diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index 4cdde19..2a94829 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -2,6 +2,22 @@
 
 Thank you for your interest in contributing to the FFE system test data repository!
 
+## Changing the contract
+
+Read [contract ownership](README.md#contract-ownership) before changing an
+expectation. These fixtures define behavior for downstream implementations;
+they are not snapshots of whichever SDK or backend happens to be deployed.
+
+Describe the rule being added or corrected, include a concrete input and expected
+result, and explain the effect on consumers. Validate against actual native
+implementations where available and record the tested revisions and capability
+coverage. A consumer mismatch calls for investigation; it does not by itself
+justify changing the expected result. Follow [handling a mismatch](README.md#handling-a-mismatch).
+
+Keep deployment prerequisites, release versions, and temporary consumer workarounds
+in downstream changes. Contract documentation should remain valid as those
+projects ship support.
+
 ## Getting Started
 
 1. Fork the repository
@@ -76,12 +92,24 @@ merge by itself.
 
 ## Updating Downstream Repositories
 
-After your changes are merged to this repository, update the submodule references in downstream repos. Each downstream repository should:
-
-1. Keep `ffe-system-test-data` as a git submodule adjacent to the OpenFeature/FFE tests
-2. Configure Dependabot with `package-ecosystem: "gitsubmodule"` so fixture updates are proposed automatically
-3. Load `ufc-config.json` and loop over every `evaluation-cases/*.json` file in unit tests
-4. Avoid copied fixture directories and programmatic-only shared evaluator cases
+After your changes are merged to this repository, update the pinned fixture
+revision in downstream repos. Each rules-based consumer should:
+
+1. Pin an exact fixture revision, normally with a git submodule adjacent to the
+   OpenFeature/FFE tests. Generated snapshots must retain the source repository
+   and commit SHA and be refreshed from that source.
+2. Automate fixture updates. For submodules, configure Dependabot with
+   `package-ecosystem: "gitsubmodule"`; for snapshots, automate regeneration from
+   the source revision.
+3. Load `ufc-config.json` and loop over every `evaluation-cases/*.json` file in
+   unit tests.
+4. Run the cases for every implemented capability as well as the baseline;
+   follow the [runner requirements](README.md#implement-the-rules-based-runner).
+5. Preserve shared expectations unchanged. Avoid hand-maintained copies and
+   programmatic-only replacements for shared evaluator cases.
+
+Precomputed-assignment consumers use the same pinning and provenance rules with
+the [precomputed fixture runner](precomputed-assignments/README.md).
 
 Current downstream consumers are `system-tests`, `dd-trace-go`, `dd-trace-java`, `dd-trace-js`, `dd-trace-py`, `dd-trace-rb`, `dd-trace-dotnet`, `dd-trace-php`, `libdatadog`, and `openfeature-js-client`.
 
diff --git a/README.md b/README.md
index 1969b1a..191bb98 100644
--- a/README.md
+++ b/README.md
@@ -1,10 +1,66 @@
 # ffe-system-test-data
 
-Canonical test data for Datadog's Feature Flags & Experimentation (FFE) SDK implementations.
+The canonical behavioral contract for Datadog's Feature Flags & Experimentation
+(FFE) implementations, expressed as shared test inputs and expected results.
 
 ## Overview
 
-This repository contains the canonical set of flag configurations and evaluation test cases used to validate FFE SDK implementations across multiple languages. It serves as a single source of truth consumed via git submodules by:
+A fixture describes what an implementation must do with a particular
+configuration, evaluation context, and set of options. The expected result can
+include a value, reason, error code, selected allocation metadata, or emitted
+events. The fixtures and their documented rules define the shared behavior
+covered here; downstream projects implement that behavior in their own language.
+
+## Contract ownership
+
+**Downstream implementations must conform to this contract.** A language's
+standard library, an existing SDK behavior, or the current backend deployment
+does not redefine an expected result. For example, if a timestamp boundary test
+distinguishes microseconds, the consumer must preserve that precision even when
+its usual date type only stores milliseconds.
+
+```mermaid
+flowchart TD
+    C["Canonical contract<br/>Fixtures, schemas, documented rules"]
+    P["Configuration producers"]
+    E["SDK and backend evaluators"]
+    T["Downstream conformance tests"]
+    C -->|implement required configuration semantics| P
+    C -->|implement required behavior| E
+    C -->|load inputs and expected results unchanged| T
+    P -->|produce configuration| E
+    E -->|return actual results| T
+```
+
+Configuration producers must emit the required fields with their defined
+meaning. Evaluators implement parsing and evaluation. A downstream test adapter
+maps the shared inputs and results to native APIs and checks the implementation
+against the fixture. Rules-based tests can load `ufc-config.json` directly;
+they do not need a running configuration compiler or deployed backend.
+
+The contract can be extended or corrected through review in this repository.
+An implementation mismatch is evidence to investigate, not an automatic reason
+to change an expectation. See [handling a mismatch](#handling-a-mismatch).
+
+Contract changes can land before every consumer implements them. SDK release
+readiness, backend rollout order, and supported product versions belong in the
+downstream projects' changes and release plans. This README describes stable
+behavior and does not need an update whenever a consumer ships support.
+
+## Start here
+
+| What you are implementing | Contract to read | Test integration |
+| --- | --- | --- |
+| An SDK or backend that evaluates UFC rules | [`ufc-config.json`](ufc-config.json), [`evaluation-cases/`](evaluation-cases/), and [optional capabilities](#optional-capabilities) | Load the configuration through the real ingestion path and evaluate each case with the native evaluator. |
+| An SDK that consumes precomputed assignments | [Precomputed fixture guide](precomputed-assignments/README.md) and [schema](schemas/precomputed-assignment.schema.json) | Supply the mocked response, run the ordered evaluations, and check results and emitted events. |
+
+These families have different formats and runners. Passing one does not
+establish coverage of the other. When changing the contract, also read
+[CONTRIBUTING.md](CONTRIBUTING.md).
+
+## Consumers
+
+Downstream projects include:
 
