Skip to content

Consenting to receive a newsletter is not consenting to the pixel that tracks it.

What we measure today, on real emails from 340 brands on the French market.

Open tracking pixel

n=2548 messages measured

93 %

of emails contain one. The open is measured individually, with no action from the reader. 2366 out of 2548.

On campaigns alone: 94% (2358 out of 2521).

By detected sending platform, the share of sends carrying a pixel

Tracking coverage

92 %

of brands track 100% of their emails. And that's a floor: for the rest, emails with no pixel are mostly service messages, not a moderation of marketing tracking.

136 brands out of 148 with at least 5 emails measured

Unsubscribe link

52 %

of brands route their unsubscribe link through a measurement domain. The click that unsubscribes is recorded like any other.

73 brands out of 140 with at least 5 campaigns measured

2548 emails from 340 brands on the French market · run on 10/26/2026 · service messages included in the base, classification being tightened · external observation, measurements replayable from the archive · sending platform read from each send's headers

Beyond the pixel

The same emails, read in full: scores, sending rhythm, sign-up, links, authentication, unsubscribing.

Scores by family

Median score of the brands, out of 100: the share of published criteria their campaigns meet.

  • Compliance 40

    half of brands between 36 and 49

    Share of criteria met, among those we measure against the CNIL recommendation and the rules on email marketing; a finding, not a legal assessment.

  • Deliverability 100

    half of brands between 82 and 100

  • Technical 72

    half of brands between 65 and 87

  • Experience 59

    half of brands between 50 and 69

147 brands with at least 5 measured campaigns · weighting v13 · the method

Campaigns per week

1.4

for the median brand, over the 8 weeks before the survey. Measured on an address that never opens or clicks: a brand that emails its inactive contacts less emails us less.

Brands, by number of campaigns per week

  • 24 < 1
  • 25 1–2
  • 15 2–3
  • 9 3–5
  • 2 ≥ 5

75 brands already followed on Aug. 31, 2026 that have sent at least one campaign since

Links in campaigns

48 %

of brands route all their links through a measurement domain. Clicks are tracked like opens, without the reader seeing it.

  • All links 48 % (71)
  • Some 33 % (48)
  • None 19 % (28)

71 brands out of 147

Published DMARC policy

37 %

reject

of brands ask mailbox providers to reject a message that claims their domain without authenticating it.

  • reject 37 % (55)
  • quarantine 19 % (28)
  • none 44 % (64)

55 brands out of the 147 whose policy was read on receipt

One-click unsubscribe (RFC 8058)

99 %

of brands declare it in their campaign headers. We read the declaration; whether it works is a separate test.

  • Declared 99 % (145)
  • Absent 1 % (2)

145 brands out of 147

Subscription confirmation

18 %

of brands ask for the sign-up to be confirmed before their first campaign. Read from our own sign-ups: an email asking us to confirm, or a campaign that arrives without one.

  • Requested 18 % (50)
  • Not requested 82 % (225)

50 brands out of the 275 whose sign-up received a request or a campaign

French market · campaign rings: brands with at least five measured campaigns, most frequent state across those campaigns · each card names its base · recorded on 10/26/2026 · external observation, measurements replayable from the archive

By sector

The day, the time and the rhythm of campaigns, sector by sector.

Share of campaigns by sending day, as a percentage

  • Fashion & retail

    48 brands

    9 10 18 15 12 19 17
    Campaigns per week
    1.9
    22 brands
    Median time
    10:00
    Pixel on every campaign
    96 %
    46 of 48

  • Beauty

    24 brands

    14 9 17 12 15 16 17
    Campaigns per week
    1.3
    12 brands
    Median time
    11:14
    Pixel on every campaign
    92 %
    22 of 24

  • Home & decor

    24 brands

    15 11 17 9 11 14 23
    Campaigns per week
    1.8
    10 brands
    Median time
    10:22
    Pixel on every campaign
    92 %
    22 of 24

  • Luxury

    18 brands

    9 17 15 13 19 12 15
    Campaigns per week
    1.0
    5 brands
    Median time
    15:26
    Pixel on every campaign
    94 %
    17 of 18

  • High-tech & electronics

    9 brands

    11 27 9 17 16 15 5
    Campaigns per week
    1.3
    6 brands
    Median time
    16:31
    Pixel on every campaign
    78 %
    7 of 9

  • Food

    7 brands

    21 20 17 11 12 11 8
    Campaigns per week
    0.6
    7 brands
    Median time
    12:31
    Pixel on every campaign
    100 %
    7 of 7

French market · brands with at least 5 measured campaigns, sectors with at least 5 of them · day and time in the brand's time zone, each brand counted once · rhythm of the median brand over the 8 weeks before the survey, brands followed beforehand · pixel: brands whose measured campaigns all carry one · recorded on 10/26/2026

Six sending platforms against the recommendation

The same starting point for all six: tracking does not wait for consent.

