## Money
Real costs come from:
— initial preprocessing (names + indicators)
— storing the small non-deterministic layers
— serving tiles globally
— occasional indicator + naming rebuilds (epochs)
— coordinating shared volunteer compute
— running paid tools (Maps + What if)
### Serving eight-hundred trillion unique hexagons
deterministic backbone
Level 4–6 generated on demand
cached briefly
discarded
oceans, deserts, ice, tundra:
use deterministic filler dictionaries
no storage, no special vocab consumption
community overrides:
names, traces, metadata
small and sparse
stored layers
Level 0–3 names, themes, indicators, metadata
L4–L5 only when overridden
L6 almost entirely virtual
epochs stored as snapshots (rare)
estimated storage
L0–L3 indicators + maps + metadata: ~200–400 GB
L4–L6 overrides + name shards: ~60–160 GB
epoch archives: +50–100 GB/epoch
total baseline (active epoch + history): ~300–500 GB
storage cost
object store (R2/S3): $7–15/month
growth slow + predictable
— deterministic hex generation
— shallow levels only stored
— tiny tiles + high cache ratio
— empty regions use filler dictionaries
— community layers light + sparse
— indicator rebuilds cheap and infrequent
result
public layer (names, grid, indicators, epochs) stays free
running costs remain tens → low hundreds per month even at scale
hexagons.world defines hex0–hex6 everywhere on Earth: from ~1,000 km down to ~1 m. In theory that’s on the order of 10¹⁵ cells. In practice, almost all of them are deterministic. Themes and wordlists are derived from open data (UNESCO, ecoregions, land-use, climate, species) plus a coordinate-based seed.
A hex only stops being “purely deterministic” when a group of people override something – a local name, a tag, an indicator edit, metadata, or a photo. Even if we gave every inhabited 1 km and 100 m cell special attention, that’s roughly a billion hexes out of hundreds of trillions – about one in a million. Everything else is maths, not storage.
tile generation
10–50 KB tiles
deterministic maths at the edge
L4–L6 tiles regenerated on demand
>95–99% cache hit ratio
request patterns
casual: 30–80 tiles/month
heavy: 300–500/month
research: 1,000+/month
cost per million requests
Workers + KV: ~$0.15–0.30/million
rounded monthly cost
10k MAU → ~$10–20
100k → ~$20–60
1M → ~$80–150
10M → ~$300–600
100M → ~$2,000–4,000
keeps cheap because:
— tiny tiles
— high cache ratios
— deterministic deep layers
#### Pre-processing
initial ingest (~200–300 GB):
UNESCO, ecoregions, species ranges, OSM, climate, terrain, admin data
name assignment
combine biome + terrain + species + settlement
LLM filtering to remove bad fits
assign L0–L3 globally; L4–L5 only where inhabited
spot compute total: ~$250–550
indicator build
preprocess + modelling + QA
total statistical build: **~$50–70**
epochs
only when world/methods change
recompute names + stats: ~$40–60/epoch
rare, not monthly
#### Ongoing updates
community naming
sparse rows only (L4–L6 mostly untouched)
reviewer time = human cost
storage impact minimal
events + facilities
small structured records keyed by hex
few GB even with millions of items
### Maps ($1/month)
Maps is inherently useful in a way that doesn’t need much explanation. For many people, there are specific reasons to keep it — safer routes at night, reliable local information, or something that feels worth supporting, even at a small cost.
The free tier can support much larger adoption than the paid tier. Everyday use can remain free:
- base maps
- ordinary directions
- offline local tiles
- events and facilities
- predictable, private defaults
The paid tier is for heavier compute and deeper personalisation:
- well-lit routes at night
- shade-first routing
- multi-stop errands
- comfort / safety / energy-aware routing
- recomputation and advanced personal layers
Maps operates as a system whose raw per-request and platform costs are extremely low and predictable; so, no matter the amount of people pay, it's more than enough to keep the Maps and API layers free and calm for everyone, and fund a subtantial surplus into direct, local work (food, housing, shade, mental health) without ever touching advertising or surveillance.
This means payment and adoption should be treated separately.
couple of factors to motivate paid adoption
#### Frictionless Agree Amount in Local Currency
we will charge one dollar a month
and feels like one dollar everywhere in the world
Frictionless-agree amount ≈ 0.5%–1% of a typical daily wage
rounded to a common small denomination.
plus market clustering
- one euro
- one pound
- one *(five halalas as the feeling)* riyal
- 10k rupiah
- 50k peso
- 10 lira
#### Relaying a month to someone
relay a month to someone
allows pooled payments in cash
(useful for people without debit cards) licenses sharable:
— irl via QR
— online via a short encoded phrase
users privately see the country and flag their current license came from
(not public; relies on honesty)
Relay/gifting does two things at once:
It increases trial: people can be handed a working license instantly (QR / phrase), so they can adopt before they’ve even decided to pay.
It increases paid-month coverage: even if a user never pays digitally, the system can still collect the dollar (parents → kids, friends → friends, donors buying stacks, cash pooling).
So adoption is not “blocked” by payments. Payment becomes a social flow, not a checkout funnel. That matters because it lets MAU rise faster than traditional “subscription conversion” would allow.
#### Who might opt-in to support the map?
Paid adoption here shouldn't be thought of as a normal premium subscription funnel.
Most people won't be making a hard, analytical decision about “whether the advanced tier is worth it” in the way they might for business software or entertainment subscriptions.
Instead, support emerges inside a particular environment:
- the map is already useful in ordinary life
- the free layer is generous
- the paid layer solves heavier or more personal problems
- the operator is visibly non-extractive
- Peaceful Foundation gives the whole thing a wider moral shape.
People aren't primarily responding to features. They're responding instead to whether the tool feels trustworthy, whether it fits into daily life, and if giving a dollar feels natural rather than coerced.