 - [system-tests](https://github.com/DataDog/system-tests) - Parametric tests
 - [dd-trace-py](https://github.com/DataDog/dd-trace-py) - Python tracer
@@ -21,22 +77,19 @@ This repository contains the canonical set of flag configurations and evaluation
 
 ```
 ffe-system-test-data/
-├── ufc-config.json          # Master flag configuration (UFC format)
+├── ufc-config.json          # Shared rules-based configuration, including malformed test inputs
 ├── evaluation-cases/
 │   └── test-*.json          # Evaluation test case files
 ├── precomputed-assignments/
 │   ├── README.md            # Precomputed assignment fixture format
-│   └── cases/*.json         # Precomputed assignment test fixtures
-└── regex-conformance/
-    ├── targeting-regex-conformance.json   # FFE authoring and matching contract
-    ├── targeting-regex-conformance.sha256 # SHA-256 of the JSON bytes
-    ├── validate-targeting-regex-conformance.jq # Canonical schema validator
-    └── test-validate-targeting-regex-conformance.sh # Validator regression tests
+│   └── cases/               # Precomputed assignment fixtures, discovered recursively
+├── schemas/                # Precomputed assignment fixture schema
+└── ci/                     # Static validators and downstream comparison runner
 ```
 
 ## Usage
 
-### As a Git Submodule
+### Pin the fixture revision
 
 Add this repository as a submodule to your project:
 
@@ -52,31 +105,75 @@ git clone --recurse-submodules <your-repo>
 git submodule update --init --recursive
 ```
 
-### In Tests
-
-1. Load `ufc-config.json` to initialize your UFC evaluator
-2. For each file in `evaluation-cases/`, parse the JSON array
-3. For each test case, call your evaluator with `flag`, `defaultValue`, `targetingKey`, and `attributes`
-4. Assert the result matches `result.value` and `result.reason`
+Commit the submodule pointer in the downstream repository so local runs and CI
+use the same fixture commit. During development, a consumer can pin a fixture PR
+commit; update to the merged fixture revision before landing the consumer change.
+If a repository uses a generated snapshot instead, retain the source repository
+and exact commit SHA and refresh it from that source. Do not hand-edit expected
+results in a downstream copy.
+
+### Implement the rules-based runner
+
+1. Load `ufc-config.json` through the consumer's real configuration parser and
+   initialize its evaluator. Preserve the malformed inputs used by isolation tests.
+2. Discover every `evaluation-cases/*.json` file and parse its array. Keep case
+   order within each file, including repeated evaluations. Fail if the fixture
+   directory is missing or empty.
+3. Run every baseline case with default provider options, using `flag`,
+   `variationType`, `defaultValue`, `targetingKey`, and `attributes` as supplied.
+   Assert the expected value and reason, and any specified error code. Consumers
+   exposing selected metadata must also check the supplied `serialId` and `doLog`
+   expectations against the actual selection.
+4. For each capability the consumer implements, map its name to the native
+   option and rerun every case carrying that capability. Assert its expected
+   result instead of the baseline result. Run capabilities separately.
+5. Report the fixture SHA and executed case counts, including counts per
+   supported capability. Fail if a supported capability discovers no cases.
+   Skipping every capability case must not look like successful capability support.
+
+The adapter may translate API names and native result containers. It must
+preserve the meaning and precision of the inputs and expected results, including
+`false`, `0`, empty values, and the distinction between absent and null fields.
+Use SDK-local tests for behavior outside these fixtures, such as concurrency,
+transport failures, and provider lifecycle.
 
 ## File Formats
 
 ### UFC Config (`ufc-config.json`)
 
-The UFC (Unified Flag Configuration) file contains flag definitions in the format used by Datadog's Remote Configuration:
+The UFC (Universal Flag Configuration) file contains a document envelope and a
+`flags` map. This is a complete minimal example of the rules-based format:
 
