Noxtica is named for identification only; no affiliation or endorsement is implied. Results are what an automated browser observed at one point in time, from one vantage point; they are not legal advice and not a legal determination.
Data as of ·How Radar measures
Scan details
- Browser
- Chrome (Chromium)153.0.8010.12
- Engine
- blink
- Operating system
- linux
- Region
- tr-workstation
- Viewport
- 1440x900 (desktop)
- Method
- NR-v0.3.0
- Scanned
- Snapshot
- NR-SN-01M7FX2EKD73EXQRH6Q6TQS6H1
How the page behaved before and after each consent choice.
How to read it: Each card compares one choice with the state before any choice. Values are observations from this scan.
Part 1 of 10
Consent impact
Observed activity in three consent states of the same page. Each state was measured in separate runs with a fresh browser profile. The numbers describe the measured environment named in this section and nothing beyond it.
Matched environment: Chrome (Chromium) 153.0.8010.12 · tr-workstation · fresh profile per run · 3 repeats · method NR-v0.3.0
Protocol: Before choice, Reject all, Accept all, GPC enabled; 3 repeats; dwell 10 s; viewport 1440x900. Each state was measured in its own fresh-profile runs, not in one session. Methodology: consent comparison
How the three states were measured
Step through one scan: each state starts from its own fresh profile.
Stage 1 of 5: Fresh profile
Every situation starts in a new browser profile of its own and loads the same entry page, so no cookie, storage or consent state carries over from one situation to the next.
One fresh browser profile branches into four situations, each run in a new profile of its own. Before choice: no consent action, the whole observation window is counted. Reject all: the most comprehensive rejection control is used, and only events after the action settled are counted. Accept all: the most comprehensive acceptance control is used, and only events after the action settled are counted. GPC enabled: the Global Privacy Control signal is sent with no consent action, and the whole window is counted. Events before a consent action are reported separately and never added.
Part 2 of 10
The three consent states
One card per state: how many runs counted, the mean of five metrics with the change from , what the state means and which events its values count.
Before choice
3 of 3 runs validMean per metric over valid runs
- Signal families
- 13
- API events
- 17,873
- Scripts
- 18.67
- Destinations
- 3
- Data-flow-linked transmissions
- 0
Start at navigation; take no consent action; observe for a fixed window.
- Values count
- Whole observation window
Reject all
3 of 3 runs validMean per metric, with the change from Before choice
- Signal families
- 8
- ▼−5 lower than Before choice
- API events
- 10,571.67
- ▼−7,301.3 lower than Before choice
- Scripts
- 11
- ▼−7.7 lower than Before choice
- Destinations
- 1.67
- ▼−1.3 lower than Before choice
- Data-flow-linked transmissions
- 0
- =0 the same as Before choice
Fresh profile; detect the consent interface; invoke the most comprehensive rejection control; observe after the action.
- Values count
- Events after the consent action only
- Before the action (reported separately)
- 4,507.67 API events, mean
Accept all
3 of 3 runs validMean per metric, with the change from Before choice
- Signal families
- 19
- ▲+6 higher than Before choice
- API events
- 12,008
- ▼−5,865 lower than Before choice
- Scripts
- 13
- ▼−5.7 lower than Before choice
- Destinations
- 3
- =0 the same as Before choice
- Data-flow-linked transmissions
- 0
- =0 the same as Before choice
Fresh profile; detect the consent interface; invoke the most comprehensive acceptance control; observe after the action.