So, paid adoption is best understood as the combined effect of three reinforcing conditions, not one isolated conversion lever. In practice, people tend to support when a few simple things line up:
- the amount feels negligible
- the product keeps proving useful
- and the surrounding story feels clean enough to trust.
Within that setting, three factors matter most.
##### Factor A — $1 Is Psychologically Not a “subscription”
$1/month:
is below most people’s “should I cancel this?” threshold
is often rounded away mentally
feels more like:
a tip
a membership
a contribution
a token of support
Especially when:
it’s framed as “one dollar everywhere”
it can be gifted
it’s visibly keeping something good alive
So churn behaves very differently.
##### Factor B — Maps' Usefulness Is Local and Recurring
Maps is not a “once a month” app.
People use maps when they:
walk
commute
run errands
meet others
explore
travel
help someone
Even light usage creates habitual presence, which supports retention.
Subset of above, but stronger:
organisers
frequent walkers
safety-conscious users (especially women)
delivery / navigation-heavy users
They:
feel real value repeatedly
anchor retention
rarely churn
This lifts the overall average.
##### Factor C — Opt-in Is Shaped by Framing
At $1/month, conversion becomes mostly a function of:
Trust (charity + open source + no ads)
Default positioning (“this is your normal map”)
A moral story that doesn’t guilt
Gifting/relay (people cover others)
A charity-run, open-source replacement:
reduces fear of enshittification
reassures people it won’t “turn evil later”
lowers switching anxiety
encourages long-term commitment
This creates identity-aligned retention, not feature-aligned retention.
People like this:
tolerate minor imperfections
don’t churn on small annoyances
feel a sense of participation, not consumption
so therefore,
price ≈ negligible
trust ≈ high
usage ≈ recurring
framing ≈ contribution, not extraction
Adoption can become very large even if only a minority of users ever pay.
So paid conversion comes from the subset of people who either rely on the map more heavily, or feel comfortable supporting something they use.
##### Who is donating towards the map?
There are three payer modes, not one.
1. Self-supporters
People who:
use the app
like it
pay $1
forget about it
stay subscribed for a long time
This is the base.
They:
use the map
feel neutral-to-positive
don’t think much about the payment
Because $1 is frictionless: This is already higher than SaaS because:
no pricing pressure
no feature gating
no extraction feeling
2. Gifters
People who:
buy months for others
relay access
pool cash
support people without cards
This is not marginal — it increases effective revenue without increasing MAU.
Because:
price is tiny
ethos is pro-social
gifting is culturally natural
This creates:
revenue > MAU conversion
Not by much, but meaningfully.
Not everyone gifts — but enough do.
3. Institutional spillover (indirect)
Some people pay because they:
trust Peaceful Foundation
participate in campaigns
see Maps as shared infrastructure
So they pay not because of maps alone, but because:
Trust the operator, like the ethos, and are more willing to leave $1 on even without heavy feature use.
“this is part of something I’m already in”
Payment should be modelled as opt-in support, not as a normal premium subscription.
The rate depends on:
- who the retained users are
- how often they rely on the map
- how embedded the map is in everyday life
- how strongly they trust Peaceful Foundation and the product’s non-extractive design
As such, conversion differs by scenario.
#### Minimal
the map is chosen deliberately
~0.25% of global maps users
~6–8 million users
adoption is values-driven
the map is mostly secondary
trust is high
most users still treat it as a secondary map
there are few social or environmental reinforcements
Peaceful Foundation exposure exists,
but is still limited
reasonable direct self-support conversion
~1.5–2.7%
reasonable effective conversion
after gifting and spillover
~1.8–3.2%
around: ~110k – 260k people
this is a small but unusually aligned user base
these are exactly the sorts of people
who care about:
privacy
calm tools
open source
non-corporate infrastructure
they are more willing than average
to support something morally clean
that lifts trust
and makes support more natural than in an ordinary niche app
but the cap is still clear
most do not rely on the map daily
most still keep another map as default
the map is chosen on principle,
more than on embedded daily usefulness
so the band lands in the low single digits
below this
you are underweighting the unusually trust-aligned audience
above this
you are pretending niche values users already behave like default-map users
working intuition
this is:
“a meaningful minority of a values-heavy niche”
#### Reasonable
the map is now used repeatedly
~0.5–1.0% of global maps users
~12–28 million users
it becomes a second default for many
and a primary task-specific map for some
retention matters more than discovery
word-of-mouth and local visibility start to matter
Peaceful Foundation exposure is now more present,
and no longer just a light adjacency
reasonable direct self-support conversion
~2.2–3.8%
reasonable effective conversion
after gifting and spillover
~2.6–4.5%
around: ~310k – 1.26M people
now the map is no longer being supported mainly
by people who simply like the idea
it is being supported by people
who keep finding themselves using it
for:
campus use
walking
transit
offline reliability
work navigation
these are repeated use cases,
not abstract preferences
that changes the feel of payment
the average user is now not just aligned,
but repeatedly relying on the map in specific situations
this lifts habitual willingness to support,
because the map has started to earn its place through ordinary use
most support still comes from:
self-supporters first
then a growing amount of gifting
then a noticeable Peaceful Foundation spillover tail
the free tier remains sufficient for many users
the map is often a second default,
not yet a universal primary one
use is still situational for many,
not fully embedded across everything
so the band rises,
but not explosively
below this
you are underweighting a retained base
that now uses the map repeatedly in ordinary life
above this
you are pretending a useful second default
already behaves like a socially assumed shared map
working intuition
this is:
“a useful recurring second default with strong trust”
#### Ambitious
the map is now used repeatedly across communities
~1.0–1.8% of global maps users
~30–80 million users
adoption is socially reinforced
the map is repeatedly encountered across:
campuses
local events
Peaceful Foundation projects
everyday local activity
in some environments,
it becomes expected or recommended
reasonable direct self-support conversion
~2.8–4.8%
reasonable effective conversion
after gifting and spillover
~3.4–5.8%
around: ~1.0M – 4.6M people
by this stage,
the map is no longer spreading mainly one person at a time
it starts spreading through places
the question is no longer just:
“do I personally find this useful?”