 ```json
 {
-  "flag-key": {
-    "key": "flag-key",
-    "enabled": true,
-    "variationType": "STRING|BOOLEAN|INTEGER|NUMERIC|JSON",
-    "variations": { ... },
-    "allocations": [ ... ]
+  "createdAt": "2026-01-01T00:00:00Z",
+  "format": "SERVER",
+  "environment": { "name": "test" },
+  "flags": {
+    "flag-key": {
+      "key": "flag-key",
+      "enabled": true,
+      "variationType": "BOOLEAN",
+      "variations": { "on": { "key": "on", "value": true } },
+      "allocations": [
+        {
+          "key": "ordinary",
+          "rules": [],
+          "splits": [{ "variationKey": "on", "shards": [] }],
+          "doLog": false
+        }
+      ]
+    }
   }
 }
 ```
 
+Valid allocations require an explicit boolean `doLog`. `false` is a valid
+logging decision; omission and `null` are not valid substitutes. This input
+requirement is separate from the optional `result.doLog` test expectation below.
+Some fixtures deliberately contain malformed configuration to test error
+handling; their presence does not make that input valid producer output.
+
 ### Evaluation Test Cases (`evaluation-cases/test-*.json`)
 
 Each evaluation case uses a universal schema with the following fields:
@@ -101,11 +198,11 @@ Example:
 [
   {
     "flag": "flag-key",
-    "variationType": "STRING",
-    "defaultValue": "default",
+    "variationType": "BOOLEAN",
+    "defaultValue": false,
     "targetingKey": "user-123",
-    "attributes": { "country": "US" },
-    "result": { "value": "expected-value", "reason": "TARGETING_MATCH" }
+    "attributes": {},
+    "result": { "value": true, "reason": "STATIC" }
   }
 ]
 ```
@@ -117,6 +214,11 @@ For each supported entry in `capabilities`, enable that capability and repeat th
 same evaluation. Check its expected result instead. Ignore unsupported entries,
 not the baseline case. Capabilities run separately, not in combination.
 
+A capability is a named optional behavior, not a release-status flag or a way
+to exempt an implemented feature from conformance. A consumer that implements
+the feature must wire its capability into the runner; updating the fixture pin
+alone only adds baseline coverage.
+
 `experimentationOptOutImmediate` enables immediate experimentation opt-out. The
 input stays in `attributes.dd_experimentation_opt_out_at`. For example:
 
