AppsFlyer decries last-click ad attribution as ‘a big lie,’ offers multi-touch tracking

Israel-based AppsFlyer has some news for mobile advertisers.

“The last click is a big lie,” CEO and cofounder Oren Kaniel told me.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":1756714,"post_type":"exclusive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,marketing,mobile,","session":"C"}']

Which is why the mobile analytics company is today announcing its Multi-Touch Attribution.

Let’s say you click to download the Hipmunk mobile app for travel bookings.

You think that’s it until you’re ready to travel somewhere. But, at that moment, a flurry of software processes are running around, probably like the characters in Inside Out.

They’re trying to figure out which ad gets credit for inducing you to download the app, because the chief marketing officer (CMO) for Hipmunk will want to spend more money with that advertiser that touched you last, before you installed the app.

The last-touch or last-click convention has held that the last ad you clicked before the install was the driver of your action.

AppsFlyer’s new Multi-Touch Attribution can now show the other ads that probably contributed to your decision. Kaniel said that, if a given ad is shown to something higher than five percent of the users installing a specific app, it is considered to be part of the chain of ads leading to the install.

---

Check out VB’s research Mobile marketing automation: How the most successful apps drive massive engagement and monetization
$499 on VB Insight, or free with your martech subscription

---

The data and time of the ad interaction, of course, determines its position in the sequence.

[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":1756714,"post_type":"exclusive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,marketing,mobile,","session":"C"}']

He gave this example:

A content-based ad through Outbrain about the Seychelles Islands might be delivered to your device, followed a few days later by an ad from Hipmunk about hotels on the Islands. Then, the following week, the user performs a search on Google for “Seychelles Islands hotels,” with search result ads that include Hipmunk’s.

Then the Hipmunk app is installed by that user.

While there’s no direct evidence that all those ads got the user interested in the beaches of the Seychelles, there is inferential evidence that they acted as drivers in the traditional sales funnel, going from inquiry to product knowledge to purchase.

[aditude-amp id="medium2" targeting='{"env":"staging","page_type":"article","post_id":1756714,"post_type":"exclusive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,marketing,mobile,","session":"C"}']

This attributable sequencing could change how ad budgets behind app installs are allocated. If AppsFlyer-utilizing CMOs buy this argument and use this tool, they could begin to spend their budget on all the ads in the chain, not just the last-touch one.

AppsFlyer said that its beta testing for an unnamed retail client showed nearly a quarter of new app installs were the result of more than one ad delivered to the user. The average was 2.7 ads. AppsFlyer says that increasing the budgets of those ads earlier in the chain increased the overall return on investment.

“If you optimize [your ad budget] on last click, you’re going to damage your ROI,” Kaniel said.

Impression-tracking for attribution

Essentially, he added, it’s like you’re giving a bonus to the guy handing out discount coupons at the mall, because you think he got people to come there. But he’s only their last marketing touch, and they were induced to come there by ads they saw all week.

[aditude-amp id="medium3" targeting='{"env":"staging","page_type":"article","post_id":1756714,"post_type":"exclusive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,marketing,mobile,","session":"C"}']

In addition to enlarging the CMO’s mobile ad spend beyond just the last-touch one, the new tool could also lead marketers to give more credit to branding and content ads, as compared to the current emphasis among app installers for performance-based clickable ads that essentially declare, “download now.” Branding and content, it appears, primes the well to get the user interested in the whole idea of, say, the Seychelles Islands.

“This is huge,” VentureBeat VP of research John Koetsier told me.

“The reality is that the ‘customer journey’ is a crazy zig-zag of data points, and marketers really don’t know which ones are influential and which are not — making it hard to allocate future marketing spend.”

“The challenge,” he said, “will be in how accurate AppsFlyers’ technology is. Bad data here could potentially do more harm than no data at all, if marketers decide to go all in on a medium or tactic that looks successful, but actually is not. But tying it to content is both smart and innovative.”

[aditude-amp id="medium4" targeting='{"env":"staging","page_type":"article","post_id":1756714,"post_type":"exclusive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,marketing,mobile,","session":"C"}']

One thing, though. AppsFlyer can track all ad clicks from about a thousand mobile ad networks, but it can only track ad impressions for about half a dozen networks.

That means that, if the Outbrain content-based ad on the Seychelles is seen but not clicked, and it’s not one of the half dozen ad networks providing impression-tracking to AppsFlyer, the ad isn’t included in the causal chain. It just doesn’t show up.

Which is a consequence of being early to this game. Kaniel said ad networks are “leaving money on the table” by not enabling attribution services to track impressions. Facebook, he noted, has included mobile ad impressions across its network for attribution stats since 2012, although it doesn’t yet do multi-touch attribution.

The reason the vast majority of ad networks don’t enable such impression attribution is because app installs have been considered click-based performance events. The install is a click, of course, and the ad leading you to the install is a click. So why bother with tracking impressions?

[aditude-amp id="medium5" targeting='{"env":"staging","page_type":"article","post_id":1756714,"post_type":"exclusive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,marketing,mobile,","session":"C"}']

Because they matter, AppsFlyer said.