it becomes:
“this is part of how things work around here”
that shifts all three payer modes upward
self-support rises
as more users treat it as a genuine default
gifting rises
as usage becomes socially visible
spillover rises
as the map begins to feel like shared infrastructure
the average user is now encountering the map
not just through personal use,
but through their surrounding environment
that makes support feel less like
paying for a feature
and more like
leaving a small amount on
for something good that is becoming part of local life
the ceiling is still real
incumbents still dominate globally
adoption remains uneven
the map is strong in pockets and regions,
but not yet universal
so the band moves up,
but does not explode
below this
you are underweighting the shift
from individual retention
to social reinforcement
above this
you are pretending every socially visible utility
converts like a hard paywall SaaS
working intuition
this is:
“a map that feels normal in some places”
#### Optimistic
the map is now used repeatedly across regions and networks
~2–4% of global maps users
~50–150 million users
it becomes a “third default”
it is widely recognised even by non-users
normal across multiple regions and demographics
social norms reinforce usage
families,
friend groups,
campuses,
and neighbourhood contexts
begin to standardise around it
reasonable direct self-support conversion
~3.0–5.0%
reasonable effective conversion
after gifting and spillover
~3.8–6.5%
around: ~1.9M – 9.8M people
here the map starts to feel familiar,
not just admirable
many people encounter it
through something useful first
a cheap meal poster
a campus noticeboard
a friend sharing a hex name
a parent installing it for safety
that strengthens:
trust
default-use
gifting behaviour
the average user is now often surrounded by other users
this lifts support again,
because the map has crossed into:
household logic
group logic
regional habit
most support still comes from:
self-supporters first
then a materially stronger amount of gifting
then a larger Peaceful Foundation spillover tail
but the free product still remains enough
for a great many people
many users still sit happily on the free tier
only some users need:
deeper safety routing
deeper personalisation
heavier compute
so the band rises again,
but still stays well short
of mass consumer payment behaviour
below this
you are underweighting a map
that has become culturally normal in some environments
above this
you are assuming mass consumer payment behaviour
across the whole global base
working intuition
this is:
“a socially normal, low-friction public utility in many environments”
#### Reasonable, visionary and optimistic
the map is now used repeatedly as infrastructure
~1.0–1.5 billion main users
plus another ~1.0–2.0 billion secondary or situational users
it is no longer mainly behaving
as a consumer replacement app
it functions as:
an addressing layer
a shared reference layer
a neutral civic substrate
at this stage,
one global conversion band stops being the cleanest model
the better question is:
what kinds of places does the map live inside,
and how does support behave there?
##### A) Full-stack Zones
here the map is woven into daily life
people keep encountering it
through:
youth posters
food systems
mutual-aid chats
local campaigns
campus spaces
the map is not just useful here
it is part of how things work
reasonable direct self-support conversion
~6–10%
reasonable effective conversion
after gifting and spillover
~8–12%
around: ~80M – 130M people
this feels less like:
buying an app upgrade
and more like:
keeping shared local infrastructure alive
##### B) Partial-stack Zones
here the map is clearly useful
and often kept
but it is not yet ambient everywhere
people often arrive through one or two strong needs:
cheap food
safe routes
hex names
the map matters,
but it is not yet the whole surrounding environment
reasonable direct self-support conversion
~3–5%
reasonable effective conversion
after gifting and spillover
~4–7%
around: ~12M – 28M
this feels like:
a strong local layer
not yet:
full infrastructure
(c) outer zones
here the map still spreads more individually
people come because they need:
addressing
safety
offline use
a calmer alternative
the surrounding environment does less of the work
reasonable direct self-support conversion
~1–2.5%
reasonable effective conversion
after gifting and spillover
~1.5–3.5%
around: ~0.45M – 3.2M
this still produces real support
but it behaves more like:
replacement-app adoption
than:
locally assumed infrastructure
overall,
conversion should be derived from zone mix,
not assumed globally
a plausible blended rate is:
~4–8%
or around:
~90M – 160M people
working intuition
this is:
“support for shared infrastructure”
not:
“a global subscription product”
| Scenario | Users (MAU) | Conversion (effective) | Supporters (opt-in) |
|--------------------------------|-------------------|------------------------|---------------------|
| Minimal | 6–8 million | 1.8–3.2% | 110k – 260k |
| Reasonable | 12–28 million | 2.6–4.5% | 310k – 1.26M |
| Ambitious | 30–80 million | 3.4–5.8% | 1.0M – 4.6M |
| Optimistic | 50–150 million | 3.8–6.5% | 1.9M – 9.8M |
| | | | |
| Infrastructure (full-stack) | 1.0–1.1 billion | 8–12% | 80M – 130M |
| Infrastructure (partial-stack) | 300–400 million | 4–7% | 12M – 28M |
| Infrastructure (outer zones) | 30–90 million | 1.5–3.5% | 0.45M – 3.2M |
| | | | |
| Infrastructure (total) | 1.33–1.59 billion | ~4–8% (blended) | 90M – 160M |
#### Revenue for Maps
The Maps revenue case should be understood as opt-in support rather than a normal premium subscription funnel.
Most users can still use the public map freely. Paid coverage comes from people who either use the map heavily, receive a gifted month, want advanced routing and personalisation, or simply trust the project enough to keep a tiny recurring contribution active.
| Scenario | Users (MAU, average) | Conversion (effective, average) | Supporters (opt-in, average) |
| ------------------------------ | -------------------: | ------------------------------: | ---------------------------: |
| Minimal | 7 million | 2.5% | 185k |
| Reasonable | 20 million | 3.55% | 785k |
| Ambitious | 55 million | 4.6% | 2.8M |
| Optimistic | 100 million | 5.15% | 5.85M |
| Infrastructure (total) | 1.46 billion | ~6% blended | 125M |
So the important number is not just “conversion”. It is how many people are covered for a paid month through self-payment, gifting, pooled cash, institutional spillover, and local relay.