@@ -134,8 +236,8 @@ input stays in `attributes.dd_experimentation_opt_out_at`. For example:
 }
 ```
 
-The new cases reuse `ufc-config.json`. Immediate mode skips allocations whose
-`doLog` is true (or omitted), then continues ordinary allocations. A valid
+The cases reuse `ufc-config.json`. Immediate mode skips allocations whose
+`doLog` is true, then continues ordinary allocations. A valid
 RFC3339 timestamp excludes immediately, including a future timestamp. A missing
 attribute leaves evaluation unchanged. An invalid value returns the caller
 default with `ERROR` / `INVALID_CONTEXT`. Default mode ignores the attribute.
@@ -154,6 +256,42 @@ behavior. A supporting SDK adds the capability mapping to its existing reader;
 updating the submodule alone does not test the enabled behavior. SDK tests retain
 responsibility for telemetry, cache state, and provider lifecycle.
 
+#### Future experiments only
+
+`experimentationOptOutFutureExperimentsOnly` uses the same subject timestamp.
+Map it to the provider's future-experiments-only mode, or
+`DD_FEATURE_FLAGS_EXPERIMENTATION_OPT_OUT_MODE=future_experiments_only`.
+
+| Allocation | With a valid subject cutoff |
+| --- | --- |
+| `doLog=false` | Evaluate normally. |
+| `doLog=true`, activation before cutoff | Evaluate normally, including normal exposures. |
+| `doLog=true`, activation at or after cutoff | Skip before targeting and sharding. Continue later allocations. |
+| `doLog=true`, missing or null activation | Skip. Unknown activation does not establish an older experiment. |
+
+The allocation field is `randomizationActivatedAt`, an optional RFC3339 timestamp.
+Compare instants and preserve the fractional precision exercised by the cases;
+rounding or truncating to milliseconds changes the required result. Choose a
+native representation that retains the required precision. Do not use `startAt`,
+`endAt`, the configuration's `createdAt`, or the SDK's clock as a substitute.
+Existing schedule checks still apply to eligible allocations. No subject history
+or SDK persistence is required.
+
+The cutoff concerns allocation activation, not subject enrollment. An older
+experiment can still assign a subject for the first time. Holdout allocations
+use the same rule; the compiler supplies their activation metadata.
+
+An absent subject timestamp preserves normal evaluation. Invalid subject input
+returns `ERROR` / `INVALID_CONTEXT` before allocation selection, as in immediate mode.
+Supply the time the subject opted out. A later cutoff keeps experiments activated
+before that later time eligible, including experiments started after the subject
+opted out. Targeting and schedules still apply.
+Removing the attribute permits experimentation again.
+
+The future-mode cases reuse existing flags. Added activation fields do not change
+baseline or immediate results. Malformed activation parsing, configuration refresh,
+exposure deduplication, and telemetry independence also need SDK-local tests.
+
 ### Selected split and allocation metadata
 
 `test-case-serial-id-metadata.json` and `test-case-span-enrichment-inputs.json`
@@ -180,9 +318,12 @@ capacity-limit assertions.
 
 ### SDK-Specific Fields
 
-The shared fixtures intentionally exclude SDK-specific fields such as `variant` and `flagMetadata`. SDKs that need these fields should compute them at test load time from the universal fields. For example:
+The shared fixtures use universal result fields rather than SDK-specific
+containers such as `variant` and `flagMetadata`. Adapters should read the native
+evaluation result and map it to these fields. For additional SDK-local assertions:
 
-- **variant**: Derive from the flag configuration in `ufc-config.json` by matching the result value
+- **variant**: Use the selected variation key. Equal result values do not
+  necessarily identify the same variation.
 - **flagMetadata**: Map the selected split's `serialId` and allocation's `doLog` to
   the SDK's metadata keys; assert the optional expected fields above directly
 
@@ -197,7 +338,7 @@ expected exposure/flagevaluation emission counts.
 |-------|------|-------------|
 | `name` | string | Fixture identifier (matches filename stem) |
 | `description` | string | Human-readable description of what the fixture covers |
-| `context` | object | Evaluation context: `targetingKey` (string) and `attributes` (object) |
+| `context` | object | Evaluation context: `targetingKey` (string or null) and `attributes` (object) |
 | `response` | object | Full precomputed-assignments JSON:API envelope served by the mock CDN |
 | `_skip` | array? | Skip fixture for listed platforms: `[{"platform": "web", "reason": "..."}]` |
 | `_include` | array? | Run fixture only for listed platforms |
@@ -213,29 +354,6 @@ expected exposure/flagevaluation emission counts.
 
 See [precomputed-assignments/README.md](precomputed-assignments/README.md) for further detail.
 
-### Targeting Regex Conformance
-
-`regex-conformance/targeting-regex-conformance.json` is a standalone, versioned
-contract for authoring targeting regular expressions in FFE. It is intentionally
-outside `evaluation-cases/`; consumers of that directory parse every JSON file
-as a complete UFC evaluation case.
-
-The fixture records native observations for four implementations: Go `regexp`,
-RE2JS, the Rust rules-based evaluator, and the Rust rkyv evaluator. The accepted
-authoring subset is narrower: accepted cases must also evaluate consistently in
-the shipped Java, JavaScript, and .NET SDK evaluators. Several SDKs share the
-Rust evaluator, so agreement across those SDKs is not evidence from independent
-regex engines.
-
-Each regex case has a stable ID, an FFE authoring `contract`, raw and normalized
-patterns, native compile observations, an input, and an unanchored match
-observation when the modeled engines agree. Go and RE2JS consumers compile
-`normalizedPattern`; Rust consumers compile `rawPattern`. Cases with differences
-between modeled engines include per-engine expectations. Downstream SDK checks
-must require consistent behavior for accepted cases. A native engine accepting
-rejected syntax does not change the authoring contract. The adjacent SHA-256
-file lets downstream tests detect fixture drift.
-
 ## Automated Validation
 
 Pull requests run a blocking static validation check over `ufc-config.json`
@@ -247,7 +365,7 @@ context validation precedes flag lookup. The baseline still requires `FLAG_NOT_F
 
 The validator deliberately does not fully schema-check allocation internals.
 Some fixtures contain malformed flag fields on purpose to verify that consumers
-reject only the affected flag. Run the same check locally with:
+reject only the affected flag.
 
 Consumers exclude malformed flags from the active evaluation map while retaining
 their rejected keys for the current configuration. Evaluating a rejected key
@@ -257,8 +375,41 @@ maps atomically so fixed or deleted flags do not leave stale rejection entries.
 