- Values count
- Events after the consent action only
- Before the action (reported separately)
- 4,495.67 API events, mean
Part 3 of 10
Stage comparison
The same four counts in every scenario, side by side, each with the number of valid runs behind it.
| Scenario | State | Valid runs | Requests | Third-party requests | External registrable domains | Third-party cookie keys | Signal families observed |
|---|---|---|---|---|---|---|---|
| Before choice | Measured | 3 of 3 valid runs | 46 | 1 | 1 | 0 | Canvas, Navigator / device, Hardware (CPU, memory), Screen / window, Fonts, Permissions, Storage, Timing / performance, CSS / media capabilities, Client Hints, Network information, Plugins and MIME types, Math / engine behaviour |
| Reject all | Measured | 3 of 3 valid runs | 8 | 0 | 0 | 0 | Canvas, Navigator / device, Screen / window, Fonts, Storage, Timing / performance, CSS / media capabilities, Math / engine behaviour |
| Accept all | Measured | 3 of 3 valid runs | 17 | 0 | 0 | 0 | Canvas, WebGL, Audio, Navigator / device, Hardware (CPU, memory), Screen / window, Fonts, Media devices and codecs, Sensors, Permissions, Storage, Timing / performance, CSS / media capabilities, Client Hints, Locale / time zone, Network information, Battery, Plugins and MIME types, Math / engine behaviour |
| GPC enabled | Measured | 3 of 3 valid runs | 46 | 1 | 1 | 0 | Canvas, Navigator / device, Hardware (CPU, memory), Screen / window, Fonts, Permissions, Storage, Timing / performance, CSS / media capabilities, Client Hints, Network information, Plugins and MIME types, Math / engine behaviour |
Before choice: 46 requests, 1 external domain, 3 of 3 runs valid; Reject all: 8 requests, 0 external domains, 3 of 3 runs valid; Accept all: 17 requests, 0 external domains, 3 of 3 runs valid; GPC enabled: 46 requests, 1 external domain, 3 of 3 runs valid.
Aggregation: median over the valid runs of the scenario, rounded up. The signal families of Reject all and Accept all are those with an event after the consent action.
Part 4 of 10
Signal families by consent state
Which (kinds of browser capability) the page used in each state, as mean API events per family. Reject all and Accept all count the events after the consent action only.
Showing 19 of 19 signal families. Changed behaviors are the families this scan marks as different between Before choice and Reject all.
Data table
| Row | Before choice | Reject all | Reject all minus Before choice | Accept all | Accept all minus Before choice |
|---|---|---|---|---|---|
| Canvas | 11,584.67 | 7,829.33 | -3,755.33 | 8,136 | -3,448.67 |
| WebGL | 0 | 0 | 0 | 162 | +162 |
| Audio | 0 | 0 | 0 | 16 | +16 |
| Navigator / device | 65.33 | 1 | -64.33 | 68 | +2.67 |
| Hardware (CPU, memory) | 9.33 | 0 | -9.33 | 11 | +1.67 |
| Screen / window | 92 | 19.33 | -72.67 | 47.33 | -44.67 |
| Fonts | 20 | 1 | -19 | 293 | +273 |
| Media devices and codecs | 0 | 0 | 0 | 30 | +30 |
| Sensors | 0 | 0 | 0 | 1 | +1 |
| Permissions | 2.33 | 0 | -2.33 | 19 | +16.67 |
| Storage | 45.33 | 7.67 | -37.67 | 58 | +12.67 |
| Timing / performance | 263.33 | 25.67 | -237.67 | 300 | +36.67 |
| CSS / media capabilities | 52.67 | 7.67 | -45 | 96 | +43.33 |
| Client Hints | 14 | 0 | -14 | 17 | +3 |
| Locale / time zone | 0 | 0 | 0 | 13 | +13 |
| Network information | 4.67 | 0 | -4.67 | 23 | +18.33 |
| Battery | 0 | 0 | 0 | 11 | +11 |
| Plugins and MIME types | 9.33 | 0 | -9.33 | 6 | -3.33 |
| Math / engine behaviour | 5,710 | 2,680 | -3,030 | 2,700.67 | -3,009.33 |
Select a value in the chart to see the family’s exact values and evidence links.
Part 5 of 10
Metric comparison
The five headline metrics and four more, as means over valid runs with the range across repeats, the change from Before choice and a link to the evidence behind each value.