Pricing is locally weighted, so the effective monthly amount is not always exactly USD $1. In wealthier regions it may sit close to $1. In lower-income regions it may be lower, because the point is that the payment should feel similarly small everywhere.
***people who end up covered for a paid month*** × ***the amount that feels like one dollar there*** × ***how much we actually keep after fees***
The estimate below uses average supporter counts from the adoption table, plus midpoint assumptions for local pricing and retained revenue.
Minimal
~185k supporters × ~$0.95 × ~92.5%
→ ~$163k / month
→ ~$2.0M / year
Reasonable
~785k supporters × ~$0.90 × ~88.5%
→ ~$625k / month
→ ~$7.5M / year
Ambitious
~2.8M supporters × ~$0.825 × ~84%
→ ~$1.94M / month
→ ~$23.3M / year
Optimistic
~5.85M supporters × ~$0.75 × ~80%
→ ~$3.51M / month
→ ~$42.1M / year
Infrastructure
~125M supporters × ~$0.75 × ~81.5%
→ ~$76.4M / month
→ ~$917M / year
**total revenue** = ***effective paid coverage*** × ***local calm price*** × ***collection efficiency***
#### × ***effective paid coverage***
effective paid coverage = self-support + gifting + relay + spillover
(this is calculated in the section above)
#### × ***local calm price***
local calm price
= the locally rounded “feels like one dollar” amount
it is:
- roughly ~0.5–1% of a typical daily wage
- rounded to a normal local denomination
- clustered to simple, familiar values
examples:
£1
€1
10k rupiah
50 peso
10 lira
the goal is:
it feels tiny,
normal,
and not worth thinking about
| Region | Median daily wage band | Local denomination cluster | Local calm price | Rail quality | Relay friendliness | Collection efficiency |
|----------------------|------------------------|----------------------------|------------------|--------------|--------------------|-----------------------|
| America | medium–high (wide) | 1 / 5 / 10 / 50 / 100 | ~$1 equiv | medium–high | medium–high | ~80–95% |
| | | | | | | |
| Atlantic Europe | high | €1 / £1 | €1 / £1 | high | medium | ~90–97% |
| Continental Europe | medium–high | €1 | €1 | high | medium | ~90–97% |
| Mediterranean Europe | medium | €1 | €1 | medium–high | medium | ~85–95% |
| Nordic Europe | high | 10–20 local units | ~$1–2 equiv | high | medium | ~90–97% |
| Eurasian Plains | low–medium | 20 / 50 / 100 | ~$0.5–1 equiv | medium | medium–high | ~75–90% |
| | | | | | | |
| Continental Asia | low–medium | 10 / 20 / 50 / 100 | ~$0.5–1 equiv | medium | high | ~70–90% |
| Monsoonic Asia | low | 5k / 10k / 20k style | ~$0.3–0.8 equiv | medium | high | ~70–90% |
| Oceanic Asia | medium | 10k / 50 / 100 style | ~$0.8–1 equiv | medium–high | high | ~75–92% |
| | | | | | | |
| Oceania | high | 1 / 2 local units | ~$1–2 equiv | high | medium | ~90–97% |
| | | | | | | |
| Mediterranean Africa | low–medium | 5 / 10 / 20 | ~$0.5–1 equiv | medium | high | ~70–90% |
| Saharan Africa | low | 5 / 10 / 20 | ~$0.3–0.8 equiv | low–medium | high | ~60–85% |
| Monsoonic Africa | low | 500 / 1k / 2k style | ~$0.3–0.7 equiv | low–medium | high | ~60–85% |
| Highland Africa | low | 5 / 10 / 20 | ~$0.3–0.7 equiv | low–medium | high | ~60–85% |
| Southern Africa | low–medium | 5 / 10 / 20 | ~$0.5–1 equiv | medium | medium–high | ~70–90% |
| | | | | | | |
| Arabia | medium–high | 1 / 5 / 10 | ~$1 equiv | medium–high | medium–high | ~80–95% |
#### × ***collection efficiency***
// more considerations here
collection efficiency
= the share actually retained after practical payment friction
where possible, minimise fees transparently
use direct bank transfer where possible
use Wise or similar low-friction rails
avoid app-store cuts where possible
avoid routing everything through card processors when local banking works fine
so in practice:
*High efficiency*
direct transfer easy
Wise or similar works
low-fee rails available
relay easy
Keep rate: ~90–97%
*Medium efficiency*
some friction
some card reliance
some wallet fragmentation
relay still workable
Keep rate: ~80–90%
*Lower efficiency*
fragmented rails
higher fee dependency
more difficult direct settlement
relay helps, but collection still leaks more
Keep rate: ~65–80%
| Region | Local calm price (USD-equiv) | Collection efficiency | Net revenue per covered month |
|------------------------|------------------------------|-----------------------|-------------------------------|
| America | 1.00 | 0.90 | 0.90 |
| Atlantic Europe | 1.00 | 0.95 | 0.95 |
| Continental Europe | 1.00 | 0.94 | 0.94 |
| Mediterranean Europe | 0.90 | 0.90 | 0.81 |
| Nordic Europe | 1.20 | 0.95 | 1.14 |
| Eurasian Plains | 0.60 | 0.82 | 0.49 |
| Continental Asia | 0.70 | 0.82 | 0.57 |
| Monsoonic Asia | 0.50 | 0.78 | 0.39 |
| Oceanic Asia | 0.80 | 0.85 | 0.68 |
| Oceania | 1.10 | 0.95 | 1.05 |
| Mediterranean Africa | 0.70 | 0.80 | 0.56 |
| Saharan Africa | 0.45 | 0.72 | 0.32 |
| Monsoonic Africa | 0.40 | 0.72 | 0.29 |
| Highland Africa | 0.40 | 0.72 | 0.29 |
| Southern Africa | 0.65 | 0.82 | 0.53 |
| Arabia | 0.90 | 0.90 | 0.81 |
From there, the scenario revenues become easier to read.
