Apple's tracking prompt and your lead ads

Apple’s developer conference was yesterday, and the part that matters to anyone buying ads has nothing to do with new phones. In iOS 14, arriving this fall, an app has to ask permission before it tracks you across other companies’ apps and websites. A plain prompt, in the app, with a decline button that is just as easy to press as the allow button.

The advertising identifier that Facebook, and everyone else, has quietly used to connect an ad tap to what happened afterwards becomes opt-in. Apple is also adding privacy labels in the App Store and an option to share approximate location instead of exact location.

Facebook has not said much publicly yet. It will. In the meantime we have had the same conversation with several clients this week, so here is our read, written for a local business that puts a real budget into Facebook and Instagram every month and counts leads, not impressions.

What this actually touches

The honest version is that local advertisers are less exposed than app companies and ecommerce brands, but nobody is untouched. Sorted by how much it should worry you:

• Lead forms themselves keep working. If you run lead ads with the form that opens inside Facebook or Instagram, the name and phone number are captured on the platform. That does not depend on any tracking identifier. The lead still arrives.
• Website audiences shrink. Retargeting pools built from pixel traffic, “people who visited the booking page and did not book”, will get smaller and staler on iPhone traffic.
• Lookalikes get thinner. A lookalike built from a pixel event is only as good as the seed list. Fewer tracked events, weaker seed.
• Optimization gets noisier before reporting does. This is the one owners underestimate. The delivery system learns from conversion signals. Fewer signals means slower learning and more wasted spend at the start of a campaign, and that shows up as cost per lead drifting up for no visible reason.
• Attribution reporting gets less generous. The 28 day click and 1 day view numbers in Ads Manager are going to report less than they do today. Some of what disappears was never as real as the dashboard implied.

The last point is worth sitting with. The lead count in Ads Manager has always been Facebook grading its own homework. When it drops this fall, the leads will not all have vanished. Some of them will just have stopped being counted.

Why this is a smaller problem for local than for retail

Most of the local Facebook spend we manage is not clever. It is a tight radius around the practice or the shop, a broad audience inside it, a decent offer, and a lead form. Geography is not tracking. Age is not tracking. The offer is the same offer. None of that breaks in the fall.

What breaks is the part of the account that leans on following people around the internet. If your entire Facebook strategy is a retargeting ladder off website traffic, you have work to do. If it is “reach people within a few miles who look like our patients and give them a reason to call”, you are mostly fine.

The other reason local is insulated is that the real conversion is not on a screen. It is a phone call and a front desk. That is also where local businesses lose the most money, and it is the part we would fix first regardless of what Apple ships.

What we are changing between now and the fall

We have three months. That is enough time to stop depending on a number that is about to change.

1. Build a baseline now. Record cost per lead, cost per booked appointment and show rate for the next three months, before the prompt exists. In the fall you will need to know what normal looked like, or every change will get blamed on Apple.
2. Start counting leads outside Ads Manager. Every lead into a CRM with its source stamped on it, every call tracked, every booking marked as kept or not. The platform will always tell you a flattering number. You need one that does not care whose ad it was.
3. Get first party lists in order. Customer list audiences, built from your own patient or customer records, are uploaded by you and do not depend on a device identifier. Most of the businesses we work with have years of past customers sitting in a practice management system and have never once used them as an audience or a lookalike seed. That list gets more valuable this fall, not less.
4. Shift more capture on platform. Lead forms inside Facebook and Instagram survive this better than a click to a landing page and a pixel. We are moving more budget towards them and making up the quality difference with faster follow-up rather than more qualifying fields.
5. Fix speed to lead. A lead form that fills in six seconds is worthless if nobody calls the person for a day. If the ad account is going to get less efficient, the follow-up has to get better to hold the same cost per booked appointment.
6. Stop buying view-through stories. If a campaign only looks profitable on a 1 day view window, it is probably not profitable. Judge on clicks, leads and booked appointments now, so the fall reporting change does not move your decisions.

This is also the direction of travel, not a one-off. Safari already blocks third party cookies. Chrome said in January it intends to phase them out within two years. Apple is doing the same thing to apps that browsers have been doing to the web. Any plan that survives only while the tracking holds is going to need rebuilding every year.

What to do this month

• Write down your current cost per lead and cost per booked appointment, by campaign, and save it somewhere you will still find in October.
• Export your customer list and upload it as an audience. Use it as a lookalike seed and as an exclusion.
• Check whether your lead ads go straight into a CRM with the source recorded, or into an email somebody checks twice a day.
• Set a follow-up standard for new leads, in minutes rather than hours, and measure against it.
• Look at how much of your account depends on website retargeting. If it is most of it, start testing a broader local audience now while you can still compare.
• Do not rebuild anything in a hurry. Nothing changes until the fall.

We will write more when Facebook says what it is doing, which will probably be soon and probably be loud. Until then, the work is unglamorous: know your real numbers, own your own data, answer the phone faster. That is most of what we do in Facebook and Instagram ads anyway, and it is why lead capture and follow-up live together in the CRM we build for clients. For dental practices reopening with a backlog of postponed appointments, the follow-up half is the one worth fixing first.