This week
The ranked plan: which open change to do first, what it is projected to be worth, and how much of that projection is calibrated against shipped outcomes.
This week is the plan. It ranks your open changes so the first row is the one to do first, with a projected lift and an effort beside each, and it is honest about which of those projections have ever been checked against reality.
What you'll learn
- How the ranking is built and why the order is more trustworthy than the sizes
- What calibrated means on a row
- How a row turns into a shipped, proven change
The rows
Each row is an open change, and each carries:
- what it involves, in a sentence
- the projected lift, in percentage points, or "no projection" where there is none
- the effort it takes
- whether the projection is calibrated against shipped outcomes, or is an estimate
- the prompt it acts on, where it acts on one
The row is the change, so opening it takes you into that change's Review in the Changes lens.
Projected, not promised
The panel says this out loud, and it is the most important sentence in the lens: each figure is the modelled lift for that action type, some of those models are calibrated against outcomes the platform has actually shipped and measured, and none of them carries a measured error band yet.
So treat the order as more reliable than the sizes. The ranking is comparing like with like. The absolute numbers are estimates, and the panel tells you how many of the rows on screen are calibrated rather than estimated.
Do not put a projected lift in front of a client as a forecast. It is a prioritisation aid. The citation-rate forecast in Intelligence is a different quantity with its own calibration record, and the two are deliberately not mixed.
A row with no projection shows "no projection" rather than a zero, and is not badged as projected, because a missing estimate is not an estimate of nothing.
The loop
- Read the order. Take the top row, or the top row you can actually resource this week.
- Approve it. Approval is what starts the clock and captures the control the outcome will be scored against. Approving through the ledger is what makes the result measurable later.
- Ship it.
- Let it be observed. The platform fetches the page itself to confirm the change is live.
- Let it be confirmed. Once there are enough checks across enough surfaces, the movement is scored against the control.
Skipping step 2 is the single most common reason a change ends up marked not measurable.
Try this in Outercite
Ask the assistant "Where should I focus first, and what is it worth?" Then open the top row to read what it involves before you commit the week to it.
Related
Ship and prove a change
The workflow, step by step.
Changes
The ledger every row here opens into.
Forecasts and ranges
The other kind of projection, and how it is calibrated.
Last updated on
Readiness
The grade on your own site: whether engines can reach it, parse it, trust it, and whether they actually cite it. Four levels and four named tiers.
The last run
What the agents did overnight: the steps they actually took, what they found, and what they wrote, with each part labelled by how solid it is.