The shift has already happened:
The mobile advertising industry is shifting and changing faster than anything that came before it. Beyond the giant leap in ad technology, possibly the most significant change is related to the speedy education process that most major advertisers underwent during the course of the recent year.
The majority of big advertisers became experienced and knowledgeable with their user acquisition campaigns. Today we can hardly see any big advertiser who performs non controlled campaigns that only measure volume and not quality, wrongly paying for incent and fraud users, or even paying overpriced rates for eyeballs and clicks, without measuring it’s return on investment.
The industry is maturing fast. It’s evolving towards sophisticated media buys, demanding much more from its suppliers – the ad-networks, such as reaching specific KPIs (imposing specific activity parameters within the app) and it’s also getting a lot more competitive.
The most significant change has not yet fully matured, but it’s rapidly becoming more common and it’s inevitable: The shift from CPI to CPA.
Here is why it has not happened yet and why it is inevitable:
The Marketing Spectrum:
CPI and CPA are just two adjacent points on the Marketing Spectrum that stretches between branding on the one hand and pure Performance marketing on the other hand.
On the branding side, we find marketing activities whose impact is extremely hard to measure, such as flat rate buys that give no guarantee to the advertiser, followed by the more common CPM and CPC, which only guarantees eyeballs or clicks but doesn’t carry any accountability to the number of users they deliver.
At the middle of the spectrum we find the CPI model, which became the standard of our industry (why that happened is touched upon below), while further towards the performance side of the spectrum we have the CPA (Cost Per Action, recently also dubbed as CPE – Event) where at the very end of the line lies the Revenue Share as the most ideal pure performance model.
Risk is the parameter that changes its value through this Marketing Spectrum:
On the branding side, the advertiser carries all of the risk while the publisher does not need to guarantee any performance. At the extreme opposite end, of pure performance, lies Revenue Share where the entire risk burden is placed on the publisher who sells its valuable inventory without getting any guarantees or any transparency to its impact, and Advertiser has 0 risk.
At the middle part of the spectrum, where the industry stands today, with CPI as the most common model to buy users (and even to internally measure effectiveness of media buys), theoretically splits the risk between the Advertiser and the Publisher, so that the Publisher is accountable for the delivery of new users to the app, and the Advertiser is then responsible to monetize them, or not, as they see fit.
So splitting the risk at CPI sound like the logical (and maybe even the justified) way to go, right?
How CPI won over CPA:
Aligning the entire mobile industry to work on CPI was not a result of a thoughtful process, where advertisers and publishers gathered for a meeting around a big round table to debate on the matter, to eventually reach a logical decision of aligning everyone to CPI. Rather, the course of events that brought the entire industry to this model is detailed hereby:
If that meeting was to happen they would have probably chosen CPA, as that was the most common model on the PC-Web. First and foremost, there is one big technical difference between mobile and PC: The mandatory app install event that only exists in mobile.
This mandatory event in the user acquisition funnel, meant that now, finally, there is a unified measurement point that exists across all devices, operating systems, countries and verticals- the Install event. This event does not exist as is in PC, where each vertical has its own user acquisition funnel, so for some services on PC the user first has to register for a free trial, while for another service the user has to first download a software or just leave their phone number or their credit card details for verification, or becomes a customer only once completing a purchase.
This variety of methods also exists today in mobile, with the only one difference- it all happens after the Install event. This simple technical fact, eased the process for this event to become the only one clear point for measurement and also the obvious point for easy comparison between different companies.
So now, our CMO from company X can easily compare his costs with the CMO from company Y. Even if he is buying on CPC, our CMO immediately understands that the effective CPI he is paying is 30% higher than the CPI his competitor is paying and he rightfully demands the same costs, leading him to ask for a fixed CPI as his favorite business model.
The other reason that solidified the CPI status, was the lack of sophistication on the advertiser’s part, so a couple of years ago you could easily find overpriced, worldwide CPI campaigns, with no KPIs other than the number of installs gained by the advertiser. Coupled with the convenience it gave to the ad networks, this model became the basis for a clattered and rapidly growing industry.
