How this works
Method
This site is a record of what the Mandate asked for, what has already happened, and what is still to come. The record is neutral: every status is backed by a linked source, and we never assign one without it. The point of view is not neutral. We think what has already been carried out is the best evidence that the rest is meant seriously, and the site is built to show that.
What counts as a recommendation
A recommendation is a single, checkable ask made by the Mandate: a concrete action (rescind, abolish, create, reinstate, transfer, appoint, repeal, withdraw, issue) on a named object (an order, statute, office, program, treaty, rule). Principles without a mechanism are not included. Every recommendation carries the document's own words, verbatim, with the printed page number. When the same ask appears in more than one chapter, it is one recommendation with several citations.
Tiers
Every recommendation has a significance tier, fixed when it was extracted and never changed by what later happened. Major would change the structure of government or national policy in a way a general reader would recognize. Notable is a specific, publicly traceable change to a named rule, program, office, order, or enforcement policy. Minor covers internal staffing, management practice, advisory posture, and asks that Congress merely "consider".
How much of this we have checked
We research in tier order — major first, then notable, then minor — because that is what tier is for: it ranks how much of the Mandate's own program an ask carries. That means that at any moment part of this site has been read closely and part has not, and we would rather say which than let you assume.
Of 2,245 recommendations, we have researched 2,245. The major and notable and minor tier s are complete.
So a recommendation nobody has reached yet is stamped Not yet checked rather than Not started. The difference is the whole point: not started is a finding — we looked, and nothing in the public record answers this ask. Not yet checked is not a finding about anything. It means the sweep has not got there.
One thing this cannot tell you: whether an ask has no public instrument because nothing was done, or because nothing about it is ever written down. Plenty of the smallest asks are about internal staffing and management practice, where there may be no document either way. Where we have looked and found nothing, the recommendation's own page records what we searched.
Status
A recommendation's status is the latest dated event in its history.
| Status | Meaning |
|---|---|
| Done | The action was taken and is in effect. |
| Partially done | Part of the action was taken, or a narrower version of it. |
| Attempted | The action was formally initiated (proposed rule, bill introduced, order signed) but is not in effect. |
| Blocked | Action was taken but is currently prevented from taking effect, typically by a court. |
| Reversed | The action was achieved and later undone. |
| Unverified | A status was claimed without evidence; not counted as anything else until sourced. |
| Not started | No action found. The starting point for every recommendation. |
There is no score
This site does not reduce the Mandate to a percentage. A roll-up is a count of recommendations by status — "12
done · 3 blocked · 40 not started" — for a chapter, a section, or the whole document. It is always computed from
the record, never hand-authored, and never collapsed into a single number. A percentage would grade the
administration against the Mandate's own checklist; that is not what we are doing here. The reasoning is
recorded in docs/adr/0001-no-score.md.
What we set aside, and why you can see it
Two kinds of recommendation get no status. Contested ones come from the two chapters that argue with themselves. Already true ones were satisfied before 20 January 2025, with nothing happening since — the vaccination rule this administration was urged to stop enforcing had already been withdrawn in 2023, and a status measuring what this administration did would read "not started", which is true and useless.
Neither is dropped. Both are counted in their own labeled bucket at the end of every roll-up, and both are
left out of Still to come, because neither is coming. We say this here because the failure we most want to
avoid is the one that got the old scorecard retired: a judgment call presented as a fact by being invisible.
Adding a third bucket takes a written decision, recorded in
docs/adr/0003-status-measures-action.md.
When we change a status
Sometimes a later pass finds that an earlier reading was wrong. When that happens we replace the old status event rather than adding a new one after it, so a recommendation's history stays a record of what happened in the world and not a record of what we thought about it. The change itself — in both directions, dated, with the reasoning — goes on the corrections page.
Still to come
The Still to come list is every recommendation that is not started, reversed or blocked — nothing of it in effect — ordered by how big it is and how easily it could happen. It is derived entirely from the record. Each recommendation is sorted by feasibility: executive action if nothing in the Mandate's own text says it needs Congress; needs Congress if it does; blocked in court if it was tried and a court stopped it. A reversed recommendation was done and then undone, so it is coming again. A recommendation that was only attempted is not on the list: it is in motion, not ahead. A blocked recommendation appears in both the record of what happened and the list of what is still to come — it is proof of intent and a live candidate for another attempt.
We do not predict. Nothing on this site says what will happen next or when, and nothing in the list is hand-picked. The ordering is a sort, not a forecast.
Evidence
No status other than "not started" is assigned without at least one dated, linked source. Primary evidence is the official record: executive orders, the Federal Register, court dockets, agency releases, enacted law. Secondary evidence is reporting from wire services and major newspapers, and is shown as such. Sources must come from a fixed allowlist; anything else is noted as a lead, not evidence. Each source records who acted: executive, Congress, the judiciary, a state, or another actor.
Contested chapters
Two chapters — the Export–Import Bank and Trade — print opposing essays side by side. Recommendations from those chapters record which essay they belong to. Because the Mandate does not speak with one voice there, we do not treat either essay as its ask: we assign contested recommendations no status, and we leave them out of Still to come, because we will not forecast an ask the document argues against itself. They are still counted. A roll-up that covers them shows them in a bucket of their own — "53 contested" — rather than quietly shrinking, and their chapter pages show both positions.
Source
Mandate for Leadership: The Conservative Promise, The Heritage Foundation, 2023. 920 pages, 30 chapters. Page numbers on this site are the printed page numbers in the book.