Facet
DeutschInstall

FACET ImpactComing soon

Can we switch it off?

Your agents ship faster than anyone can keep track. When someone asks what a feature is worth, what depends on it or what replacing it would take, Impact answers from your code, records and tests, and marks which part of the answer is observed, inferred or still unknown.

catalog/import/product-importexample

Can we replace our own product import?

Replacement · an external product catalogue takes over

  • observed

    Search, product page and checkout read prices and stock from the import.

    3 records · 11 scenarios
  • inferred

    The external catalogue delivers the same fields. Discount rules need a mapping.

    data contract of catalog/import
  • owner-stated

    Prices may be up to 15 minutes old.

    answer from the product owner
  • unknown

    How many merchants use the bulk upload.

    no usage data

14 inputschallengednot reviewed yet

Illustrative example: an online shop

One feature. Four situations. Four answers.

“Is the import important?” has no single answer. It depends on what happens to it.

  1. 01

    Sudden outage

    Prices and stock go stale. Once the data is too old, checkout refuses orders.

  2. 02

    Controlled shutdown

    Possible with notice: checkout must show the last known stock instead of failing.

  3. 03

    Removal without replacement

    Search, product page and checkout lose their data. Not an option.

  4. 04

    Replacement

    Customers keep everything. The work is the mapping, the cutover and a way back.

Impact gives no overall importance score. User value, technical dependence and replacement effort stay separate answers.

For product, engineering and risk

Six questions, each with its own answer.

NeedDo we need this capability?
User goals, the product's purpose, obligations, alternatives
User consequencesWho loses what if it goes?
Affected tasks and user groups, workarounds, usage data
MaintenanceWhat does keeping it cost?
Operations, changes, support, incidents, specialist knowledge
ReplacementHow do we move away in a controlled way?
Decoupling, substitute, migration, transition, rollback
OutageWhat stops working if it fails?
Direct and transitive dependencies, fallbacks, recovery
Operational riskWhich risks follow from that?
Likelihood where evidence exists, consequence, detection, countermeasures

Evidence

Every sentence says where it comes from.

A second agent challenges each assessment and turns what it cannot confirm into questions for the owner. Only a person marks an assessment as reviewed.

  • observed

    Found in code, records or test results, with the exact revision.

  • inferred

    Derived from observations, with the reasoning that leads there.

  • owner-stated

    An answer from the person responsible, attributed to them.

  • unknown

    No evidence. It stays unknown and never counts as zero.

Change

When the product changes, so can the answer.

Each assessment records the inputs it used: records, capability membership, owner answers, rubric version. If one of them changes, or a new feature joins the capability, Impact flags the assessment instead of showing a stale answer.

  • taxonomy:catalogpurposeunchanged
  • capability:catalog/importmembershipnew feature: fallback feed
  • record:catalog/import/product-importintent · contract · depends_onunchanged
  • owner:pricing-freshnessanswerunchanged

Reassess: a new fallback can change the outage answer.

Coming soon

Impact assesses the features you describe today.

Impact builds on FACET records. Every feature you capture with FACET Core now is one Impact can answer questions about later.

Impact is released once its answers hold up against independently labelled reference cases.