What does Results record?
Every message the platform sent on the owner's behalf: a manager email about a stuck backlog, a win-back list pushed to the email tool, a markdown proposal that was approved, an article published, a campaign approved. For each, what was sent, when, to whom, and on what evidence.
Then the measurement. For each send, the engine compares the relevant figure in the window before with the window after: did the stuck lines move, did the lapsed customers buy, did the page fix reduce that product's returns.
- What was sent, when, to whom, and on what evidence.
- The figure the send was meant to move, and the window it is measured over.
- The before-and-after comparison on your own data, with its assumption stated.
- Whether the recipient acted, where the send went to a manager.
- The roll-up into money identified, released, at risk and forgone.
What a verdict is, and what it is not
A verdict is a before-and-after comparison on the brand's own data, with the window, the metric and the assumption stated. It is not a causal claim. Other things happened in the same window, and the ledger says so rather than awarding credit it cannot prove.
That honesty is the point. An owner reading a verdict knows whether the number moved after the action, which is a fact, and can weigh it against what else was happening, which is judgement.
What is the value ledger?
Verdicts roll up into a running tally: money identified, meaning found and proposed; money released, meaning measured after an approved action; money at risk, meaning still exposed; and money forgone, meaning proposals that were skipped. Each total drills to the individual decisions that make it up.
Forgone is recorded deliberately. A platform that only counts its wins would let an owner forget the decisions they declined, and the pattern of what gets skipped is one of the more useful things a brand can learn about itself.
How does manager accountability work?
Where a send went to a manager, the ledger records whether the action was taken and what followed. A store manager who received a staffing plan and a merchandiser who received a markdown list both appear with the outcome, not to assign blame but so the next proposal goes to the person who acts on it.
It also closes the loop on the proposal itself. A recommendation that is repeatedly sent and repeatedly not acted on is usually a bad recommendation, and the ledger makes that visible.
What is shipped and what is roadmap
Shipped: the record of every send, the before-and-after verdict per send, and the value ledger with its four totals drillable to the decision. Roadmap: requiring a recommendation to declare its expected outcome, measurement method and date before it can be shown, then publishing the accuracy of those declarations, and the outcome-based pricing and found-money guarantee that become possible once that exists.
The distinction matters and the engine keeps it. What is measured today is what happened after a send; what is coming is the discipline of declaring the prediction first.
How it changes the way an owner works
After a month, the ledger is the honest record of whether the platform earned its place: what it found, what was approved, what moved, and what was left. It turns a series of recommendations into a track record, and a track record is what makes the next recommendation easier to judge.
On ftes.ai the same discipline runs per role: each Digital FTE is judged on one number, and the ledger is where the money side of that judgement is recorded.
How should an owner read the ledger monthly?
Start with forgone. The proposals that were skipped say more about the business's constraints than the ones that were approved, and a pattern of skipping the same class of decision usually means the proposal is wrong for this brand rather than that the owner is wrong.
Then read released against identified. A wide gap means findings are not turning into actions, which is a workflow problem rather than an analytics one. A narrow gap with small numbers means the engine is finding little, which is a different and more honest problem.