The social model tells us how many people end up covered.
The regional model tells us roughly what each covered month is worth on average.
That gives us a grounded way to estimate revenue under each adoption pattern.
| Scenario | Avg supporters | Avg monthly value | Avg retained/payment capture | Estimated monthly revenue | Estimated yearly revenue |
| -------------- | -------------: | ----------------: | ---------------------------: | ------------------------: | -----------------------: |
| Minimal | 185k | $0.95 | 92.5% | ~$163k | ~$2.0M |
| Reasonable | 785k | $0.90 | 88.5% | ~$625k | ~$7.5M |
| Ambitious | 2.8M | $0.825 | 84% | ~$1.94M | ~$23.3M |
| Optimistic | 5.85M | $0.75 | 80% | ~$3.51M | ~$42.1M |
| Infrastructure | 125M | $0.75 | 81.5% | ~$76.4M | ~$917M |
#### Minimal
effective paid coverage
~110k – 260k people covered
there is enough support here
for the map to clearly stand on its own
the early base forms where:
trying alternative software is culturally legible
small recurring payments are easy
and the moral shape of the project is immediately understood
this naturally concentrates early support
in higher-trust, lower-friction environments
so the average net revenue per covered month stays relatively high
average net revenue per covered month
~$0.82 – $0.92
monthly revenue
~$90k – $240k
annual revenue
~$1.1M – $2.9M
what this feels like
not mass adoption
a small but real layer of loyal support
strong enough to keep the map calm,
free,
and maintained
at this level,
the system has already proved
that a morally clean map can survive
#### Reasonable
effective paid coverage
~310k – 1.26M people covered
the map now starts to earn support
not only because people like the idea,
but because they keep finding themselves using it
the support base broadens
early high-trust environments still anchor it
but repeated everyday usefulness now matters more:
walking
transit
campuses
reliability
local coordination
this begins to pull in support
from a wider mix of places
where the map fits into ordinary life
so the average net revenue per covered month softens slightly
even as total coverage rises much faster
average net revenue per covered month
~$0.74 – $0.86
monthly revenue
~$230k – $1.08M
annual revenue
~$2.8M – $13.0M
what this feels like
the map has moved beyond curiosity
it is now part of ordinary life
for a real minority of people
not everyone pays
but enough people rely on it often enough
that support begins to feel routine
#### Ambitious
effective paid coverage
~1.0M – 4.6M people covered
by this stage,
the map is no longer spreading mainly one person at a time
it begins to spread through places
campuses
friend groups
local scenes
repeated public encounters
the support base becomes structurally broader
some places contribute more per person
because payment is simple and consistent
others contribute more people
because the map solves stronger coordination problems
this is where:
relay
gifting
and pooled support
start to matter more
so the average net revenue per covered month falls a little further
not because the model weakens
but because it is becoming more global in behaviour
average net revenue per covered month
~$0.64 – $0.78
monthly revenue
~$640k – $3.6M
annual revenue
~$7.7M – $43M
#### Optimistic
effective paid coverage
~1.9M – 9.8M people covered
now the map is carried
by a visibly mixed world
it is no longer strongest only
where payment is easiest
it is strongest
where usefulness,
trust,
and social reinforcement overlap
this creates a layered support base
some places contribute more
because collection is clean and reliable
others contribute more
because the map has become too useful
not to be part of ordinary life
this pushes coverage much higher,
even while the average net revenue per covered month softens again
average net revenue per covered month
~$0.58 – $0.74
monthly revenue
~$1.1M – $7.3M
annual revenue
~$13M – $88M
the map is no longer just useful
it is familiar
people hear about it through:
friends
family
campuses
neighbourhood habits
something useful they needed anyway
#### Reasonable, Visionary and Optimistic
##### A) Full-stack Zones
effective paid coverage
~80M – 130M people covered
these are places where the map is not merely used
it is woven into local life
addressing
food
campus spaces
safe routes
and shared reference points
all reinforce each other
support behaves less like paying for software
and more like keeping infrastructure alive
that is why coverage is so large
and why the average net revenue per covered month
can be lower than earlier scenarios
while total revenue becomes enormous
average net revenue per covered month
~$0.52 – $0.72
monthly revenue
~$42M – $94M
annual revenue
~$500M – $1.1B
##### B) Partial-stack Zones
effective paid coverage
~12M – 28M people covered
these are places where the map is clearly useful
but not yet ambient everywhere
people often arrive through one strong wedge:
safe routes
food
addressing
local reliability
the system is important,
but not yet the whole environment
this is why the average net revenue per covered month
sits slightly higher than in full-stack zones
while total scale remains smaller
average net revenue per covered month
~$0.60 – $0.78
monthly revenue
~$7M – $22M
annual revenue
~$85M – $260M
##### C) Outer Zones
effective paid coverage
~0.45M – 3.2M people covered
here the map is still spreading more individually
people come because they need:
better addressing
offline use
safety
a calmer alternative
the surrounding environment does less of the work
so support is thinner,
and collection is less consistent
but even here,
a real minority still supports it
because the map solves something tangible
average net revenue per covered month
~$0.42 – $0.62
monthly revenue
~$0.2M – $2.0M
annual revenue
~$2.5M – $24M
total
effective paid coverage
~90M – 160M people covered
monthly revenue
~$49M – $117M
annual revenue
~$590M – $1.4B
the map has crossed the line from useful product to quiet infrastructure
| Scenario | Covered people | Avg. net revenue / covered month | Monthly revenue |
|---------------------------------------|----------------------|----------------------------------|----------------------|
| Minimal | 110k – 260k | $0.82 – $0.92 | $90k – $240k |
| Reasonable | 310k – 1.26M | $0.74 – $0.86 | $230k – $1.08M |
| Ambitious | 1.0M – 4.6M | $0.64 – $0.78 | $640k – $3.6M |
| Optimistic | 1.9M – 9.8M | $0.58 – $0.74 | $1.1M – $7.3M |
| | | | |
| Infrastructure — full-stack zones | 80M – 130M | $0.52 – $0.72 | $42M – $94M |
| Infrastructure — partial-stack zones | 12M – 28M | $0.60 – $0.78 | $7M – $22M |
| Infrastructure — outer zones | 0.45M – 3.2M | $0.42 – $0.62 | $0.2M – $2.0M |
| | | | |
| Infrastructure — total | 90M – 160M | blended | $49M – $117M |
Within different people's lives, this feels like:
* A 19-year-old in Jakarta is meeting up for potato with her friends tonight, and walks there along a shaded, well-lit route, and sees her campus hexagon slowly filling with tiny traces of good
* She shares a hexagon name in a WhatsApp chat so a friend can find the right gate in a maze of streets.
