The recorded baseline, the new signal, its context and any counter-evidence.
What changes in the earlier judgement, what stays and what now needs review.
Anything beyond what the available evidence can support.
A signal matters when it changes an assumption behind the plan.
New information should not reopen a strategic decision merely because it is recent. It becomes material when it challenges an assumption the plan depends on, changes the confidence behind an earlier rating, or alters the role, ownership or dependency expected to carry the work.
A key person resigns. Survey results change in one unit. A milestone slips for the second time. Each event has a plausible explanation. Together, or in the right context, they may show that an assumption behind the plan no longer holds.
That is the point to compare the evidence with the earlier judgement instead of continuing with the plan because it was already agreed.
Change the judgement without rewriting history.
Atlas preserves the earlier forecast, then tests the new signal against its source, context and counter-evidence. Only the parts of the judgement affected by the evidence should move.
Keep the baseline
Retain the original judgement, evidence, confidence, owner and review window. A reforecast needs a visible point of comparison.
Test the new signal
Ask what the evidence directly shows, which earlier assumption it challenges and what credible counter-evidence remains.
Update only what changed
Revise the affected rating and recommendation. Leave conclusions untouched where the new evidence has no bearing.
Set the next evaluation
Record who acts, what outcome will test the revised call and when the evidence will be reviewed again.
For each new signal, write what it challenges, what counter-evidence exists and which earlier conclusion it leaves untouched. If everything changes at once, the new evidence may be carrying too much weight.
Old conclusions can absorb new information.
Existing conclusions shape how new signals are read. One visible event receives too much weight, a quieter pattern gets dismissed, and the quickest explanation often centres on a person rather than the plan.
The test is specific. Which earlier assumption does the evidence challenge? What credible counter-evidence remains? Would the new information change a rating, recommendation, owner or review date? If the answer to all three is no, record the signal and keep watching. If the decision case changes, reforecast.
The expensive mistake is not caution. It is continuing to decide from an outdated view when the evidence needed to revisit the strategic decision already exists.
A reforecast changes the judgement, not the history.
Atlas compares the new evidence with the plan, the earlier judgement and any counter-evidence. The previous view stays in the record, so the client can see which assumption moved, what still holds and why the recommendation changed.
A reforecast is warranted when the new signal changes the decision case. Where the signal is too thin to support a revised judgement, Atlas marks the view blind and names the evidence needed next. This keeps missing evidence separate from positive evidence.