Safari will soon limit script-set cookies to one day when a visitor arrives from a tracking domain through a link carrying identifiers, as ad clicks do. Expect fewer credited iPhone conversions, and read the gap before cutting ads.

A second change in two months

A month ago, Safari 12.1 and iOS 12.2 began limiting cookies set by scripts to seven days. Analytics on iPhones started counting some returning visitors as new.

Yesterday, Apple's WebKit team announced the next step, Intelligent Tracking Prevention 2.2. In one specific situation, the limit drops to a single day.

The situation is a visitor arriving from a site Safari considers a tracker, through a link with extra information added to its address. That is how most ad clicks arrive.

It comes with the next Safari updates, so there is time to understand it before the reports shift.

The conditions

The second is the hard one to predict. Safari classifies trackers on each device, from how often a domain appears in the background of other sites, and the large advertising and social platforms are obvious candidates.

Why ad clicks fit the description

Advertising platforms add identifiers to the address of the page an ad sends people to. Google Ads adds one called gclid when auto-tagging is on, and Facebook adds its own to links people follow off the site.

That is the extra information after the question mark. It is how the platform connects a later enquiry to the click that started it.

Analytics, ad conversion tracking and remarketing tags then set their cookies with scripts on the landing page, which is exactly the kind of cookie the new rule limits.

The campaign tags you add to links yourself count as extra information too, so a tagged link in a social post can meet the same condition.

On a second page within the site, the seven-day limit generally applies again. So the visitors most affected are the ones who read one page and leave, which describes a great many ad clicks.

What it does to the numbers

Take a visitor who clicks an ad on an iPhone, reads the landing page, and comes back three days later by typing your address. To Safari's cookie rules, the first visit has been forgotten.

Analytics counts a new visitor arriving directly. If they enquire, the enquiry is credited to direct traffic, and the ad that brought them gets nothing.

The ad platform's own conversion count misses the enquiry too, since the cookie that linked it to the click has gone.

Remarketing lists that rely on those cookies shrink for Safari users as well. Nothing about the customers has changed; only the record of them has.

A worked example

A company that builds backyard sheds and garden offices advertises on Facebook and Google from late winter, when people start planning summer projects.

Its customers almost never buy on the first visit. They click an ad on a phone in the evening, look at the models, and come back days or weeks later to request a quote.

Since early April, both ad dashboards had shown fewer quote requests than the inbox did, much of the gap on iPhones. Much of that was the seven-day limit at work.

The Time Lag report in Analytics showed most quote requests arriving several days after the first visit. Under the one-day rule, many of those from iPhones will lose their link to the ad entirely.

The owner kept the campaigns running and changed the landing pages instead, so the first visit could produce something.

Each now carries a short quote form, a tappable phone number, and an offer to email the full price list. Somebody not ready for a quote can still leave an address on day one.

That request is recorded the same day, whatever Safari does to cookies afterwards, and the company can follow it up directly.

Some of the gap was always there

It is worth keeping this in proportion. A visitor who clicks on a phone and enquires on a laptop was never connected by cookies, and neither was somebody who cleared their browser.

Ad reports have always undercounted the slow, considered purchase. Safari's changes widen the gap and concentrate it on one browser, which is what makes it visible.

Reading the gap

Two reports say how much this matters. The Time Lag report, under Conversions in Universal Analytics, shows how many days pass between a first visit and an enquiry.

If most enquiries arrive the same day, little changes. If most take a week, a large share of your Safari conversions from ads is about to go uncredited.

The browser report under Audience shows how large that share is. Compare enquiries credited to ads with your own monthly count, and read a fall on Safari as lost records before lost customers.

Give the first visit something to keep

One change to the site itself helps. The less your advertising depends on a second visit, the less any cookie rule matters.

Put a short enquiry form and a tappable phone number on the landing page itself, not two clicks away. For considered purchases, offer something worth an email address, such as a price list or a buying guide.

An enquiry made during the first visit is credited to the click that brought it. An address collected then lets the business follow up without relying on any browser to remember anybody.

The counter-case

Not every business will notice this.

A business that does not advertise online loses only some accuracy in returning visitor counts. One whose customers act within the same visit, like somebody calling a locksmith from the ad, is barely touched.

The businesses most affected are those with long consideration and paid traffic: renovations, furniture, vehicles, courses, anything people think about for days.

For them, the answer is to read the gap rather than react to it, and to let the first visit produce something worth recording.

Before the reports change

  1. Open the Time Lag report in Analytics.
  2. Note how many enquiries come after day one.
  3. Check what share of visits comes from Safari.
  4. Put the enquiry form on the landing page.
  5. Offer something worth an email address.
  6. Compare ad reports with your own count.
  7. Hold off cutting ads on reported conversions alone.

Step four is the one that takes cookies out of the question. An enquiry sent from the page the ad opened is recorded while the click is still remembered, in every browser.

The first of these changes, and what it did to returning visitors, is covered in Safari is about to shorten your cookies.

Frequently asked questions

What did Apple announce this week?

A further Safari change: when a visitor arrives from a domain classified as a tracker, through a link carrying identifiers, cookies set by scripts on that landing page last only one day.

Why does this affect advertising in particular?

Because ad clicks arrive exactly that way. Ad platforms add a click identifier to the address, and analytics and conversion tracking set their cookies with scripts on the landing page.

Which visitors are most affected?

Safari users who click an ad, read one page and leave, then return more than a day later. On their return they look like new visitors from somewhere else.

How do I know whether my business is affected?

Open the Time Lag report under Conversions in Universal Analytics. If most enquiries arrive the same day, little changes; if they take days, the gap on Safari will widen.

Should I cut ads that report fewer iPhone conversions?

Not on that evidence alone. Reported conversions from Safari will fall even when nothing about the customers has changed. Check them first against the enquiries you count yourself.

How can a business reduce the effect?

Let the first visit produce something: a short enquiry form and a tappable phone number on the landing page, and an offer worth an email address for slower decisions.

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.

Ad reports and the inbox disagreeing?

Send us the last three months of both. We will show you where the gap comes from and how to read it.

Start a Conversation