* Her aunt in São Paulo uses the free tier just to see which local market is cheapest this week and where the nearest community garden is.
* A municipal worker in Barcelona reads the tree-cover indicator by hexagon and uses What-if to test where to plant another 1,000 trees.
They’re all using the same map, for different reasons, without being asked to “join a platform”. It’s just infrastructure that quietly helps.
At that point, large numbers stop feeling too surprising. They simply reflect that many people have found the map useful and worth keeping around.
### Hexagons API
100k/month free
generous limits
// running costs
// static
// aim is to not operate at a loss
why an API will be useful
and for whom
public health researchers
environmental science departments
poverty / development economics
urban planning
psychology / loneliness studies
social policy programmes
climate modelling labs
student projects
Who uses the API:
researchers
students
NGOs
civic hackers
journalists
small startups
APIs are about embedding hexagons.world as a standard, not profit.
#### Usage tiers and pricing
generous API limits
100k/mo for free
- 100k API calls per month free
- public-interest use encouraged
- open-source, student, journalist, researcher, and civic projects treated kindly
- no surprise bills
- no lock-in
A student project, small civic app, or local journalist should be able to use the API without thinking about cost.
A company, institution, or high-volume service should pay enough that their usage cannot quietly become a burden on the commons.
Free
100k calls/month
public indicators
basic hex lookup
basic tiles / metadata
rate-limited bulk access
attribution required
You get a generous free tier. Once you go over it, you move into a fixed monthly band that is priced from the actual cost profile we already defined.
then you pay
#### Costs
As most of the map is deterministic and cache-heavy, the marginal cost per normal API call should be tiny.
raw costs
$0.00000030 / per api call
or, 0.30 / per million requests
practical loaded cost per API call:
$0.000002
1M calls
raw cost: $0.30/month
loaded cost: $2/month
10M calls
raw cost: $3/month
loaded cost: $20/month
100M calls
raw cost: $30/month
loaded cost: $200/month
```
monthly_api_cost =
fixed_api_cost
+ (api_calls × marginal_cost_per_call)
+ (heavy_jobs × marginal_cost_per_heavy_job)
+ support/admin_cost
```
The aim isn't to make the API expensive, but to make sure each extra layer of usage pays for itself.
#### Modelling
API profit should be modelled as a linear function of usage, where the fixed platform cost is already mostly covered by the public map infrastructure, and the marginal cost is mostly requests, bandwidth, storage reads, and any heavier indicator/statistical queries.
as a function
revenue
```
monthly_api_revenue =
paid_api_calls × price_per_call
+ paid_heavy_jobs × price_per_heavy_job
+ institutional/API plans
```
profit
```
monthly_api_profit =
monthly_api_revenue - monthly_api_cost```
```
### What if
Public What if should stay generous for most people.
Explorer — free
Can create unlimited drafts.
Can run a few small public scenarios per month on the shared pool.
Can fork public scenarios and share links.
No team features.
No guaranteed scheduling.
Good enough for students, curious locals, and journalists trying things out.
Builder — about $5/month
More monthly runs.
Can schedule recurring public scenarios.
Can compare multiple versions side by side.
Can save personal libraries and templates.
Gets slightly higher queue allowance, but still no “fast track.”
This tier is about reducing friction, not buying power.
Most users stay on Explorer, the free tier lets people play, learn, share scenarios, and build the habit of thinking in public indicators.
Paid adoption comes from regular users who hit recurring friction.
workloads:
choose compute mode:
— volunteer-first (slow, extremely cheap)
— datacentre-only (fast, guaranteed)
— mixed
billing:
datacentre compute charged at near-cost
volunteer jobs almost free
Workspace
Something light like $10/month for a shared public workspace with permissions, templates, shared scenario libraries, and comments. This is for small NGOs, student groups, local councils testing the waters, and media teams.
use cases:
shade impact, bus network changes, food waste, pollution checks, local scenarios
| Scenario | Civic monthly users | Student monthly users | Practitioner monthly users | Total monthly users |
| --------- | ------------------: | --------------------: | -------------------------: | ------------------: |
| Low | ~83k | ~85k | ~89k | ~258k |
| Middle | ~817k | ~781k | ~723k | ~2.32M |
| Ambitious | ~4.43M | ~4.83M | ~3.40M | ~12.67M |
#### Builders
Builder is the individual paid tier.
Builder is attached to an individual person, giving them more runs, saved templates, comparison views, recurring public scenarios, and a smoother personal workflow.