Data table
| Metric | Before choice | Reject all | Accept all | Reject all minus Before choice | Accept all minus Before choice |
|---|---|---|---|---|---|
| Signal families | 13 | 8 | 19 | -5 | +6 |
| API events | 17,873 | 10,571.67 | 12,008 | -7,301.33 | -5,865 |
| Scripts | 18.67 | 11 | 13 | -7.67 | -5.67 |
| Destinations | 3 | 1.67 | 3 | -1.33 | 0 |
| Data-flow-linked transmissions | 0 | 0 | 0 | 0 | 0 |
Values are means over valid runs; the range across repeats is shown next to each value. “View evidence” opens the first valid repeat of that state in the evidence explorer, filtered to what the metric counts. Every repeat is listed in the run table.
| Metric | Before choice | Reject all | Accept all |
|---|---|---|---|
| Signal families | |||
| API events | |||
| Scripts | |||
| Destinations | |||
| Data-flow-linked transmissions |
More measured metrics (4)
| Metric | Before choice | Reject all | Accept all |
|---|---|---|---|
| API surfaces | |||
| Third-party destinations | |||
| Correlated transmissions | |||
| Fingerprint-like behaviors |
Part 6 of 10
Statements
Plain sentences generated from the numbers above, each with the runs behind it.
Reject all control labelled “Reject all”; Accept all control labelled “Accept all”
7,301.3 fewer API events after Reject all than before choice (mean of 3 valid runs each)
5 fewer signal families after Reject all than before choice (mean of 3 valid runs each)
No data-flow-linked transmissions were observed after Reject all
Before the Reject all action, 4,507.7 API events were recorded (mean of 3 valid runs)
5,865 fewer API events after Accept all than before choice (mean of 3 valid runs each)
6 more signal families after Accept all than before choice (mean of 3 valid runs each)
No data-flow-linked transmissions were observed after Accept all
Before the Accept all action, 4,495.7 API events were recorded (mean of 3 valid runs)
Part 7 of 10
Consent interaction
How the consent interface was operated in each state: the consent management platform (CMP), the control that was clicked, how many layers were opened and what happened in every run. Outcomes are listed as recorded; a missing interface, a failed interaction and a reappearing banner are different states and are never combined.
Before choice
- Consent management platform
- Interface detected, platform not identified
- Control label clicked
- None: no consent action in this scenario
- Layers traversed
- No control invoked
- Banner reappeared flag
- Not set in any run
- Outcome per run
not_applicable3 of 3 runsNo consent action was part of this scenario.
Start at navigation; take no consent action; observe for a fixed window.
- Before the consent action (Repeat 1)
- End of the observation window (Repeat 1)
Reject all
- Consent management platform
- Interface detected, platform not identified
- Control label clicked
Reject all
- Layers traversed
- 0 (first layer)
- Banner reappeared flag
- Not set in any run
- Outcome per run
action_succeeded3 of 3 runsThe control was invoked and the consent interface closed.
Fresh profile; detect the consent interface; invoke the most comprehensive rejection control; observe after the action.
- Before the consent action (Repeat 1)
- After the consent action (Repeat 1)
- End of the observation window (Repeat 1)
Accept all
- Consent management platform
- Interface detected, platform not identified
- Control label clicked
Accept all
- Layers traversed
- 0 (first layer)
- Banner reappeared flag
- Not set in any run
- Outcome per run
action_succeeded3 of 3 runsThe control was invoked and the consent interface closed.
Fresh profile; detect the consent interface; invoke the most comprehensive acceptance control; observe after the action.
- Before the consent action (Repeat 1)
- After the consent action (Repeat 1)
- End of the observation window (Repeat 1)
Part 8 of 10
GPC experiment
A separate experiment: the same page loaded with the signal sent, compared with Before choice. It reports observed network behavior only.
GPC was sent in the GPC enabled runs of this scan (Sec-GPC: 1 and navigator.globalPrivacyControl).
Fresh profile; send the Global Privacy Control signal (Sec-GPC: 1 and navigator.globalPrivacyControl); take no consent action. Measured in Chrome (Chromium) 153.0.8010.12 · tr-workstation · fresh profile per run · 3 repeats · method NR-v0.3.0.
