Shadow mode
Shadow mode is how you decide to turn enforcement on, or to deploy an update, without any risk. The feed runs in log-only mode and produces a report showing exactly what would change on your own transactions. Nothing is enforced; every number is a projection your team approves.
How it works
You run the feed against your own transactions in log-only mode. It records what it would have blocked or sent to review, with zero customer impact, and produces an impact report. You compare that projection against what your current setup does, then decide.
What the report shows
- What would newly block, and the dollar volume, that your current setup does not catch. This is the headline change.
- What would newly go to review.
- Which accounts and merchants drive the change, so you can see whether it is a few big merchants or a broad tail.
- What would stop being flagged, shown honestly, so the change is never one-sided.
- Breakdowns by category and evidence strength, and the projected decline-rate change.
Two comparisons
- Detect The Bet versus your existing block. The first question: what does the feed add on top of what you already do? It runs alongside your block, so it only adds coverage and never unblocks a transaction your block already declines.
- Update A versus update B. For an existing deployment: what does a new release change versus the one running now, including any merchants that were reclassified in either direction.
Setting success up front
Before a pilot starts, we agree on two or three numeric success criteria, measured from your own shadow-mode results (for example, coverage your existing block missed, and a ceiling on false flags). Shadow mode is what produces those numbers, so the decision to enforce is made on evidence, not on our claims.