``Builder subscribers = monthly users × Builder conversion``
| Demographic | Builder conversion | Why |
| ------------- | -----------------: | ----------------------------------------------------------------------------------------------------------------------------------------- |
| Civic | 4% | Some regular local organisers and civic users want saved templates, recurring public scenarios, and comparison views, but many stay free. |
| Students | 3% | Lots of usage, but more price-sensitive. Some pay because they use it repeatedly for projects, organising, or coursework. |
| Practitioners | 8% | Strongest individual paid fit. They use it for reports, workshops, explanations, planning, and repeated public-facing scenarios. |
Civic users pay when they are regularly testing local ideas, comparing public scenarios, or keeping small campaign templates.
Some regular local organisers and civic users want saved templates, recurring public scenarios, and comparison views, but many stay free.
Students pay less often, but some pay when What if becomes part of coursework, campus organising, student journalism, or repeated project work.
Lots of usage, but more price-sensitive. Some pay because they use it repeatedly for projects, organising, or coursework.
Practitioners pay most often because What if helps them explain things to other people: in workshops, reports, grant applications, planning sessions, and public briefings.
Strongest individual paid fit. They use it for reports, workshops, explanations, planning, and repeated public-facing scenarios.
| Scenario | Civic Builders | Student Builders | Practitioner Builders | Total Builders |
| --------- | -------------: | ---------------: | --------------------: | -------------: |
| Low | ~3.3k | ~2.6k | ~7.1k | ~13k |
| Middle | ~32.7k | ~23.4k | ~57.8k | ~114k |
| Ambitious | ~177k | ~145k | ~272k | ~594k |
At ~$5/month:
| Scenario | Builder revenue / month |
| --------- | ----------------------: |
| Low | ~$65k |
| Middle | ~$570k |
| Ambitious | ~$2.97M |
#### Workspace
Workspaces are used for easily working with others.
One person, group, organisation, teacher, organiser, or practitioner pays for the room. The people inside might be Builders working on the workspace, or explorer (free) users with read-only access. Workspace is attached to a shared group space. One person or organisation pays for the workspace, and multiple people can join it.
| Demographic | Workspace-involved rate | Avg users per workspace | Why |
| ------------- | ----------------------: | ----------------------: | --------------------------------------------------------------------------------------------- |
| Civic | 8% | 6 | Local groups, campaigns, civic teams, neighbourhood projects. |
| Students | 6% | 8 | Student project groups, campus organisers, class teams. Bigger groups, lower payer intensity. |
| Practitioners | 12% | 6 | NGO teams, journalists, researchers, council-adjacent staff, public planning teams. |
- Civic workspaces are for local groups, neighbourhood projects, public campaigns, and civic teams.
- Student workspaces are for class projects, campus groups, student organisers, and research groups.
- Practitioner workspaces are for NGO teams, journalists, researchers, council-adjacent workers, public health people, planners, and people who need to explain a scenario with others.
``Workspace subscriptions = monthly users × Workspace participation rate ÷ average people per workspace``
We don't treat workspace members as extra individual paid users, since workspace participation is not the same as workspace payment. A share of regular users participate in group-shaped work, and those group-shaped users produce a smaller number of paid workspaces.
| Scenario | Civic Workspaces | Student Workspaces | Practitioner Workspaces | Total Workspaces |
| --------- | ---------------: | -----------------: | ----------------------: | ---------------: |
| Low | ~1.1k | ~640 | ~1.8k | ~3.5k |
| Middle | ~10.9k | ~5.9k | ~14.5k | ~31.3k |
| Ambitious | ~59k | ~36k | ~68k | ~163k |
At ~$10/month:
| Scenario | Workspace revenue / month |
| --------- | ------------------------: |
| Low | ~$35k |
| Middle | ~$313k |
| Ambitious | ~$1.63M |
#### Revenue
The simple read is that making things free creates culture, the builder assists enables individual use, and workspace assists collaboration.
Civic users pay when they are trying to keep local scenarios alive.
Students pay when the tool becomes part of a project.
Practitioners pay when the tool becomes part of their work.
| Scenario | Builder revenue | Workspace revenue | Total monthly revenue | Total annual revenue |
| --------- | --------------: | ----------------: | --------------------: | -------------------: |
| Low | ~$65k | ~$35k | ~$100k | ~$1.2M |
| Middle | ~$570k | ~$313k | ~$883k | ~$10.6M |
| Ambitious | ~$2.97M | ~$1.63M | ~$4.6M | ~$55M |
### What if? - Private
non-government organisations
companies
small companies
local governments
how many people and organisations might be interested?
- who do geospacial already
- who could be demographics who might be interested
| Segment | Users universe | Likely payer | Adoption style |
| ----------------------------- | -------------: | ----------------------------------------------------------- | --------------------------------- |
| Institutions | ~30M users | councils, universities, NGOs, public health, planning teams | slower adoption, higher retention |
| Operational local businesses | ~31M users | logistics, trades, agriculture, services, property | practical, ROI-driven |
| Retail & hospitality | ~27M users | chains, markets, hospitality groups, local retailers | lighter usage, local planning |
| Consultants / scenario people | ~12M users | consultants, analysts, grant writers, planners | high influence per user |
#### Costs
system administration
support staff
training
hosting costs
cloud providers typical pricing
#### Pricing
$20 for a workspace + $5 per active contributor
(viewers are free)
For researchers, civic hackers, NGO staff, and consultants using only public data.
Larger scenario scopes, more active jobs, better exports, better collaboration, more workspace features.
Still on public/shared compute unless they move into private.
At $5/user, a 10-person team costs $50/month. That feels like nothing for an organisation, and still meaningful for Peaceful Foundation at scale. More importantly, nobody thinks too hard about inviting the intern, the volunteer, the researcher, the local organiser, the comms person, or the student group. That matters because What if only becomes valuable when a bunch of different people are thinking together.
