Case Study
Consumer App Attribution Engineering Multi-Channel Offline Events

iOS privacy updates cut Thursday's attribution to zero. We rebuilt the data pipeline from the ad click to the venue door.

When you cannot see which ads drive real ticket buyers, the algorithm optimizes for the wrong thing. We closed both gaps - app installs and in-app purchases - with a custom server-side solution.

TL;DR
  • iOS privacy changes made Meta, Google, and TikTok completely blind to Thursday's two-stage funnel: app download then ticket purchase
  • Built a first-party server-side solution that connected a single ad click to both conversion milestones
  • Ran multi-channel creative across TikTok, Meta, and Google, each mapped to a specific role in the funnel
  • Recovered 28% of previously lost attribution data, dropping app acquisition costs by 22%
  • Ticket sales up 32% as the platforms finally learned to find real buyers, not passive downloaders
28%
Lost attribution data recovered
-22%
App user acquisition cost
+32%
Physical event ticket sales
3
Channels: Meta, Google, TikTok

Your ad platform is optimizing for the wrong customer.

When Meta, Google, or TikTok cannot see what happens after someone clicks your ad, they optimize for whatever they can measure - usually the cheapest, easiest action available. For Thursday, that meant platforms were spending budget finding people who might click an ad rather than people who would download the app and buy a ticket to a live event. Post-iOS privacy updates wiped out conversion visibility across both steps of their funnel simultaneously, and the platforms filled that gap by chasing volume on the wrong metric entirely.

Thursday's funnel has two milestones that both matter to the business: a user downloads the app, then buys a ticket to attend a live weekly event. Standard tracking breaks at the App Store handoff, so the platforms had no visibility into either step. And recovering one without the other would not have moved the needle, because cheaper installs that never convert to ticket buyers do not drive revenue. The fix required closing both gaps at once and feeding actual purchase values back to the platforms, so the algorithms could learn what a real Thursday customer looked like.

Ad Examples
Static and video creatives coming soon

A server-side attribution system across three channels.

Each channel had a specific role in the funnel. TikTok drove cold audience awareness through high-energy event content built around the experience of attending. Meta reached active urban event-goers with app-install creative targeted against nightlife and social interest stacks. Google intercepted high-intent local searches from people already looking for things to do and routed them to optimized landing pages. All three fed into a custom first-party server solution that tracked the complete journey from ad click through app install through ticket purchase, feeding real conversion values back to each platform - not just signals that a conversion happened, but the actual value of what was purchased. That shift is what moved the algorithms from chasing download volume to finding people who show up and spend money.

With clean conversion data flowing back from both milestones, we could also see exactly where different user segments were dropping off and what messaging they were responding to. That visibility informed a dedicated creative layer built around two distinct messages - what downloading the app gets you versus what attending the event gets you and how the two connect into a single experience. Static and video creatives were developed for each stage, so a user encountering Thursday for the first time saw different messaging than someone who had already installed the app but had not yet bought a ticket.

Ad Examples
Static and video creatives coming soon

Once the platforms could see the right customer, they found more of them.

With both conversion milestones reporting real data back, acquisition costs dropped 22% and ticket sales rose 32%. The 28% attribution recovery was the data input that gave the algorithms something real to learn from, and the downstream revenue impact followed directly. The creative split between app value and event value compounded those results further, because users arriving at each stage already understood what the next step was worth to them, which reduced drop-off between install and ticket purchase.

Ad Examples
Static and video creatives coming soon

Common questions about this engagement.

How do I know if my current tracking setup has this problem?
If the conversion that actually matters to your business happens somewhere other than the webpage your ad points to - inside an app, over the phone, in a physical location - there is almost certainly a gap between what your platforms think is working and what is actually driving revenue.
Is this only relevant for app businesses?
No. The attribution gap is most visible in app funnels because the App Store handoff is a hard break in the data trail, but the same problem appears anywhere the customer journey crosses a boundary the platform pixel cannot follow - phone-based sales, in-store purchases, or multi-step online funnels all face versions of the same issue.
What is the difference between fixing the signal and just getting more data?
Sending a binary conversion signal tells the platform something happened. Sending the actual value of what happened tells it which customers are worth finding more of. The first improves attribution. The second changes who the algorithm targets, which is where the revenue impact comes from.
How quickly does the algorithm respond once better data starts flowing?
Meaningful optimization shifts typically show within two to four weeks of clean conversion data flowing consistently, as platforms need a minimum volume of conversion events before their learning phase completes and targeting begins to shift.
Think your ads are flying blind?
Most businesses running post-iOS campaigns without server-side infrastructure are spending more than they need to and attributing results to the wrong sources. A short audit is usually enough to identify where the gaps are and what they are costing.