 360 individually assigned vehicles across 2008 and 2009, an 8% fleet reduction, reporting roughly $1M a year and a projected $6.6M over five years. The city paid $115 to $125 per vehicle per month for the managed reservation and access layer. Note the revenue model: this is software licensing against someone else's assets, not utilization of the operator's own. Recency caveat: this evidence is 2008–2014 and current program status could not be verified.

  • Government Technology
  • DC Department of Energy and Environment release
  • Greater Greater Washington
  • Government Fleet, March 2012

Equity Residential

FACT
Constraint
Structured parking cost plus resident mobility expectations
Distribution move
Shared vehicles placed inside an apartment portfolio
What happened
Expanded from 30 properties to more than 100 locations
What it teaches
Distribution is stronger when the partner supplies the location
View evidence +

Equity Residential launched in 2011 across 30 US properties and later reported vehicles at over 100 locations. In San Francisco, Veritas Investments expanded from a pilot to more than 30 vehicles across its properties. The property owner's motive is financial rather than goodwill: structured parking is expensive to build, and municipalities reduce parking minimums in exchange for shared vehicles. Austin reportedly eliminated roughly 1,100 required parking spaces by reducing minimums by 20 spaces per fleet vehicle, saving developers over $38.5M — a secondary source, flagged. Research compiled by the Victoria Transport Policy Institute finds no multifamily development in its sample recovered the full cost of on-site parking through parking fees.

  • Multi-Housing News
  • Mobiag (secondary source, flagged)
  • Victoria Transport Policy Institute

Hertz and Uber

FACT
Constraint
Fleet utilization against weekly rather than daily demand
Distribution move
Gig-driver rentals at scale
What happened
~50,000 drivers rented; 24M+ trips reported by June 2022
What it teaches
High utilization can arrive with high depreciation and claims exposure
View evidence +

Announced October 27, 2021 with up to 50,000 vehicles targeted for gig drivers by 2023. By June 2022 nearly 50,000 drivers had rented, completing more than 24 million fully-electric trips and over 260 million electric miles. Extended to Europe in January 2023. Hertz's FY2023 10-K describes rideshare rental as longer-duration than traditional business rentals and as a source of higher-mileage vehicles for disposition. Disconfirming: Avis took an approximately $500m EV write-down at year-end 2025 and shortened assumed EV economic life from 36 months to roughly 18. High-mileage usage is not free.

  • Hertz press release, October 27, 2021
  • Uber newsroom, June 28, 2022 and January 17, 2023
  • Avis Q4 2025 earnings call, February 19, 2026

HyreCar

FACT
Constraint
Needed recurring, high-hour demand for third-party vehicles
Distribution move
Gig-driver marketplace as the primary business
What happened
Chapter 11 in February 2023; assets sold, then wound down
What it teaches
Maximum utilization can still produce terrible economics
View evidence +

HyreCar filed Chapter 11 on February 27, 2023 in Delaware, case 23-10259, with $5m in debtor-in-possession financing. Getaround acquired the assets in March 2023 for $9.45M, then shut all US operations in early 2025 citing lack of liquidity and risk to its US liability insurance coverage. Mechanism of failure was insurance and claims: gig mileage breaks consumer-rental actuarial assumptions, rising deposits pushed the product beyond its intended customer, and slow claims handling destroyed supply. The same failure appeared earlier at Uber's Xchange Leasing and at Fair.com.

  • HyreCar press release via GlobeNewswire, February 27, 2023
  • Coverager
  • TechCrunch
  • The Rideshare Guy, 2020 (secondary source, flagged)

GIG Car Share

FACT
Constraint
Per-vehicle cost against hours sold
Distribution move
Consumer distribution backed by a ~60M-member motoring organization
What happened
Shutdown announced August 2024; operations ceased December 27, 2024
What it teaches
Access to a huge membership base is not enough
View evidence +

GIG Car Share, backed by AAA Mountain West Group through A3 Ventures, announced shutdown in August 2024 citing decreased demand and rising operational costs, ceasing operations effective December 27, 2024. AAA is precisely the demand-aggregating membership organization a partnership thesis would nominate: a car-centric brand with existing roadside and insurance relationships and local density. It was not enough.