Honey
for heavier compute
honey is priced in a transparent structure
#### Uptake
from hexagons people section
| Scenario | Rounded active users |
| ----------------------- | -------------------: |
| Very conservative | ~450k |
| Middle | ~1.4M |
| Ambitious | ~3.9M |
| Global / infrastructure | ~10M |
| Region | Very conservative | Middle | Ambitious | Global / infrastructure |
| --------- | ----------------: | --------: | --------: | ----------------------: |
| Americas | 72k | 233k | 630k | 1.61M |
| Europe | 70k | 221k | 592k | 1.50M |
| Asia | 240k | 768k | 2.10M | 5.49M |
| Oceania | 4k | 12k | 32k | 81k |
| Africa | 41k | 132k | 354k | 903k |
| Arabia | 21k | 66k | 178k | 452k |
| **Total** | **448k** | **1.43M** | **3.89M** | **10.04M** |
speed of uptake within different regions
#### Revenue
#### Regions
Different regions also adopt through different institutional shapes.
This is important because What if? isn't only a modelling product. It's also a coordination layer between local knowledge, operational planning, and shared indicators. So the shape of adoption depends partly on how decisions already move through a region
| Region | Rough labour force | Plausible decision-maker share | Decision-maker universe |
| --------- | -----------------: | -----------------------------: | ----------------------: |
| Americas | ~520M | 1.5–4% | ~8M–21M |
| Europe | ~350M | 2–5% | ~7M–18M |
| Asia | ~2.15B | 1–4% | ~22M–86M |
| Oceania | ~18M | 2.5–5% | ~0.45M–0.9M |
| Africa | ~520M | 0.5–2.5% | ~3M–13M |
| Arabia | ~140M | 1.5–4.5% | ~2M–6M |
| **Total** | **~3.7B** | | **~42M–145M** |
Even this is still conservative in one sense: it only captures people directly making operational decisions themselves. In reality, many more people influence those decisions indirectly through reports, proposals, dashboards, grant applications, campus planning, logistics coordination, or local advocacy work.
So the true “influence surface” around place-based modelling is larger than the direct paid seat count. What if? spreads partly because one person running a scenario often affects many other people nearby.
| Region | Institutions | Operational local businesses | Retail & hospitality | Consultants |
| --------- | -----------: | ---------------------------: | -------------------: | ----------: |
| Americas | 4.6M | 4.6M | 3.9M | 2.3M |
| Europe | 6.0M | 2.7M | 2.7M | 2.0M |
| Asia | 14.4M | 20.2M | 17.3M | 5.8M |
| Oceania | 0.3M | 0.2M | 0.1M | 0.1M |
| Africa | 3.0M | 1.7M | 2.6M | 1.3M |
| Arabia | 1.9M | 1.1M | 0.9M | 0.4M |
| **Total** | **30.3M** | **30.4M** | **27.4M** | **11.9M** |
To avoid overcomplicating things, we'll use 100 million people as the working global decision-maker universe for What if – Private, split roughly into 30M institutional users, 31M operational local-business users, 27M retail/hospitality users, and 12M consultants or scenario people.
### Additional costs running paid offerings
Developers
web
android
ios
secure software
engineers
software
geospacial
data
//
legal and compliance
lawyer
paralegals
administration staff
billing and financial
project managers
//
Domain experts
physics
statistics
//
2. Missing cost buckets (important gaps)
You’ve covered compute and storage very well. What’s missing are human and operational costs, which is where most “cheap” platforms fail.
A. People costs (not optional)
Even with volunteers, you will need paid continuity roles.
Minimum realistic baseline (very lean):
Role Why Rough cost (monthly, USD)
1 core engineer keeps infra alive $6k–10k
0.5 geospatial/data engineer indicators, pipelines $3k–5k
0.5 product/generalist maps, APIs, What-If UX $3k–5k
0.5 ops/admin billing, emails, orgs $2k–3k
Bare minimum $14k–23k / month
You can defer this early, but not indefinitely once:
universities
NGOs
enterprises
start depending on the system.
👉 This is the real floor your paid offerings must cover.
B. Trust, governance, and liability
Not huge costs, but must be named:
Legal review (privacy, disclaimers, liability)
Data governance docs (especially for What-If)
Occasional insurance (orgs will ask)
Ballpark:
$3k–10k/year initially
grows slowly, not exponentially
C. Maps-specific hidden costs
Not infra-heavy, but support-heavy.
3. Who actually pays (this is the key missing clarity)
3.5 What-If Private / Enterprise
This is slower, but very high leverage.
Even:
5 enterprise clients
at $120k/year
= $600k/year
You do not need many.
This tier exists to:
fund long-term stewardship
subsidise free global access
absorb legal/ops complexity
4. Overall financial picture (realistic, not hype)
Monthly costs (steady state, lean team)
Infra (all public layers): $200–1,000
Tools + APIs: $200–500
Staff (bare minimum): $15k–20k
Misc/legal/admin: $500–1k
👉 Real operating floor: ~$18k–22k/month
Revenue pathways stacked (important)
Layer Purpose Revenue
Public hexagons legitimacy, adoption $0
Maps $1 fairness, shared compute $1k–5k
API embedding $0.5k–3k
What-If core funding $7k–30k
Enterprise long-term sustainability $50k+/mo (eventual)
This is healthy because:
no single layer has to do everything
failure of one doesn’t kill the system
public good stays public
5. Biggest remaining gaps to fill (next steps)
These are not numbers — they’re clarity gaps:
Who maintains indicators long-term?
named roles
escalation path
What happens when volunteers disagree with results?
governance cost, not compute
What support promise does $30/month actually include?
email SLA?
response times?
What is explicitly not supported?
critical to avoid burnout
Bottom line (plain and honest)
Your infra cost model is unusually strong and realistic
Maps is a trust layer, not the main revenue engine
What if public and private are what make everything else possible
You only need dozens, not thousands, of paying institutions
It allows gifting months, which:
pulls in people without cards
builds social flow
breaks the “subscription = obligation” feeling
It anchors the idea that:
“This is real infrastructure, not an extractive platform”
That’s not cheapness — that’s intentional lightness.