Safari will soon cap cookies set by scripts at seven days, so Analytics will count Safari visitors who return after a week as new, and ad conversions after a week will go unlinked. Read the numbers accordingly.

Seven days

Apple's WebKit team, which builds the engine behind Safari, announced yesterday that the next versions of Safari will limit cookies created by scripts on a page to seven days.

The change is called Intelligent Tracking Prevention 2.1, or ITP 2.1. It will reach Safari on Macs, iPhones and iPads with a routine software update. Apple has not said exactly when.

Most of the attention has gone to advertisers. But the tool most affected on a small business site is the one counting your visitors.

Nothing changes in your reports today. What changes is how to read them once the update arrives.

How Analytics recognises a returning visitor

Google Analytics has no idea who anybody is. On a first visit, its script stores a random ID in a small cookie in the visitor's browser.

When the same browser returns, the script finds the ID and counts a returning visitor rather than a new one.

That cookie is created by the script on the page, which is exactly the kind Safari will cap. It normally lasts two years and is renewed on every visit.

Under the new rule, the seven days will run from the last visit. A Safari visitor who comes back within a week is recognised; one who comes back after ten days is a stranger again.

What will change in the reports

The fourth is the one with money attached. It changes how well your advertising appears to work, and automated bidding learns from whatever it is told.

Why a week matters

Plenty of small business purchases take longer than a week to decide.

A kitchen renovation, a wedding venue, a set of dental implants, a used car: people visit, leave, compare, and come back weeks later to enquire.

After the update, every one of those return visits from Safari will look like somebody arriving for the first time.

So will regular customers who come back every month or two to book again or reorder. Loyal visitors will look like strangers each time.

For a business whose customers decide the same afternoon, such as a takeout order or an emergency plumber, the week hardly matters.

What happens to ad attribution

When somebody clicks an ad, the click carries an identifier to your site. The advertising tag on the page stores it in a cookie, created by a script.

If that person later sends an enquiry, the tag finds the stored click and credits the ad with the conversion.

After the update, that cookie will last seven days in Safari. An iPhone visitor who clicks an ad and enquires twelve days later still enquires, but the ad may no longer get the credit.

Reports will show fewer conversions from Safari than really happened, and any bidding that optimises towards conversions will learn from the shortened record.

A worked example

A kitchen and bath showroom in a light industrial area sells to people who take weeks to decide, and it runs search ads to bring them in.

This morning its owner looked up two things: how many phone visits come from iPhones, and how long customers usually take between a first visit and a quote request.

Most phone visits came from iPhones. And the quote requests credited to the ads often arrived two or three weeks after the first click, from somebody coming back to compare finishes.

After the update, many of those returns will look like first visits, and the ads will be credited with fewer of the quotes they started.

The showroom settled on three things: annotate the date the update ships, read ad results by device rather than in total, and add a question to the quote form asking how the customer first heard of them.

None of that stops the change. It stops the change being mistaken for the ads getting worse.

Reading the numbers afterwards

Universal Analytics lets you add an annotation to the timeline. Add one the day the update arrives, so nobody reads the jump in new visitors as a sudden rush of strangers.

Compare Safari with Safari, and Chrome with Chrome. A segment for each browser keeps the change in one place.

Lean on sessions and on enquiries rather than on users and returning visitors. Sessions are counted the same way before and after; user counts are not.

When this year is compared with last, expect the share of new visitors to look higher from the update onwards, and say so in any report that shows it.

Count what does not depend on cookies

The most reliable figures on a small business site were never cookie-based: form enquiries, phone calls, bookings, orders.

Count those directly, every month. Ask new customers how they first heard of you, and write the answer down.

A business that already does this will barely notice the update. A business that judges its ads only by the ad platform's own conversion count will notice a great deal.

The counter-case

For some sites, this is a footnote.

A business whose customers decide within a day, or whose visitors mostly use Chrome on Android or Windows, will see little change. The browser report in Analytics, under Audience and then Technology, answers that in a minute.

Nor is there a clever workaround worth chasing first. Cookies set by your own server are not covered by this change, but moving analytics onto them is a technical project, and the first job is reading the reports correctly.

Annotate, segment by browser, and count enquiries directly: that covers most of what this change means for a small business.

Before the update

  1. Check your Safari share in the browser report.
  2. Note how long customers take to decide.
  3. Create a segment for each main browser.
  4. Annotate the date the update ships.
  5. Read ad results by device.
  6. Count enquiries and calls directly each month.
  7. Ask new customers where they heard of you.

Step two is the one that tells you how much any of this matters. A business whose customers decide in an afternoon can relax; one whose customers take a month needs every other step.

Another reason the visitor counts can mislead is covered in referral spam in your analytics reports.

Frequently asked questions

What is Safari changing?

Cookies created by scripts on a page will last seven days at most. Apple announced it this week for the next Safari updates on Macs, iPhones and iPads, without a date yet.

How does this affect Google Analytics?

Analytics recognises returning visitors through a cookie its script creates. In Safari, a visitor who stays away longer than a week will be counted as new.

Will my visitor numbers look higher?

User counts may rise, and new visitors may outnumber returning ones on Safari, without any more people visiting. Sessions and enquiries are steadier measures to rely on.

What happens to ad conversion tracking?

Conversions that happen more than a week after an ad click, from Safari visitors, may no longer be linked to the click, so the ads can look weaker than they are.

Which businesses are affected most?

Those whose customers take more than a week to decide, such as renovations, weddings or larger purchases, and those with many iPhone visitors. Same-day decisions are barely affected.

What should I do before the update?

Check your Safari share, note how long customers take to decide, segment reports by browser, annotate the date the update arrives, and count enquiries and calls directly.

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.

Customers who take weeks to decide?

Send us access to your Analytics and ads. We will set up reports that still make sense after the change.

Start a Conversation