 ```bash
 python3 ci/validate-fixtures.py
+python3 -m unittest discover -s ci -p 'test_*.py'
 ```
 
+For precomputed fixtures, install `jsonschema` and run
+`python3 ci/validate-precomputed-fixtures.py`. See the
+[precomputed fixture guide](precomputed-assignments/README.md#schema-and-validation).
+Static validation checks fixture structure; it does not execute an SDK or prove
+the expected behavior is correct.
+
+The [downstream compatibility workflow](.github/workflows/downstream-compatibility.yml)
+runs native consumer tests against the base and proposed fixture revisions.
+Its results are advisory evidence for review, even when a new mismatch makes
+the comparison job fail. Inspect the tested consumer revision, runner, and
+capability coverage; a passing baseline run does not prove a new capability was
+executed. Local or CI conformance also does not prove deployment readiness.
+
+## Handling a mismatch
+
+1. Record the fixture commit, consumer commit, case file and index, enabled
+   capability, expected result, and actual result. Run through the real parser
+   and evaluator to distinguish an adapter bug from an implementation bug.
+2. Check the documented rule and reproduce the relevant boundary. Once the
+   expectation is confirmed, fix the downstream implementation to meet it.
+   Keep any temporary downstream exclusion explicit and tracked; it does not
+   change the shared expectation or establish conformance.
+3. If the fixture is wrong or the intended behavior needs to change, propose
+   that correction here with a concrete example, rationale, and consumer impact.
+   Obtain contract review before propagating changed expectations downstream.
+
+For example, a consumer treating `.123001` seconds as equal to `.123` has lost
+precision required by the opt-out boundary cases. Fix its timestamp handling;
+do not round the fixture to make that implementation pass. Likewise, backend
+support that has not shipped yet is a rollout dependency, not a defect in a
+fixture defining the intended behavior.
+
 ## Evaluation Cases
 
 | File | Description |
@@ -269,6 +420,9 @@ python3 ci/validate-fixtures.py
 | `test-case-disabled-flag.json` | Disabled flag returning DISABLED reason |
 | `test-case-empty-flag.json` | Flag with empty configuration |
 | `test-case-empty-string-variation.json` | Flag returning empty string as value |
+| `test-case-experimentation-opt-out.json` | Baseline, immediate, and future opt-out behavior, including invalid subject input and ordinary fallbacks |
+| `test-case-experimentation-opt-out-future.json` | Activation cutoffs, fractional precision, timezone offsets, and unknown activation |
+| `test-case-experimentation-opt-out-do-log.json` | Ordinary allocations remain eligible when activation equals or follows the cutoff |
 | `test-case-falsy-value-assignments.json` | Flags returning falsy values (0, false, empty) |
 | `test-case-flag-with-empty-string.json` | Flag with empty string in configuration |
 | `test-case-integer-flag.json` | Integer-typed flag evaluation |
@@ -296,6 +450,8 @@ python3 ci/validate-fixtures.py
 | `test-case-semver-validation-flag.json` | SemVer 2 syntax boundaries retained by the extended parser, invalid syntax, non-string attributes, and invalid configured comparands |
 | `test-case-semver-invalid-comparand-waterfall.json` | An invalid configured SemVer comparand in a non-last allocation of a waterfall must reject the whole flag at parse time rather than skipping it and falling through to a later valid default |
 | `test-case-semver-invalid-comparand-categories.json` | Configured SemVer comparand boundaries: a valid hyphen-heavy prerelease matches, while invalid syntax (including consecutive dots and a Ruby-style `.DEV` suffix), v-prefixed, leading-zero, and non-string comparands reject their flag at parse time with PARSE_ERROR without poisoning the rest of the configuration |
+| `test-case-serial-id-metadata.json` | Selected split IDs, including zero, absence, targeting, and repeated evaluations |
+| `test-case-span-enrichment-inputs.json` | Selected IDs and logging decisions used by downstream span-enrichment tests |
 | `test-case-start-and-end-date-flag.json` | Flag with start/end date time bounds |
 | `test-case-unknown-fields-tolerance.json` | Unknown UFC object fields are ignored |
 | `test-case-unknown-operator-isolation.json` | A flag with an unknown operator is removed without poisoning valid flags |
@@ -307,11 +463,16 @@ python3 ci/validate-fixtures.py
 
 ## Origin
 
-These fixtures are derived from the Go SDK (`dd-trace-go`) reference implementation, which was the first to implement and validate the OpenFeature `reason` field. The Go fixtures serve as the canonical source of truth for expected evaluation behavior across all Datadog SDKs.
+These fixtures originated in the Go SDK (`dd-trace-go`). This repository now
+owns the shared contract; Go and every other downstream implementation consume
+it and are subject to the same conformance requirements.
 
 ## Downstream Updates
 
-Downstream repositories should consume this repository as a git submodule and run their fixture coverage by loading every JSON file in `evaluation-cases/`. Shared evaluator behavior should be added here first, then downstream repositories should update their submodule SHA. Do not add copied JSON fixture directories or language-only programmatic cases for behavior that belongs in this shared fixture set.
+Add shared evaluator behavior here, then update each consumer's
+[pinned fixture revision](#pin-the-fixture-revision) and native test runner.
+See [Updating Downstream Repositories](CONTRIBUTING.md#updating-downstream-repositories)
+for the integration checklist.
 
 ### Informational Compatibility Preview
 
diff --git a/evaluation-cases/test-case-experimentation-opt-out-do-log.json b/evaluation-cases/test-case-experimentation-opt-out-do-log.json