No difference in the measured network behavior between the baseline runs and the GPC enabled runs, for the metrics below.
| Metric | Baseline (Before choice, GPC not sent) | GPC enabled | Difference |
|---|---|---|---|
| Signal families | 13 | 13 | 0 |
| API surfaces | 64.33 | 64 | -0.33 |
| API events | 17,873 | 14,823.67 | -3,049.33 |
| Scripts | 18.67 | 15 | -3.67 |
| Destinations | 3 | 3 | 0 |
| Third-party destinations | 1 | 1 | 0 |
| Data-flow-linked transmissions | 0 | 0 | 0 |
| Correlated transmissions | 9.67 | 8 | -1.67 |
| Fingerprint-like behaviors | 5 | 3 | -2 |
Data table
| State | Value | Change versus Baseline |
|---|---|---|
| Baseline | 13 | Baseline |
| GPC enabled | 13 | 0 |
Data table
| State | Value | Change versus Baseline |
|---|---|---|
| Baseline | 64.33 | Baseline |
| GPC enabled | 64 | -0.33 |
Data table
| State | Value | Change versus Baseline |
|---|---|---|
| Baseline | 17,873 | Baseline |
| GPC enabled | 14,823.67 | -3,049.33 |
Data table
| State | Value | Change versus Baseline |
|---|---|---|
| Baseline | 18.67 | Baseline |
| GPC enabled | 15 | -3.67 |
Data table
| State | Value | Change versus Baseline |
|---|---|---|
| Baseline | 3 | Baseline |
| GPC enabled | 3 | 0 |
Data table
| State | Value | Change versus Baseline |
|---|---|---|
| Baseline | 1 | Baseline |
| GPC enabled | 1 | 0 |
Data table
| State | Value | Change versus Baseline |
|---|---|---|
| Baseline | 0 | Baseline |
| GPC enabled | 0 | 0 |
Data table
| State | Value | Change versus Baseline |
|---|---|---|
| Baseline | 9.67 | Baseline |
| GPC enabled | 8 | -1.67 |
Data table
| State | Value | Change versus Baseline |
|---|---|---|
| Baseline | 5 | Baseline |
| GPC enabled | 3 | -2 |
The GPC signal was sent in 3 valid GPC enabled runs
Destinations: 3 without GPC, 3 with GPC (mean of 3 valid runs each)
Third-party destinations: 1 without GPC, 1 with GPC (mean of 3 valid runs each)
Data-flow-linked transmissions: 0 without GPC, 0 with GPC (mean of 3 valid runs each)
Correlated transmissions: 9.7 without GPC, 8 with GPC (mean of 3 valid runs each)
No observable change in network behaviour was measured with GPC enabled (measured traffic only, not a legal conclusion)
Mean differences within run-to-run variation, because the valid runs overlap: correlated transmissions (9.7 without GPC, 8 with GPC)
Part 9 of 10
Runs behind these values
Every scenario run of this scan, with its status and reason. Only runs marked Included enter a comparison; an excluded run is shown here and never replaced by another.
Runs at a glance
One square per run: filled when the run is included in comparisons, hollow with a cross when it is excluded.