  • Next City, November 2024
  • UC Berkeley Parking and Transportation notice

IKEA mobility partnerships

FACT
Constraint
Customers unable to transport large purchases
Distribution move
Vehicles stationed at the retail destination, sold at point of purchase
What happened
Programs persisted across multiple European markets
What it teaches
The strongest acquisition moment attaches directly to the need
View evidence +

Renault launched a van sharing service for IKEA customers in France from Paris-Villiers in 2018, targeting 300 vehicles by September as an extension of Renault Mobility's 180 self-service locations. IKEA Germany currently lists van rental at stores through Cambio Car Sharing, with additional stores served by Kölner Flitzer and CarlundCarla.de. IKEA Switzerland directs customers to vans through Europcar provided by Allride, located in store car parks. Current listing on the retailer's own service pages is evidence of continued operation rather than of announcement. Disconfirming: the same retailer has been building delivery, lockers, and e-cargo-bike rental, which may cap the channel.

  • The Drive, April 2018
  • ikea.com/de and ikea.com/ch service pages
  • Supply Chain Digital, April 2025

04 / The mechanisms

Not all distribution is equal

01

Fill the trough

Bring recurring demand into periods where capacity would otherwise sit unused.

  • employers
  • institutions
  • government
  • business customers
  • weekday demand

Key question
Does this usage happen when the asset would otherwise be idle?

02

Remove a fixed cost

Structure the partnership so someone else supplies something the business currently pays for.

  • parking
  • real estate
  • infrastructure
  • customer acquisition
  • charging
  • capital
  • distribution

Key question
What cost disappears because the partner exists?

03

Contract the demand

Replace uncertain individual demand with an institution capable of repeatedly assigning customers or usage.

  • insurers
  • employers
  • governments
  • universities
  • property portfolios
  • purchasing cooperatives

Key question
Can one relationship produce recurring demand from hundreds or thousands of users?

05 / Failure modes

Six ways partnership strategies fail

01

More customers, same economics

Demand increases but contribution per asset does not. The channel is reported as growth and shows up nowhere in the unit economics.

02

Discounting the wrong inventory

The channel fills capacity that would have sold anyway, or prices below marginal cost in periods that were never the problem.

03

High-utilization, high-cost customers

Some customers maximize usage while simultaneously maximizing insurance, maintenance, claims, or depreciation.

04

Geographic mismatch

A national audience is worth little when supply is constrained to individual physical locations.

05

Subsidy dependence

The model works only while parking, infrastructure, tax treatment, or public funding support remains in place.

06

Aggregation without structural advantage

A large membership organization or platform supplies reach without changing anything about the underlying business.

The historical record is full of businesses that solved distribution without solving economics.

06 / Decision framework

Before adding another acquisition channel, ask seven questions

  1. 1

    Where is capacity actually unused?

    Not theoretically. By location, hour, day, and asset.

  2. 2

    What does an additional unit of usage cost?

    Include servicing, labor, claims, depreciation, support, energy, and other variable costs.

  3. 3

    Who aggregates demand during the trough?

    Look for organizations, not audiences.

  4. 4

    Is that demand geographically aligned with supply?

    The tighter the physical match, the more valuable the channel.

  5. 5

    Can the partner absorb an existing cost?

    Distribution is more valuable when it changes both sides of the equation.

  6. 6

    Is the demand recurring or promotional?

    Contracted usage compounds. One-time discounts usually do not.

  7. 7

    Does contribution per asset improve?

    This is the final test. Everything above is preparation for it.

07 / Channel attractiveness

Where a channel sits determines what it can do

Recurring / contracted demand
National membership organizations
Credit cards / loyalty
Consumer marketplaces
Retail destinations
Employer benefits
Universities
Government / institutions
Property portfolios
Geographic alignment: lowHigh

The closer a channel gets to recurring demand at the exact location of unused capacity, the more interesting the economics become.

Channels in the lower-left produce reach. Channels in the upper-right produce contracted usage against specific assets. Both are described as partnerships. Only one of them reliably changes contribution.

Positions are analytical judgements drawn from the case set, not measured values.

08 / The counterintuitive finding

The best growth strategy may start with a cost line

Most acquisition strategies begin with the customer. For businesses with fixed physical capacity, history suggests starting somewhere else:

What are we already paying for that isn’t earning enough?

Then work backward.

  • Which hours are empty?
  • Which locations underperform?
  • Which fixed costs could a partner absorb?
  • Who already aggregates demand around those exact assets?

That changes partnership strategy from a reach exercise into an economics exercise. Large audiences become less important. Fit becomes more important.

09 / From evidence to action

Finding the channel is only the beginning

The historical cases suggest where to look.

Turning that into a growth channel requires a second step: connecting a specific capacity problem to an organization with the demand, economics, and incentive to solve it.

The strongest partnership strategies followed a similar sequence.

  1. 01

    Map the capacity problem

    Start with the asset, not the audience. Identify:

    • which locations underperform
    • which hours or days are empty
    • which capacity is already paid for
    • which fixed costs are creating pressure
    • where incremental usage has attractive contribution

    Output
    A precise definition of the capacity that needs demand.