new file mode 100644
index 0000000..0d2e8d5
--- /dev/null
+++ b/evaluation-cases/test-case-experimentation-opt-out-do-log.json
@@ -0,0 +1,22 @@
+[
+  {
+    "description": "ordinary allocation remains eligible when activation equals cutoff",
+    "flag": "no-log-flag", "variationType": "BOOLEAN", "defaultValue": false, "targetingKey": "subject",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
+    "result": { "value": true, "reason": "STATIC" },
+    "capabilities": {
+      "experimentationOptOutImmediate": { "value": true, "reason": "STATIC" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": true, "reason": "STATIC" }
+    }
+  },
+  {
+    "description": "ordinary allocation remains eligible when activation is after cutoff",
+    "flag": "no-log-flag", "variationType": "BOOLEAN", "defaultValue": false, "targetingKey": "subject",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T11:59:59Z" },
+    "result": { "value": true, "reason": "STATIC" },
+    "capabilities": {
+      "experimentationOptOutImmediate": { "value": true, "reason": "STATIC" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": true, "reason": "STATIC" }
+    }
+  }
+]
diff --git a/evaluation-cases/test-case-experimentation-opt-out-future.json b/evaluation-cases/test-case-experimentation-opt-out-future.json
new file mode 100644
index 0000000..ea33f20
--- /dev/null
+++ b/evaluation-cases/test-case-experimentation-opt-out-future.json
@@ -0,0 +1,146 @@
+[
+  {
+    "description": "older allocation continues",
+    "flag": "integer-flag", "variationType": "INTEGER", "defaultValue": 0, "targetingKey": "5",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-01T12:00:00.000001Z" },
+    "result": { "value": 1, "reason": "SPLIT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": 1, "reason": "SPLIT" }
+    }
+  },
+  {
+    "description": "equal activation is excluded",
+    "flag": "integer-flag", "variationType": "INTEGER", "defaultValue": 0, "targetingKey": "5",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-01T12:00:00Z" },
+    "result": { "value": 1, "reason": "SPLIT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "DEFAULT" }
+    }
+  },
+  {
+    "description": "newer allocation is excluded",
+    "flag": "integer-flag", "variationType": "INTEGER", "defaultValue": 0, "targetingKey": "5",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-01T11:59:59.999999Z" },
+    "result": { "value": 1, "reason": "SPLIT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "DEFAULT" }
+    }
+  },
+  {
+    "description": "equal instant with negative offset",
+    "flag": "integer-flag", "variationType": "INTEGER", "defaultValue": 0, "targetingKey": "5",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-01T08:00:00-04:00" },
+    "result": { "value": 1, "reason": "SPLIT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "DEFAULT" }
+    }
+  },
+  {
+    "description": "equal instant with positive offset",
+    "flag": "integer-flag", "variationType": "INTEGER", "defaultValue": 0, "targetingKey": "5",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-01T14:00:00+02:00" },
+    "result": { "value": 1, "reason": "SPLIT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "DEFAULT" }
+    }
+  },
+  {
+    "description": "future cutoff is not an evaluation schedule",
+    "flag": "integer-flag", "variationType": "INTEGER", "defaultValue": 0, "targetingKey": "5",
+    "attributes": { "dd_experimentation_opt_out_at": "2099-01-01T00:00:00Z" },
+    "result": { "value": 1, "reason": "SPLIT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": 1, "reason": "SPLIT" }
+    }
+  },
+  {
+    "description": "skip newer targeted allocation and continue older randomization",
+    "flag": "integer-flag", "variationType": "INTEGER", "defaultValue": 0, "targetingKey": "5",
+    "attributes": { "country": "US", "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
+    "result": { "value": 3, "reason": "TARGETING_MATCH" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": 1, "reason": "SPLIT" }
+    }
+  },
+  {
+    "description": "targeted allocation remains eligible before cutoff",
+    "flag": "integer-flag", "variationType": "INTEGER", "defaultValue": 0, "targetingKey": "5",
+    "attributes": { "country": "US", "dd_experimentation_opt_out_at": "2026-10-21T12:00:00Z" },
+    "result": { "value": 3, "reason": "TARGETING_MATCH" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": 3, "reason": "TARGETING_MATCH" }
+    }
+  },
+  {
+    "description": "skip both allocations",
+    "flag": "integer-flag", "variationType": "INTEGER", "defaultValue": 0, "targetingKey": "5",
+    "attributes": { "country": "US", "dd_experimentation_opt_out_at": "2026-10-01T12:00:00Z" },
+    "result": { "value": 3, "reason": "TARGETING_MATCH" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "DEFAULT" }
+    }
+  },
+  {
+    "description": "fractional activation equals cutoff",
+    "flag": "special-characters", "variationType": "JSON", "defaultValue": {}, "targetingKey": "ben",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00.123Z" },
+    "result": { "value": { "a": "піклуватися", "b": "любов" }, "reason": "SPLIT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": { "a": "kümmert", "b": "schön" }, "reason": "STATIC" }
+    }
+  },
+  {
+    "description": "allocation is one microsecond older",
+    "flag": "special-characters", "variationType": "JSON", "defaultValue": {}, "targetingKey": "ben",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00.123001Z" },
+    "result": { "value": { "a": "піклуватися", "b": "любов" }, "reason": "SPLIT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": { "a": "піклуватися", "b": "любов" }, "reason": "SPLIT" }
+    }
+  },
+  {
+    "description": "allocation is one microsecond newer",