- Before choice3 of 3 runs included
- Reject all3 of 3 runs included
- Accept all3 of 3 runs included
- GPC enabled3 of 3 runs included
| Scenario and repeat | Status and reason | In comparisons | Consent outcome | API events | Run |
|---|---|---|---|---|---|
| Before choice Repeat 1 | Valid | Included | not_applicable | 23,910 | NR-RN-01M78Z5ARQ2P86S0AW6AP8JJ5Q |
| Before choice Repeat 2 | Valid | Included | not_applicable | 14,855 | NR-RN-01M78Z7TVTT8NMA8TW1ZFDGDHG |
| Before choice Repeat 3 | Valid | Included | not_applicable | 14,854 | NR-RN-01M78Z9QG8Y7JEDJ9404N60ZHE |
| Reject all Repeat 1 | Valid | Included | action_succeeded | 15,052 | NR-RN-01M78Z6BT3XJQQM3RDE2FCECR5 |
| Reject all Repeat 2 | Valid | Included | action_succeeded | 15,101 | NR-RN-01M78Z89RKA65XA48NBHCY55JK |
| Reject all Repeat 3 | Valid | Included | action_succeeded | 15,085 | NR-RN-01M78ZA6C6HHSC8JNVX4X80GTQ |
| Accept all Repeat 1 | Valid | Included | action_succeeded | 16,530 | NR-RN-01M78Z6WJEPW9RTEC9W7MV253W |
| Accept all Repeat 2 | Valid | Included | action_succeeded | 16,511 | NR-RN-01M78Z8S5SBZBBRKR8W1RQH63V |
| Accept all Repeat 3 | Valid | Included | action_succeeded | 16,470 | NR-RN-01M78ZANS0BWBBKCZH029R94TP |
| GPC enabled Repeat 1 | Valid | Included | not_applicable | 14,860 | NR-RN-01M78Z7BYZCH29A3ZCZP2A3GQF |
| GPC enabled Repeat 2 | Valid | Included | not_applicable | 14,767 | NR-RN-01M78Z98NA6H9N1S9KQ896GCFJ |
| GPC enabled Repeat 3 | Valid | Included | not_applicable | 14,844 | NR-RN-01M78ZB54YHMFR8FVKGCX489MQ |
Part 10 of 10
Limitations and method
What this scan can and cannot say, and where each definition is written down.
Limitations of this scan (16)
- Measured in one browser, region and viewport with a fresh profile per run; the numbers describe that measured environment only.
- The entry URL was loaded in every run; there was no sign-in and no form was submitted. Extra pages were loaded as described below.
- Interaction probes typed made-up values with key events into at most 5 email, telephone and text fields of the entry page and never submitted them: no Enter key was pressed and nothing was clicked after typing. Password, payment, search, one-time-code, hidden, file, disabled, read-only and cross-origin-frame fields were skipped.
- The extra-page visits loaded up to 4 same-site pages after the entry page, in the Before choice scenario only. robots.txt was checked for each URL before it was loaded, and a disallowed URL was not loaded. Page loads were at least 5 seconds apart. The requests of an extra page are listed with that page and are not part of the scenario counts.
- The policy-page check read the visible text of one policy page among the extra pages in memory (at most 1 MiB), kept no page text and compared organisation names of the editorial entity map with the third-party organisations observed; it says nothing about what a policy promises.
- An observed API access shows that the page called the API, not what the page used the result for.
- Correlated links show timing and script association only; data-flow-linked links show a value found in an outgoing request or storage write.
- The document policy section reads the response headers of the entry document only. It states what the response declared; it does not test that the browser enforced a policy or that the site follows it.
- TLS facts are passive: the negotiated protocol and the certificate validity dates the browser itself reported. No connection was opened to test the server.
- Links, forms and embedded destinations are declarations read from the page DOM at two checkpoints, without any interaction. A declaration is not a request: only a matching request row shows that a destination was requested. At most 256 links, 128 forms and 128 embedded destinations are listed per checkpoint; the rest are counted.
- The page hooks run in documents only: dedicated and shared workers are outside them and service workers are blocked by the scan profile, so activity there is not observable.
- The storage inventory lists the keys the entry page wrote during the observation window. A set event is an attempt or an observation: the acceptance column says whether the browser cookie jar held the cookie at the end of the run.
- Candidate identifier sharing compares cookie values with the request URLs of the same run in browser memory and keeps no value. The candidate set is unvalidated and is not the published ID-cookie validation; a match can be accidental, and what a server does with a value is not observed.
- DNS alias observations list the alias chain the resolver of the scan reported for hosts the egress proxy allowed by name. The lookup is best effort and never changed a connection; an alias is an observation, not proof of cloaking, tracking or ownership.
- This scan was run from a single workstation with direct internet egress: no attested network boundary was used and the measurement reflects one network location only.
- The compliance relevance checks are technical observations mapped to published rules; they are not legal advice and not a legal determination.