The Shift to CPA
The forces against the shift:
The shift is already happening everywhere, yet there are several forces trying to prevent it from happening.
These would be mainly the ad networks who obviously don’t want to take more of the risk upon themselves, to be held accountable for the overall performance of their traffic. They can’t be blamed. They want to keep selling impressions, clicks and installs without the burden of tying them to the actual monetary benefits of the advertiser.
But the fault is not only theirs. They are not incentivized the right way. Instead of tightly tying performance to reward and paying the networks accordingly, advertisers still base their decisions on a variety of parameters that matter far less than performance, such as premium placements, apps chart rankings, reach, targeting specific audiences, transparency, native ad formats, non-skippable ads and videos, etc. When these parameters are not tied into performance they don’t aid in reaching ROI, rather they distract the advertiser from being able to focus and measure what matters most.
The gradual shift started happening during the last 12 months. Today, the biggest mobile campaigns carry with them very clear KPIs. Most of our advertisers specify their minimum KPIs as part of the campaign terms. For example, advertisers demand a specific retention rate, or a clear conversion rate (CR) from install to any specific in-app event, such as completion of a level, registration, or first in-app purchase.
These are scattered steps in the process of aligning the industry to CPA. Advertisers and Publishers might not fully realize it, but these KPIs actually mean they are already working based on CPA. Here is why:
If the advertiser pays $X CPI and requires Y% CR to the Action in the app, then their CPA = CPI/CR.
For example, if Advertiser pays $1 CPI and demands 50% CR to registration, then they have a CPA of 1/0.5 = $2 CPA. (They require 2 installs to make a registration, so they are willing to pay $2 for Registration).
These campaigns already become more common, yet they are listed as CPI campaigns, while in fact they are CPA campaigns. The next step is for the Advertiser to say, forget about KPIs, I want to pay $X CPA for a user that performs his first in-app purchase (or any other main event he cares for), regardless of how many other users you brought me.
This shift has already started and it is not before long that most advertisers will switch to CPA. The ad-networks will simply have to align themselves accordingly, despite the fact that currently, they don’t have this capability, as the Advertiser is writing the cheques and therefore the Advertiser is King. Ad-networks who will not be able to align will go out of business.
The benefits of CPA:
Running user acquisition campaigns based on CPA carries several significant advantages:
The barriers for CPA and the solution:
The first barrier we already touched upon is the lack of education on both the Advertiser side and the supplier side. This is changing fast, as the driving forces of our industry realize that paying on CPI, while demanding strict KPIs, actually, equals paying on CPA. When Advertisers will start demanding to work based on CPA, this already existing trend will become the standard.
There is one big technological barrier which still remains: Most ad-networks are not equipped to work based on CPA. They don’t have the knowledge, nor the technology to do so. This poses a real problem for them, they have to account for the traffic they buy on CPM, CPC and CPI, calculate the conversion rate to the Action that the Advertiser is paying for, and figure out if it is making a profit or a loss. Unlike CPI to CPI where the network doesn’t take any risk, in this case they can’t know in advance which traffic source will end up making them money or losing them money, so they must change many processes and technologies internally to be able to handle this shift.
Through great foresight, and as a real performance-based network, Performance Revenues is the first true mobile CPA network. Finally, it is possible to guarantee CPA campaign to Advertisers from all verticals.
This has become possible through an in-house technology dubbed KPI Hero which automates the process of converting CPI to CPA. The technology constantly monitors the performance of each campaign, calculates the effective CPA for each source, and performs the required actions to increase or pause publishers based on their performance.
Once we started to deliver through CPA campaigns, our advertisers found it to be so much better for them, as it was more convenient, much less demanding and ultimately it aligns our own interests tightly with theirs, leading to a trustful and mutually beneficial partnership.
We believe that this is a real game changer and a leap forward into the not-so-far future, where most campaigns will run on CPA.
Don’t fear change, embrace it, and stay
ahead of the curve