  2. 02

    Find who aggregates the demand

    Do not start by asking which companies have large audiences. Ask who already controls access to the people who need that capacity at the right location and time. Potential aggregators include:

    • employers
    • property owners
    • institutions
    • governments
    • universities
    • retailers
    • insurers
    • platforms
    • purchasing cooperatives

    Output
    A map of potential distribution channels and the organizations inside them.

  3. 03

    Find the partner's economics

    A partnership needs a reason to exist for both sides. Determine what the potential partner gains:

    • lower costs
    • less owned capacity
    • higher customer value
    • employee benefits
    • better utilization
    • avoided infrastructure
    • additional revenue
    • differentiated service
    • operational efficiency

    Output
    The reason each partner should care.

  4. 04

    Design the partnership

    Turn the shared economics into an actual model. Define:

    • who pays
    • what gets bundled
    • where capacity sits
    • who owns the customer relationship
    • what behavior the partnership should create
    • what each party contributes
    • what a pilot looks like
    • how success will be measured

    Output
    A partnership proposition that can actually be taken to market.

  5. 05

    Build the channel

    Strategy becomes useful when someone takes it into the market. That means:

    • prioritize target accounts
    • identify the right decision-makers
    • develop partner-specific propositions
    • open conversations
    • structure pilots
    • negotiate economics
    • measure results
    • expand what works

    Output
    A repeatable distribution channel, not a list of partnership ideas.

A partnership opportunity isn’t a company that could theoretically work with you. It’s a specific counterparty with demand you need, economics you can improve, and a reason to act.

That distinction turns partnership strategy from brainstorming into business development.

Five places the evidence says to look

These are not recommendations for any particular company. They are recurring structures found across the historical cases.

01

Property portfolios

They aggregate
People around a specific physical location.
They can contribute
Customers, real estate, parking, infrastructure.
Works when
Location is part of the capacity problem.

02

Institutional capacity substitution

They aggregate
Recurring organizational usage.
They can contribute
Contracted demand and predictable weekday utilization.
Works when
The institution currently owns or pays for underused capacity of its own.

03

Employer distribution

They aggregate
People around predictable locations and schedules.
They can contribute
Recurring users, subsidy, distribution.
Works when
The operator's trough overlaps with working hours.

04

Cooperative purchasing

They aggregate
Many institutional buyers behind one procurement relationship.
They can contribute
Scalable access and lower business-development cost.
Works when
Selling one organization at a time is too expensive.

05

Point-of-need distribution

They aggregate
Customers at the exact moment another transaction creates demand.
They can contribute
Location, intent, distribution, and sometimes infrastructure.
Works when
The need for capacity is triggered by another purchase or activity — retail, travel, healthcare, moving, events.

10 / Building the channel

What this looks like in practice

  1. Capacity problem
  2. Channel hypothesis
  3. Partner universe
  4. Priority accounts
  5. Partner proposition
  6. Pilot
  7. Repeatable channel

A useful partnership strategy should eventually produce something tangible:

  • Which channel should we pursue?
  • Which companies should we approach?
  • Why should each of them care?
  • What should we propose?
  • What would we test first?
  • What would tell us to scale it?

Research identifies the pattern.
Strategy determines where it applies.
Business development turns it into a channel.

11 / The rule

Don’t ask how many customers a partnership can bring. Ask what happens to contribution from the capacity you already own.

That is the difference between distribution as marketing and distribution as strategy.

And once the right distribution model is identified, the work shifts from analysis to building it.

Method, evidence, and limits

Case set
Twenty-four historical cases across mobility, rental, fitness, real estate, government, marketplaces, employer benefits, and institutional distribution. Ten are displayed above.
Evidence weighting
SEC filings, earnings calls, company disclosures, government records, partner disclosures, and credible business reporting. Partnership announcements alone were not treated as evidence of success.
Evidence labels
FACT — directly supported by a dated source. INFERENCE — a conclusion supported by multiple facts. HYPOTHESIS — plausible but not demonstrated.
Stated caveats
Wellhub operator outcome data comes from a vendor-commissioned study of that vendor’s own partners. Some shared-fleet government evidence dates to 2008–2014 and current program status could not be verified. Some real-estate savings figures rely on secondary sources. Partnership announcements are not proof of success. The case set documents failures more reliably than quietly profitable programs. Where the record establishes correlation, no causal claim is made.
Citations
Sources are listed inside each case row as publication and date. Where the underlying research did not record a URL, none is shown rather than reconstructed.

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 you would have to build to act on them.

Talk through a decision