+    "flag": "special-characters", "variationType": "JSON", "defaultValue": {}, "targetingKey": "ben",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00.122999Z" },
+    "result": { "value": { "a": "піклуватися", "b": "любов" }, "reason": "SPLIT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": { "a": "kümmert", "b": "schön" }, "reason": "STATIC" }
+    }
+  },
+  {
+    "description": "missing activation is not treated as older",
+    "flag": "kill-switch", "variationType": "BOOLEAN", "defaultValue": false, "targetingKey": "alice",
+    "attributes": { "country": "US", "dd_experimentation_opt_out_at": "2099-01-01T00:00:00Z" },
+    "result": { "value": true, "reason": "TARGETING_MATCH" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": false, "reason": "DEFAULT" }
+    }
+  },
+  {
+    "description": "null activation is not treated as older",
+    "flag": "numeric_flag", "variationType": "NUMERIC", "defaultValue": 0, "targetingKey": "alice",
+    "attributes": { "dd_experimentation_opt_out_at": "2099-01-01T00:00:00Z" },
+    "result": { "value": 3.1415926, "reason": "STATIC" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "DEFAULT" }
+    }
+  },
+  {
+    "description": "activation is not startAt or document createdAt",
+    "flag": "start-and-end-date-test", "variationType": "STRING", "defaultValue": "default", "targetingKey": "subject",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
+    "result": { "value": "current", "reason": "DEFAULT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": "default", "reason": "DEFAULT" }
+    }
+  },
+  {
+    "description": "old activation does not override endAt or startAt",
+    "flag": "start-and-end-date-test", "variationType": "STRING", "defaultValue": "default", "targetingKey": "subject",
+    "attributes": { "dd_experimentation_opt_out_at": "2026-10-21T12:00:00Z" },
+    "result": { "value": "current", "reason": "DEFAULT" },
+    "capabilities": {
+      "experimentationOptOutFutureExperimentsOnly": { "value": "current", "reason": "DEFAULT" }
+    }
+  }
+]
diff --git a/evaluation-cases/test-case-experimentation-opt-out.json b/evaluation-cases/test-case-experimentation-opt-out.json
index 8120c5f..daa4f72 100644
--- a/evaluation-cases/test-case-experimentation-opt-out.json
+++ b/evaluation-cases/test-case-experimentation-opt-out.json
@@ -4,7 +4,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "not-a-timestamp" },
     "result": { "value": "default", "reason": "ERROR", "errorCode": "FLAG_NOT_FOUND" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -12,7 +13,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
     "result": { "value": {}, "reason": "ERROR", "errorCode": "TARGETING_KEY_MISSING" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": { "a": "kümmert", "b": "schön" }, "reason": "STATIC" }
+      "experimentationOptOutImmediate": { "value": { "a": "kümmert", "b": "schön" }, "reason": "STATIC" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": { "a": "kümmert", "b": "schön" }, "reason": "STATIC" }
     }
   },
   {
@@ -20,7 +22,8 @@
     "attributes": { "country": "UK" },
     "result": { "value": "red", "reason": "TARGETING_MATCH" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": "red", "reason": "TARGETING_MATCH" }
+      "experimentationOptOutImmediate": { "value": "red", "reason": "TARGETING_MATCH" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": "red", "reason": "TARGETING_MATCH" }
     }
   },
   {
@@ -28,7 +31,8 @@
     "attributes": { "country": "UK", "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
     "result": { "value": "red", "reason": "TARGETING_MATCH" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": "default", "reason": "DEFAULT" }
+      "experimentationOptOutImmediate": { "value": "default", "reason": "DEFAULT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": "red", "reason": "TARGETING_MATCH" }
     }
   },
   {
@@ -36,7 +40,8 @@
     "attributes": { "email": "alice@mycompany.com", "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
     "result": { "value": "green", "reason": "TARGETING_MATCH" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": "green", "reason": "TARGETING_MATCH" }
+      "experimentationOptOutImmediate": { "value": "green", "reason": "TARGETING_MATCH" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": "green", "reason": "TARGETING_MATCH" }
     }
   },
   {
@@ -44,7 +49,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
     "result": { "value": { "a": "піклуватися", "b": "любов" }, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": { "a": "kümmert", "b": "schön" }, "reason": "STATIC" }
+      "experimentationOptOutImmediate": { "value": { "a": "kümmert", "b": "schön" }, "reason": "STATIC" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": { "a": "kümmert", "b": "schön" }, "reason": "STATIC" }
     }
   },
   {
@@ -52,7 +58,8 @@
     "attributes": {},
     "result": { "value": 2, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 2, "reason": "SPLIT" }
+      "experimentationOptOutImmediate": { "value": 2, "reason": "SPLIT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 2, "reason": "SPLIT" }
     }
   },
   {
@@ -60,7 +67,8 @@
     "attributes": {},
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 1, "reason": "SPLIT" }
+      "experimentationOptOutImmediate": { "value": 1, "reason": "SPLIT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 1, "reason": "SPLIT" }
     }
   },
   {
@@ -68,7 +76,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
     "result": { "value": 2, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 2, "reason": "SPLIT" }
     }
   },
   {
@@ -76,7 +85,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 1, "reason": "SPLIT" }
     }
   },
   {
@@ -84,7 +94,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
     "result": { "value": 3.1415926, "reason": "STATIC" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "DEFAULT" }
