1. What a report actually is

A report is one claim, made checkable: this institution, measured against these people’s needs, moved this much need in this direction.

You pick a subject (an institution, a project, a policy, an agent). You pick a denominator (the people whose needs are the measuring stick). Then you score five rows, one per level of human need, on two numbers each. The tool does the arithmetic and puts a dot on a chart.

What makes it a report rather than an opinion is that every number carries a source and every assumption is written down. Someone who disagrees opens your link, changes the number they dispute, and gets a new link showing exactly what they changed. Disagreement becomes a visible difference in arithmetic instead of an argument about adjectives.

You do not need an API key. The builder has an optional panel that asks an AI to draft a report using a key you supply and pay for. It is optional. Everything in this guide is done by typing numbers into the table. If the AI panel is in your way, leave it closed.

2. The denominator: who you are measuring for

This is the one field beginners skip, and skipping it makes every score meaningless. The denominator names whose needs count in this report.

“A hospital” is not scorable. “A hospital, measured against the residents of the county it serves” is. Change the denominator to “the hospital’s shareholders” and every score in the table legitimately changes. There is no neutral choice here, which is exactly why the tool makes you write it down: your reader can see the frame you chose and argue with it.

Good denominators name a group of people: “BC taxpayers”, “tenants in the building”, “residents facing food insecurity in this neighbourhood”. Bad denominators name an abstraction: “society”, “the economy”, “stakeholders”.

3. F and A, the only two numbers you score

Each of the five need levels gets two scores, both running from −10 to +10 in half-steps. Leave a score at 0 when the record is silent; 0 is an honest answer and the tool never treats it as a gap.

Score	Question it answers	Positive when	Negative when
F filtering	Does the subject take in hardship that would otherwise land on the denominator?	The denominator’s hardship goes down because the subject absorbed it.	The subject creates hardship or passes it through: fees, rate hikes, displacement, known harms left unaddressed.
A amplification	Does the subject emit fulfillment beyond what it received from the denominator?	Something genuinely new flows out: capability, connection, recognition, opportunity.	The subject consumes more fulfillment than it returns.

The mistake almost everyone makes first. Grants, tax exemptions, donated goods, public funding, and volunteer hours are things the subject received. They are intake, not output. Net them off before you score A positive. An organisation that received a great deal and passed a little of it along is not amplifying much.

The second most common mistake: an institution quietly absorbing criticism, a scandal, or budget pressure has not filtered anything, because the denominator’s hardship did not change. Unmet need scores 0 or negative, never positive absorption.

Which level does a flow belong to? Book each flow where the need actually lives:

Level	What gets booked here
Physiological	Housing, food, water, health, heat.
Safety	Affordability, fees, debt, rates, physical safety, job security.
Belonging	Community, inclusion, connection, participation.
Esteem	Recognition and dignity, but only where recognition is the flow being measured.
Actualization	Capability, learning, research, the ability to become more.

Every nonzero score wants a source. One line is enough: a link, a statute, a filing, a report name. The tool does not block you if you skip it, but it flags the level as unsourced on the reading, and anyone you share with sees the flag. That is deliberate: the framework’s standing invitation is “substitute my number with a source”.

4. Weights, and why they start at 5-4-3-2-1

The five levels are not equally urgent. Someone who cannot eat is in a different situation from someone who cannot self-actualise, so Physiological starts at weight 5 and Actualization at 1. That ordering is the framework’s default, not a law.

Every weight is yours to change, and the report records what you chose, so a reader can see that you weighted Belonging at 5 and decide whether they agree. If you have no reason to change them, leave them.

5. Worked example: a food bank (simple)

Riverside Community Food Bank

ILLUSTRATIVE. Every figure below is invented to teach the method. This is not a scored real organisation, and the report says so in its own assumptions.

Denominator: residents of the Riverside neighbourhood facing food insecurity.

