Safari ITP ecommerce tracking problems are often blamed on "iOS privacy" in general, and the details get lost. Intelligent Tracking Prevention (ITP) is a specific set of rules in WebKit, the engine behind Safari. Some of those rules change how long cookies and click IDs survive, and that matters when a shopper clicks an ad on Monday and buys on Friday.
This post sticks to what WebKit documents, links the sources, and then explains where server-side tracking helps and where it does not. Where we are not sure of a detail, we say so or leave it out.
ITP and App Tracking Transparency are different things
People mix these up in iOS tracking ecommerce discussions. They are separate:
- App Tracking Transparency (ATT) is a permission prompt for apps. It governs tracking inside iOS apps.
- Intelligent Tracking Prevention (ITP) applies to the web, in Safari and other WebKit-based browsers. It is what affects your store when someone browses it on an iPhone, iPad or Mac.
Historically every browser on iPhone has been built on WebKit, so ITP rules shaped the whole iOS web experience. That is changing in some regions, so check Apple's current rules for your market before you assume.
What WebKit documents
WebKit publishes its behaviour on its Tracking Prevention in WebKit page. These are the parts that touch ecommerce measurement.
| WebKit behaviour | What it means for a store |
|---|---|
| Third-party cookies are fully blocked. | A pixel that depends on cookies from the ad platform's own domain cannot read them in Safari. |
| Link decoration detection: when click IDs are added to links and read by script, cookies created in JavaScript on the landing page are capped to 24 hours. | A click ID stored by a script can be gone a day later. A purchase after that has nothing stored to match against. |
| All cookies created in JavaScript and other script-writeable storage (such as LocalStorage and IndexedDB) are deleted after 7 days of no user interaction with the site. | A shopper who clicks an ad and returns after more than a week of not visiting loses script-written cookies. |
| Cookies set in an HTTP response by third-party CNAME cloaking or third-party IP address cloaking are capped to 7 days. | Pointing a subdomain at a tracking vendor does not give you a longer-lived cookie in Safari. |
| Third-party referrers are downgraded to the origin. | Query strings in the referrer are not passed on. |
One more item is specific to Safari's Private Browsing. In the release notes for Safari 17.0, WebKit lists blocking of known tracking query parameters in links, blocking of known trackers and fingerprinting noise as Private Browsing protections. See WebKit features in Safari 17.0. That is narrower than a blanket statement that Safari strips every click ID, and we do not claim the broader version.
First-party cookies are not permanent cookies
A common assumption is that a first-party cookie is safe from ITP. The WebKit page is more precise. The 7-day cap applies to cookies created in JavaScript and to other script-writeable storage. The 24-hour cap applies to JavaScript-created cookies on a landing page where ITP detects link decoration. A cookie your own server sets in an HTTP response is treated differently, with the CNAME and IP-cloaking exception above.
That has a practical consequence for tracking scripts. Meta's _fbp cookie, written by a script, is a script-created first-party cookie. So is the first-party cookie SyncBeacon's own browser tag writes. Both fall under these caps in Safari. We would rather say that plainly than imply the tag escapes ITP. It does not.
What this means for an ecommerce funnel
Think about a typical journey on an iPhone:
- A shopper taps a Meta or Google ad. The URL carries a click ID.
- They browse, leave, and come back directly a few days later.
- They buy. Your store creates an order.
In a browser-only setup, step 3 depends on what the script stored at step 1. If the click ID cookie was capped at 24 hours, or the storage was cleared after seven days without interaction, the purchase arrives without it. The ad platform may then fail to link the order to the ad. The order is not lost in your shop. It is orphaned in the ad platform.
What server-side tracking iOS setups change
Server-side tracking iOS setups change two things, and leave a third alone.
1. The purchase no longer depends on the browser
SyncBeacon takes purchases from your store's orders (Shopify order webhooks, WooCommerce webhooks, Shopware order sync), so ITP does not touch whether the purchase is reported. See the Shopify setup post for one example.
2. Click IDs can be remembered on the server
The browser tag captures click IDs (such as fbclid, gclid and wbraid) when the visitor lands, and SyncBeacon remembers them between the visit and the order. How long depends on your plan's click-ID memory, and this is skipped when marketing consent is denied. The point is that the server copy does not expire when Safari clears the browser copy.
3. What it does not change
- The click ID still has to be captured while the visitor is on the landing page. If the browser or a Private Browsing rule removes the parameter before the script sees it, there is nothing to remember.
- The tag's own first-party cookie and storage are still subject to the caps above.
- Consent still applies. If a visitor explicitly denies marketing consent, the event is stored for your records but not forwarded.
- Matching a later order to an earlier visit may still rely on a customer identifier such as a hashed email. When none is available, a match may fail.
Three myths worth clearing up
- "ITP blocks first-party cookies." It does not. It limits how long script-created cookies and other script-writeable storage live, and WebKit's page describes the exact conditions.
- "Server-side tracking bypasses ITP." It does not. It moves the purchase report off the browser and keeps a server-side copy of what it captured. The browser rules are unchanged.
- "The iOS 14.5 prompt and Safari limits are one change." They are different mechanisms, covering apps and the web respectively. Opt-in rate statistics for the app prompt say nothing about how many of your web orders are affected by ITP, and we do not quote any here.
How to check your own store
- Test on a real iPhone in Safari. Click a test ad link with a click ID, then check what the SyncBeacon Conversions page records for the order.
- Compare conversion volume by device and browser in your analytics with orders in your store. A larger gap on Safari points to ITP effects.
- Look at the ad platform's match quality and diagnostics, which show whether events arrive with usable identifiers. Details in the Meta Conversions API page.
- Remember that attribution windows also make platform totals differ from your store. See why Meta shows fewer conversions than Shopify.
A simple Safari test you can run today
- On an iPhone, open a test ad link that carries a click ID and land on your store. Browse one product page.
- Close Safari without buying. Come back to the store directly one or two days later and place a small order.
- Check the order in the SyncBeacon Conversions page, then check in the ad platform whether the purchase was matched to the ad.
- Repeat with a click ID the same day. If only the same-day order matches, you are seeing the effect of the storage caps on your own store.
Related reading
- Ad blockers, conversion tracking and ROAS
- Why Google Ads conversions go invisible on iOS and with ad blockers
- Google Ads Enhanced Conversions
- Ecommerce conversion tracking overview
Sources
- WebKit, Tracking Prevention in WebKit. webkit.org/tracking-prevention
- WebKit, WebKit Features in Safari 17.0. webkit.org/blog/14445
Get started
Close your conversion gap today
Connect Meta CAPI, Google Enhanced Conversions, and more - in one setup, no developer required.
Browser behaviour changes with new releases. We summarised WebKit's documentation as published when this post was written. Check the sources for the current rules. Not legal advice.