     }
   },
   {
@@ -92,7 +103,8 @@
     "attributes": { "country": "US", "dd_experimentation_opt_out_at": "2026-10-19T12:00:00Z" },
     "result": { "value": true, "reason": "TARGETING_MATCH" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": false, "reason": "DEFAULT" }
+      "experimentationOptOutImmediate": { "value": false, "reason": "DEFAULT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": false, "reason": "DEFAULT" }
     }
   },
   {
@@ -100,7 +112,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T12:00:00.123456789Z" },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 1, "reason": "SPLIT" }
     }
   },
   {
@@ -108,7 +121,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "2026-10-19T08:00:00-04:00" },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 1, "reason": "SPLIT" }
     }
   },
   {
@@ -116,7 +130,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "2099-01-01T00:00:00Z" },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "DEFAULT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 1, "reason": "SPLIT" }
     }
   },
   {
@@ -124,7 +139,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "not-a-timestamp" },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -132,7 +148,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "" },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -140,7 +157,8 @@
     "attributes": { "dd_experimentation_opt_out_at": true },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -148,7 +166,8 @@
     "attributes": { "dd_experimentation_opt_out_at": null },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -156,7 +175,8 @@
     "attributes": { "dd_experimentation_opt_out_at": 1792411200 },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -164,7 +184,8 @@
     "attributes": { "email": "alice@mycompany.com", "dd_experimentation_opt_out_at": "not-a-timestamp" },
     "result": { "value": "green", "reason": "TARGETING_MATCH" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -172,7 +193,8 @@
     "attributes": { "email": "alice@mycompany.com", "dd_experimentation_opt_out_at": false },
     "result": { "value": "green", "reason": "TARGETING_MATCH" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -180,7 +202,8 @@
     "attributes": { "email": "alice@mycompany.com", "dd_experimentation_opt_out_at": null },
     "result": { "value": "green", "reason": "TARGETING_MATCH" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": "default", "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -188,7 +211,8 @@
     "attributes": { "dd_experimentation_opt_out_at": "not-a-timestamp" },
     "result": { "value": { "a": "піклуватися", "b": "любов" }, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": {}, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": {}, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": {  }, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -196,7 +220,8 @@
     "attributes": { "dd_experimentation_opt_out_at": { "timestamp": "2026-10-19T12:00:00Z" } },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   },
   {
@@ -204,7 +229,8 @@
     "attributes": { "dd_experimentation_opt_out_at": ["2026-10-19T12:00:00Z"] },
     "result": { "value": 1, "reason": "SPLIT" },
     "capabilities": {
-      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
+      "experimentationOptOutImmediate": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" },
+      "experimentationOptOutFutureExperimentsOnly": { "value": 0, "reason": "ERROR", "errorCode": "INVALID_CONTEXT" }
     }
   }
 ]
diff --git a/ufc-config.json b/ufc-config.json
index a436b88..e435a37 100644
--- a/ufc-config.json
+++ b/ufc-config.json
@@ -65,7 +65,8 @@
               "shards": []
             }
           ],
-          "doLog": true
+          "doLog": true,
+          "randomizationActivatedAt": null
         }
       ]
     },
@@ -1016,7 +1017,8 @@
             }
           ],
           "endAt": "2002-10-31T09:00:00.594Z",
-          "doLog": true
+          "doLog": true,
+          "randomizationActivatedAt": "2000-01-01T00:00:00Z"
         },
         {
           "key": "future-versions",
@@ -1027,7 +1029,8 @@
             }
           ],
           "startAt": "2052-10-31T09:00:00.594Z",
-          "doLog": true
+          "doLog": true,
+          "randomizationActivatedAt": "2026-10-01T12:00:00Z"
         },
         {
           "key": "current-versions",
@@ -1039,7 +1042,8 @@
           ],
           "startAt": "2022-10-31T09:00:00.594Z",
           "endAt": "2050-10-31T09:00:00.594Z",
-          "doLog": true
+          "doLog": true,
+          "randomizationActivatedAt": "2026-10-20T12:00:00Z"
         }
       ]
     },
@@ -1278,7 +1282,8 @@
               ]
             }
           ],
-          "doLog": true
+          "doLog": true,
+          "randomizationActivatedAt": "2026-10-01T12:00:00Z"
         },
         {
           "key": "rollout",
@@ -1384,7 +1389,8 @@
               ]
             }
           ],
-          "doLog": true
+          "doLog": true,
+          "randomizationActivatedAt": "2026-10-20T12:00:00Z"
         },
         {
           "key": "50/50 split",
@@ -1421,7 +1427,8 @@
               ]
             }
           ],
-          "doLog": true
+          "doLog": true,
+          "randomizationActivatedAt": "2026-10-01T12:00:00Z"
         }
       ]
     },
@@ -1644,7 +1651,8 @@
               ]
             }
           ],
-          "doLog": true
+          "doLog": true,
+          "randomizationActivatedAt": "2026-10-19T12:00:00.123Z"
         },
         {
           "key": "allocation-default",
@@ -5341,6 +5349,7 @@
       "allocations": [
         {
           "key": "no-log-allocation",
+          "randomizationActivatedAt": "2026-10-19T12:00:00Z",
           "rules": [],
           "splits": [
             {