Level	w	F	A	Why
Physiological	5	6	2	Its core act is absorbing hunger that would otherwise land on residents, so F is high. A is small: the surplus-produce share-out puts a little food out beyond what anyone came asking for.
Safety	4	2	1	Predictable weekly access removes some of the precarity of not knowing where the next meal comes from.
Belonging	3	1	3	The strongest amplification here: the Thursday community meal creates connection that did not exist and was not received from the denominator.
Esteem	2	0	1	Volunteers are recruited from among the people served, which returns recognition to the denominator.
Actualization	1	0	0	The record says nothing, so this stays 0 rather than being guessed.

Result: F_agg = 27.3, A_agg = 16.7, W = +46.7. A Transmuter: it both absorbs deprivation and emits fulfillment, weighted heavily toward absorbing.

Notice what the scores are not doing. Donated food and volunteer hours never appear as amplification, because the food bank received them. The community meal is booked at Belonging, not Physiological, because the flow being measured there is company, not calories.

Open this example in the builder

Change one number and watch the dot move. Set Physiological F to −6 and the same organisation becomes an Extractor.

6. Worked example: a provincial power regime (real)

Manitoba datacenter-power regime, present needs

A real, sourced Foundation report, published in the gallery. Every score carries a citation.

Denominator: Manitoba residents — households, existing ratepayers, and affected communities including Indigenous Nations.

Level	w	F	A	Why
Physiological	5	4	0	Hyperscale load the system could not firmly serve was refused rather than passed to households.
Safety	4	3	1.5	A default +100% levy on datacenter and crypto rate classes protects rates; 500 MW of expiring export contracts were repatriated for domestic use.
Belonging	3	1	0	Environmental grounds named in the rejection; the rate review carries Indigenous interveners.
Esteem	2	0	0	Silent in the record.
Actualization	1	0	0	Silent in the record.

Result: F_agg = 23.3, A_agg = 4.0, W = +29.0. A Transmuter that is almost entirely a filter: it prevented harm rather than generating new fulfillment.

This is the shape a regulatory decision usually takes. Refusing to make things worse is real moral work and scores positively on F, but it is not the same as making things better, and A stays near zero to say so. Two of the five levels are 0 because the record genuinely says nothing about them.

Open this example in the builder

7. τ and ρ, the two dials you can ignore at first

Both live under Advanced and both default to 1.0, which is the neutral setting. You can write a complete report without touching either.

τ (tau), the asymmetry coefficient. It sets how the tool trades filtering against amplification. τ < 1 prioritizes fulfillment generation. τ > 1 emphasizes cycle-breaking. τ = 1 weights them equally. It rotates the contours on the chart without changing their spacing.

ρ (rho), the power coefficient. It raises the bar for a powerful agent. ρ = 1 is parity, with the neutral origin at the centre. As ρ rises, the M = 0 line slides up the moral axis, so identical flows score lower — the argument being that more power should mean more is expected.

If you change either one, say so in your assumptions. A reader comparing two reports needs to know they were scored on the same dials.

8. Saving, sharing, and being disagreed with

Your report is saved from the first keystroke. The whole thing lives in the address bar: reload the page, bookmark it, or copy the URL at any moment and nothing is lost. The browser also keeps its own copy, so closing the tab and coming back later offers your work back.

Action	What it does
Bookmark the page	Saves the report as it stands. No button needed.
Copy link	Puts the same URL on your clipboard to hand to someone else.
Download .json	Saves a file you own. Safest for long reports and for archiving.
Open .json	Brings a downloaded report back into the builder, on any machine.
Submit to gallery	Shows you the submission text to send to the Foundation. Publication is by human review.

When someone opens your link, your report is pinned as their baseline. The moment they change anything, the chart shows two dots — yours as a ghost, theirs solid — and a panel lists every field they changed as theirs → yours. Their Copy my version mints a new link that still carries yours inside it.

That is the point of the whole tool. You are not asking anyone to accept your scores. You are handing them arithmetic they can edit, with your sources attached, and asking them to show their work when they disagree.