Methodology
What an efficiency rating means, where it comes from, and what it cannot tell you.
Current content is example data. Every project on the site is a fictional placeholder labeled "Example data - not sourced". The process below describes how real entries will be handled.
What the rating shows
The efficiency rating is the share of each $1 that reaches the organization's actual programs, as opposed to overhead, advertising and fundraising, and executive pay. A rating of 90¢ means roughly 90 cents of each dollar goes to the work and 10 cents to everything else.
We rely on existing evaluators
We do not do our own audits. Each real project will cite the evaluator its numbers come from, and link to it.
- Charity Navigator: financial health and accountability ratings built from public tax filings.
- Candid / GuideStar: nonprofit profiles and transparency data, including Form 990 financials.
- GiveWell: cost-effectiveness research, which looks at outcomes per dollar rather than expense ratios.
Where an evaluator gives no figure, we show no rating rather than estimate one.
Limits to keep in mind
- Efficiency is not impact. A lean organization can pursue an approach that does not work, and a costly one can do excellent work.
- Small organizations may be too new or small to be rated, and ratings can lag the latest filings.
- Categories of "overhead" are defined differently by different evaluators.
- We do not judge which charities are good. We show data and let you choose.
Letter grades
Each project's efficiency rating maps to an A-F grade. Every grade cites the data source of the underlying rating; example entries say "Example data - not sourced".
| Grade | Efficiency (¢ of each $1 reaching the work) |
|---|---|
| A | 90 or higher |
| B | 80 to 89 |
| C | 70 to 79 |
| D | 60 to 69 |
| F | below 60 |
A grade is a summary of an expense ratio, not of impact. The thresholds are defined in one place in the code (src/lib/rating.ts).
Not yet rated
Where no evaluator gives a figure, a project shows "Not yet rated" instead of a letter or an estimate. It sorts after all rated projects and the calculators show no estimate for it, because inventing an efficiency would be misleading.
Example data
All current projects are fictional. Their rating source reads "Example data - not sourced" and points here. Raised amounts and donor counts carry an "as of" date and, for real projects, will each need a dated source. Example projects cannot accept donations.
Goal and ongoing projects
Goal projects have a finish line: they show money raised against the goal, a percentage (which can pass 100%), and a "Funded" stamp once the goal is met. Ongoing projects, such as feeding or monitoring programs, have no finish line, so they have no progress bar. They show what a monthly gift does (for example "$25/month feeds 1.9 colony cats") and a count of monthly donors.
The dollar calculators
The site-wide calculator and the per-project calculators use one formula: units funded = amount × efficiency ÷ unit cost. For example, $500 at 94% efficiency and $2,000 per acre funds 0.2 acres. The minimum amount is $20; any amount above that works. Goal projects are shown as a one-time gift, with the amount also expressed as a share of the goal. Ongoing projects use a monthly unit cost and are shown as monthly gifts ("$25/month feeds 1.9 colony cats", plus the yearly total) in their own section. Unrated projects get no estimate. Unit costs are round placeholders on example data. This is an illustration, not a guarantee.
How donating works
Unconventional Impacts never handles money. The Donate button on a real project opens a donation form hosted by Every.org, which sends the gift straight to the charity, with one-time and monthly options. The Every.org script loads only when you click Donate. If a project has no Every.org page, the button goes to the charity's own donation page. We never see your payment details. See Every.org's own terms for how they handle fees.
Neglectedness
Each category and project shows a neglectedness score from 1 to 10. A 10 means the cause gets very little funding compared with mainstream causes, so a gift is less likely to be one of many covering the same need. A 1 means it is already well funded. The labels are: 1-3 "Well funded", 4-6 "Somewhat overlooked", 7-8 "Overlooked" and 9-10 "Highly neglected".
The current scores are illustrative example data and are not measured. Real scores will compare funding per person helped, or per unit of need, with mainstream comparators, using sources such as Giving USA sector totals, Candid and IRS Form 990 data, and foundation grant databases. Each score will carry a date and a source, and be reviewed on a schedule. Neglectedness is not quality: a cause can be overlooked for good reasons, so read it next to the efficiency rating.
Charts
The donut, the fund-use bars and the impact charts are drawn from the same data as the numbers beside them, and each has a text table or list. All sample values are fictional.
Trending and newly added
The home page lists projects by a trending score and by the date they were added. Both are example values today. Real trending would come from aggregate Every.org totals, with no personal data.
Recent donor activity
Lines such as "12 people gave this week" are example data and say so. A real version would use aggregate counts only, never names, and would show the date they were last updated.
Funding progress
Progress bars show amount raised against the funding goal, with the date of the figure. Current figures are example data.
Sorting
Projects are sorted by efficiency rating, highest first, everywhere on the site. Ties are broken alphabetically; projects not yet rated come last.