L3 · Cascade
Cascade, arrival and remediation
One approved corporate position, pushed to every market, franchise partner and employee cohort, in the right language and register, inside one window. This surface measures whether it arrived and was understood, and opens the remediation job when it was not.
Markets
14
Languages
10
Cohorts
5
Human touches
2
Distribution
Distribution is not arrival
In a system that is ~95% franchised, the company reaches most of the people who must act on a position through an independent business owner. Distribution is solved; arrival is not.
“What matters is what arrived, to whom, and whether they understood it. A position that reaches 61% of franchise owners in Brazil and is logged as delivered is not a communications success, it is an undetected failure with a two-week fuse.”
Supports · The measurement runs whether or not anyone is looking at it. A cohort below floor becomes a named job with an owner, automatically.
Markets landed
0
of 14 in this cascade
Never landed
0
detected by the platform, not by a person
Divergence reports
0
1 contradiction · the rest permitted and logged
Jobs auto-opened
0
6 still open
The cascade on the board
Value-menu pricing architecture — corporate position
G1 POSITION committed on the value-menu pricing architecture. The change cascade opened automatically as a first-class job with a named owner — step 01, unattended.
US
1 markets
United States
~95% franchised. Direct corporate channel to owners; no direct channel to crew.
Reach
97%
Compr.
91%
Failed
0
IOM
7 markets
International Operated Markets
~89% franchised. Company-operated market structure, corporate channel intact, works councils in scope.
Reach
92%
Compr.
79%
Failed
0
IDL
6 markets
International Developmental Licensed
~99% through strategic licensees in roughly 80 markets. Corporate does not own the channel.
Reach
83%
Compr.
70%
Failed
3
Read
The licensee operates its own owner-communication channel. The corporate cascade reached 61% of franchise owners and was not detected as a failure by any pre-platform process.
Market lead · Developmental licensee — corporate affairs — weakest cohort · Franchise owners
Divergence report
No divergence. The source variant ran unchanged in this market.
Cascade record
Cascade
Situation
Source position
Workflow
Owner
Window
Variants
Landing floors
Reach
85%
Comprehension
70%
Floors are configuration, set per cascade at intake and visible in the audit trail. Step 15 opens a remediation job automatically whenever a cohort falls below either floor — no human has to notice.
Comprehension read completes at 72h · divergence monitored for 14 days
Where the humans are
Two touches in seventeen steps
Fifteen of the seventeen specified steps run unattended — 88.2%, the highest of the five workflows. The two that do not: someone owns a claim that failed substantiation, and someone signs the release in their own market.
Claim owner substantiates
Claim owner
Fires only on failure. Silent when the Evidence Ledger is clean for that market. On this cascade it fired twice — once on a US comparison claim outside UK scope, once on a nutrition adjacency in Germany.
Market leads confirm release
Market comms leads
Confirm in parallel, not in series. Below the materiality floor this is one click. Fourteen leads, one window — the serial version of this step is what used to take nine days.
The specified sequence
Seventeen steps, read from the specification
Each row carries the autonomy level and owner the specification assigns it, including the two steps where a person is in the path.
Against the way it is done now
What changes, line by line
Market-by-market, this position previously took nine to fourteen working days and no one measured whether it arrived.
Elapsed time to last market
9–14 working days
18 hours
Serial market-by-market handoff replaced by one parallel release window.
Variants produced
Market teams rewrite from the source deck
63 variants from one approved position
Five cohorts × fourteen markets, minus cohorts a market does not have.
Divergence visibility
None. Adaptation is silent.
Mandatory divergence report per variant
Emphasis permitted, adaptation logged, contradiction blocked.
Landing measurement
Cascade reported as sent
Reach and comprehension per cohort per market
The difference between distribution and arrival.
Failure detection
Surfaces weeks later, if at all
6 remediation jobs opened automatically
Two markets never landed. Both were detected by the platform, not by a person.