CNIL recommendation adopted on 12 March, published on 14 April 2026

Publicly documented capabilities, as of Oct. 17, 2026 — the delivered state is observed when we were able to open the screen. Never a compliance claim: responsibility lies with the sender, the ESP acts as a processor.

Brevo ex-Sendinblue France

Per-contact consent is shipped, to be switched on; Brevo then advises not tracking contacts who haven't answered.

  • Delivered: Tracking cutoff (anonymous mode, off by default)
  • Delivered: Per-contact tracking consent confirmed on-account on 23 Aug, off by default: “when this option is disabled, all contacts are tracked” (interface text)
  • Documented, not observed: Native consent collection documented (a checkbox in forms, a revocation link in the email footer), not confirmed on screen
Klaviyo United States

The controls are shipped; without recorded consent, tracking continues.

  • Delivered: Tracking cutoff account-wide and per recipient; cuts the recording: the pixel stays in the email
  • Delivered: Per-contact tracking consent mechanism shipped, default = tracked
  • Documented, not observed: Native consent collection documented at sign-up (a form block), with no native revocation; not confirmed on screen
Adobe Journey Optimizer United States

The toolbox is there; assembly is up to the customer.

  • Delivered: Tracking cutoff per message, the pixel is removed from the email
  • Documented, not observed: Per-contact tracking consent framework to implement
  • Documented, not observed: Native consent collection a landing-page preference center writes the tracking choice to the profile; its effect on the pixel isn't described
Salesforce Marketing Cloud Engagement United States

A per-subscriber Do Not Track preference removes the pixel at send time; the link to consent is still to be assembled.

  • Delivered: Tracking cutoff
  • Documented, not observed: Per-contact tracking consent mechanism documented: pixel and click tracking removed at send time, tracked by default; consent left to assemble
  • Absent: Native consent collection the preference is changed through the application, a preference center, the API or an import; no collection from the recipient is described
HubSpot United States

Tracking consent per email address, in beta and to switch on, collected on the preference page; with no recorded answer, no tracking.

  • Delivered: Tracking cutoff account or email; switched off on an email, the pixel is removed from it
  • Documented, not observed: Per-contact tracking consent documented in beta, to switch on: an address with no status is not tracked; not confirmed on screen
  • Documented, not observed: Native consent collection documented in beta (a checkbox on the preference page, not in forms), not confirmed on screen
Mailjet Sinch France

Each contact can reduce the tracking the account sets, and sending applies their choice; with no choice recorded, the account decides.

  • Delivered: Tracking cutoff (whole account, transactional and campaign emails together)
  • Delivered: Anonymous tracking depending on the account's configuration
  • Documented, not observed: Per-contact tracking consent documented: a contact's preference can only reduce the account's tracking; not confirmed on screen
  • Documented, not observed: Native consent collection documented (a form field, a preference page hosted by the vendor), not confirmed on screen

Three tiers: consent collection and enforcement documented, to switch on or configure (Brevo, Mailjet, and HubSpot in beta) · collection at sign-up, with no native revocation, that stops the recording and leaves the pixel (Klaviyo) · controls shipped, to connect yourself to the consent you collect (Adobe, Salesforce). In every case, the switch is an action the customer takes — never the out-of-the-box state.

And once tracking is switched off, does the pixel really disappear? At Mailjet and Adobe, yes: it's no longer added to the email. At Salesforce too, for a subscriber set to Do Not Track, and at HubSpot, when switched off on an email; switched off for the account or for lack of consent (beta), HubSpot says nothing is tracked, without saying whether the image goes. At Klaviyo and Brevo (anonymous mode), it stays in the message and loads on every open — Klaviyo discards the request, Brevo counts it without linking it to the recipient. In these last two cases, the reader's device keeps calling the platform's servers.

And in your own systems?

These figures are the market's. Not yours.

You state your purposes, your legal basis, and your settings. We send back what the recommendation means for you, point by point.

Take the diagnostic

postmetry.com/en/pixels-cnil/diagnostic/

Postmetry · publicly documented capabilities, as of Oct. 17, 2026 · corpus snapshot as of Oct. 26, 2026, French market · deterministic measurements, replayable from the raw archive · no data drawn from customer areas postmetry.com