How the decision figures are made

Short version: decisions counted when they are made, timed from receipt, approved over approved plus refused, and no council published until its data passes our checks.

The cohort

A window holds the decisions a council made in it, by decision date. The headline window is the twelve months to the last complete month. Counting by decision date is how the official statistics work, and it means a slow application does not go missing from the figures while it is still undecided.

Timing

Days from the date the council validated the application to the date it decided it, for applications approved or refused in the window. That is the clock councils report against and the official statistics use, so our figures sit beside theirs. Each council's page also gives the count from receipt, which is what the applicant actually waited: validation typically adds a few days. A negative duration, or one over ten years, is a data error and is left out. Withdrawn applications and undecided applications are not timed.

Where our figures still differ from a council's own, two things explain it. Councils count an agreed extension of time as decided on time; we treat the calendar as the calendar. And ours are live and rolling where official statistics are quarterly and annual. Years before the live window come from our archive, which carries no validation date, so those years are timed from receipt and say so.

Rates

Approval rate is approvals divided by approvals plus refusals. Withdrawals, and decisions whose outcome the register does not state, are counted in the decided total but not in the rate.

Kinds of application

We sort applications by procedure from the register's own application type: full, householder, outline, reserved matters, prior approval, listed building consent, advertisement consent, tree works, discharge of conditions, amendments and lawful development certificates. A kind is shown for a council when it has 25 timed decisions of that kind in the window.

Decision route

Where the register says how a decision was made, it is mapped to delegated (decided by officers) or committee. Unknown stays unknown and is not counted in the share.

Appeals

Appeals come from the Planning Inspectorate's casework database for England, joined to the council and counted by the Inspectorate's decision date.

Publication floor and exclusions

A council is published when it has at least 25 timed decisions in the window. It is not published when its register's decision data fails our consistency checks: a year with hundreds of decisions and almost no refusals against a normal year before it, or a live refusal rate that disagrees several times over with the same council's own history. The reason is shown on the council's page. Every council is re-tested nightly.

Ranks

Ranks are among published councils only, and say so. The speed rank orders councils by median days, ties broken by the share within eight weeks. Change is the movement against the twelve months before.

Questions about a figure: contact us. The underlying table is available through the API.