A useful weekly intelligence brief is a decision queue, not a news digest. It contains only verified movements that may affect a named buyer, deal, product boundary, or commercial choice. Every item shows the source, observed fact, bounded implication, unknowns, owner, and response deadline.
What this guide covers
- the editorial threshold for including an item;
- a one-page brief structure;
- severity and confidence without fake precision;
- a copyable template and meeting routine.
Exclude most things
The brief should be shorter than the underlying review ledger. Exclude:
- unreviewed alerts;
- visual or navigation changes without commercial meaning;
- routine releases that do not alter the buyer job;
- rumors without a primary record;
- repeated observations already assigned;
- true facts with no decision relevance this week.
Official pages answer different questions. The Cursor changelog supports release observations, Cursor status supports disclosed incidents, and GitHub Copilot plans supports current offer boundaries. Do not blend them into a general narrative about company momentum.
Give every item a fixed anatomy
Use six blocks:
- Verified movement: one sentence containing the exact change and scope.
- Why it matters now: named segment, decision, or active assumption.
- Evidence: primary URL, observation date, and comparable before state.
- Interpretation: plausible implication, explicitly labeled.
- Unknowns: what the source does not establish.
- Action: owner, review, due date, and no-action option.
If an item cannot fill those fields, keep it in the research queue.
Use severity and confidence for different things
Severity describes possible decision impact. Confidence describes the evidence record. A high-confidence change can be low severity. A potentially severe rumor can have insufficient confidence for circulation.
| Evidence confidence | Decision severity | Brief treatment |
|---|---|---|
| high | high | lead item, named owner and deadline |
| high | low | context line or ledger only |
| low | high | research request; do not present as fact |
| low | low | reject |
Avoid an unexplained 8.7/10. Use defined states and a reason.
Keep one stable section for “no action”
Record rejected and reviewed-no-action items in a compact appendix. This protects institutional memory and shows that silence was a decision rather than missed monitoring.
Useful reason codes:
- outside target segment;
- no material field changed;
- baseline only;
- incomparable or source fault;
- already covered by prior decision;
- insufficient corroboration;
- true but not currently actionable.
Copyable weekly brief
# Weekly market-change brief — [week ending]
**Owner:**
**Coverage:** [companies / market / decisions]
**Evidence cutoff:** [date/time]
## Decisions required
### 1. [Movement in plain language]
**Observed fact:**
**Primary evidence:**
**Before / after:**
**Affected buyer, deal, or assumption:**
**Interpretation:**
**Unknowns / alternatives:**
**Decision requested:**
**Owner / due:**
## Context worth retaining
- [verified movement, why it may matter later, source]
## Research needed
- [question, evidence gap, owner, due]
## Reviewed — no action
| Item | Reason code | Revisit condition |
|---|---|---|
| | | |
## Previous decisions closed this week
| Evidence | Decision | Outcome | Follow-up |
|---|---|---|---|
| | | | |Run a 20-minute decision meeting
Do not read the brief aloud. Participants review it asynchronously. In the meeting:
- resolve only the “decisions required” section;
- assign missing evidence rather than debate unsupported stories;
- record no action as a valid outcome;
- close prior actions with an observed result;
- reject any new topic that lacks a decision owner.
The brief owner maintains evidence quality, not organizational authority. Product, pricing, sales, security, or the founder owns the actual response.
Do not turn the brief into pricing mimicry
Stripe's pricing and packaging guide shows that price interacts with value metric and tier design. If a competitor moves price, the brief should surface the affected workload and segment—not request an automatic match.
Use the EXVIV company directory and market map for context, Evidence Compare for public before/after inspection, and the monitoring guide for source rules. The startup operating system keeps the weekly process bounded.
Related EXVIV research
Sources and further reading
Method note
This template routes public-source evidence. It does not convert an observation into proof of motive, adoption, revenue, or private roadmap. Preserve the underlying record and current source state.