This week. Cost / Interest against the project's target (target_cp_interest; Submitted where a project is sold on interest): on target = 100, 50% cheaper = 150 (the most), twice as dear = 0. Spam rate ≤ 10% and UNQ rate ≤ 35% are maximums: at or under = 100, twice the max = 0.

A week under 5 Interest or 20 leads reads the weeks before it (up to 4) until it has enough; the hover says how many weeks were read.

Booking — campaign to date: CPA 60 (spend vs booking value × the project's target CPA) · Cost / SVD 40 (vs target_cp_svd), same scale as C/I. Not rated in the first 25% of the campaign; after that, spend with no booking or no visit = 0.

Non Booking — Spending Pace 100: spend vs 20% of the MO spread over the campaign days; 60–100% of the allowance = 100, less falls toward 0, more falls to 0 at twice it. Its funnel targets are a reference for effectiveness only, never its contract KPI.

Coverage 50 — the share of the owner's rated projects that got at least one Meta or Google change this week, any day. Results impact 50 — of the projects changed last week, how much C/I · Spam · UNQ (60/20/20) improved this week; no improvement counts 0.

Improvements — C/I · Spam · UNQ (60/20/20) this week vs the project's previous 4 weeks: the same = 50, 25% better = 100, 25% worse = 0. A change counts toward a project only once its campaign is mapped to the project (NPS code / Ads Mapping).

A project with no spend this week, no Expectation or Idle is greyed and not quantified. A part with nothing to rate prints a dash and is left out, so it pushes the score neither up nor down. Agency, freelance and Organic owners are not ranked. If the change log was read short, the page refuses to rank at all.

A week is computed live until 7 days after it ends — unqualified leads need days of call attempts to settle — then locked by a manager's view: it is saved and never moves again. A manager can re-lock a week whose data arrived late; a week locked after the fact says "backfilled".

An owner's projects are read as they were that week (the Project Index change history; a handover counts from the week after it is made), so someone who has since left still appears in their weeks — greyed, ranked as they were, never the week's Bottom — and nobody else's weight moves.

Per PIC, blended: total spend ÷ total booking value × 100, each project measured over its own campaign window. Target is the project's own — max usage ÷ booking target value; 0.9% where none is set — and a PIC's target is the booking-value-weighted blend of their projects'.

No bookings prints ≈ %: the CPA if the first booking lands at the project's average product price. Projects are sorted by distance to their own target.

A contract whose KPI type is Non Booking (set in the Project Index; unset = read from Product inventory) has no CPA. Its KPI is spend ≤ 20% of the MO value, printed as % of MO, with the same ON TARGET / OVER reading. Any funnel target it carries is a reference for effectiveness only.

Campaign to date, against the target pro-rata by campaign days. Booking contracts: booking value (units where no value target is set) and site visits done. Non Booking: the one volume the contract is sold on — SVD for a visit product, Submitted / Interest, else leads. On pace = 100, 1.5× ahead = 150 (the most), nothing = 0.

Not rated in the first 25% of a Booking campaign (10% for Non Booking) — bookings need time.

SVS ÷ Interest and SVD ÷ SVS, each against the project's own target ratio (target_svs ÷ target_interest, target_svd ÷ target_svs): on target = 100, up to 150. Read on the leads of the 8 weeks that ended 2 weeks before the week's end, so they have had time to visit; at least 10 Interest / 5 SVS.

Booking Recap — of the contract's counted bookings, the share with a real booking date in the CSS sheet as it stands today (a locked week keeps what it read): 95% or more = 100, 50% or less = 0. No booking yet: left out.

Improvements — SVS/INT and SVD/SVS against the 8 weeks before that window: the same = 50, 25% better = 100, 25% worse = 0.

A contract that is Idle, has no Expectation, has not started or has no target is greyed and not quantified. A part with nothing to read prints a dash and is left out, so it pushes the score neither up nor down. A CSS Freelance is listed, never ranked.

The same rule as PMKT: a week is computed live until 7 days after it ends, then locked by a manager's view and never moves again. Automatic locking starts when the CSS scorecard goes live — until then every week is computed from today's data.

The CSS name in the Project Index seven days before the week's end (its change history), so a handover counts from the week after it is made. A contract naming two CSS counts for each of them.

The contract's label, shared with PMKT and set on the PMKT Bandwidth board or PMKT Monitoring — one label per contract, so both teams weigh the same contract the same way.

Ad spend and cost per lead — PMKT's. Calls, chats and response times — Tele's, which gets its own scorecard. Ending-soon and renewals stay on Bandwidth.

Weights, windows and floors are a starting point (formula c1). Any of them can change; until locking goes on, every past week is recomputed with the new values.