Being on the sell side of the advertising business we hear about CPA or Cost Per Action a bit too often—a shiny promise of analytics companies and the wet dream of product marketers, CPA implies a fixed reward for the desired outcome. Only pay for a result—how this can sound bad to anyone? And yet, no ad providers accept it.
“Heads I Win, Tails You Lose” pretty much explains the biggest weakness of this model—it shifts all the risk of the business venture's success from a business owner to an ad publisher, who has no control over it.
It’s like asking a hacker to rob a bank and share money 50-50 with you so this way you both make a profit.
For those of you who are still in doubt, we’ll go through three main problems with the CPA advertising model:
The distribution of risk and reward in advertising
If your CPA rate is profitable, you should be the one capturing the margin
No one knows how to sell your product better than you
The distribution of risk and reward in advertising
Before a user lands on your landing page and fills out that signup form their attention typically passes a number of steps, on each of which there is a responsible party working hard to make sure the user gets to the next stage. Unfortunately for you, an advertiser, the user’s attention does not come for free and does not materialize from the air.
First and foremost we have publishers, they are builders working hard to create a valuable product that users will want to come back to over and over. Their job is to generate user attention (ad impressions). That’s the value they produce and get paid for.
But publishers usually don’t have the capabilities to acquire advertisers and sell ads at scale, nor can they provide them with technical tools to utilize impressions most efficiently—that’s where the ad networks come in.
Good networks like Slise take raw traffic and turn it into diamonds, via cleaning, data enrichment, targeting, analytics, and convenient user experience, allowing advertisers to launch campaigns across millions of users with just a few clicks. Connecting many-to-many, the ad networks provide the most optimal way for advertisers to reach user attention and take the risks off publishers.
That’s what they get paid their margin for, and it shouldn’t be surprising that it’s proportional to inventory processed and not the performance of the ads, as neither ad networks nor publishers have control over what’s being advertised.
But it’s not just enough to obtain user attention, one needs to know how to utilize it, that’s where experienced marketers and media buyers come into play.
Whether it’s an in-house team or a marketing agency, they use their expertise to communicate an advertised product to the users in the best way and produce the maximum result from the limited amount of impressions, which can be expressed in CTR—how many clicks are generated per impression.
The difference between good and bad ads shown to the same audience can easily be 10-fold. Marketers get paid for the impact on ad performance they produce thanks to their expertise. They can make bad CTR better, but even they can’t control the final ROAS, as it depends on the actual product and how it captures value.
After the click happens, the user lands on the product page, where a product and marketing team continue to work with it. The role of advertising ends here, as no ad platform has control over what happens next.
Landing pages can vary in their efficiency, and many services have been built around AB-testing and conversion optimization. It also depends on whether a product fulfils the promise made in advertising. If the funnel is not optimized, or a user doesn’t understand what to do next, valuable clicks can easily be lost, and that’s the responsibility of product and marketing to make sure it doesn’t happen.
Finally, when a user signs up, it depends on the product and the value it brings, as well as the onboarding flow and other minor product tricks, to define whether the user will want to pay for it.
At the end of the day, to sell something valuable you need to produce something valuable, and it doesn’t matter how effective the whole marketing funnel is if, at the end of the funnel, the user gets disappointed by the product and doesn’t get their needs met.
The process of generating revenue from raw ad impressions involves many parties each of whom adds their contribution to common success by de-risking their part of the funnel and getting rewarded for their job.
Each link in this chain is critical because as you can see in the final formula, the final revenue is a product of the conversions on each step, and if one of them malfunctions the outcome of all will go to zero, and, of course, none of them want to bet their compensation on work of others.
In the CPA model, however, an advertiser wants to cut contributors on steps 2-5, and make a publisher solely responsible for user acquisition.
The advertiser in this case just hands over a referral link and waits for sales to come, while the publisher needs to ensure they deliver ads to the right audience, communicate the product in the best way, AB-tests different versions of creatives, make sure that the landing converts, or even build their own landing, and hope that the anyone actually will pay for the advertised product or they will waste all their traffic for nothing.
I can’t emphasize enough how nonsensical it is.
In fairness, that’s exactly how it works in traditional affiliate marketing: the media buyers only get a product order link and they create their own landing pages and onboarding flows to ensure that the traffic converts at the highest rates, with one small but important difference: They choose the product they want to distribute based on the historical data of the performance of different offers in affiliate network. Not even traditional affiliate marketers would bet their traffic on the random product handed to them.
✨ CPA narrative pictures the shiny dream, that advertisers can build anything they want, and even if they themselves can’t sell it, there is a whole bunch of humble publishers with freely available traffic who will be happy to sell it for them, and get 10% of the profit, with no downside ✨
No, Karen, no one needs your generous reward of $1/signup to your Uber for Dogs app, which takes $100 and 20k real impressions to generate because your app is garbage and no one wants to use it. Get real.
In fact, there is always a demand for user attention, it’s one of the if not “the” most valuable assets in a digital world, and people who know how to turn it into revenue are ready to pay for it a lot (especially at overheated and overfunded markets like crypto).
Which brings us to the second point:
If your CPA rate is profitable, you should be the one capturing the margin
Have you noticed what products are usually distributed through the affiliate networks? Gamblings sites, pp creams, microloans, weight loss pills… Not Gucci bags or Enterprise software. Most of those products have problems with growth and have to rely on grey methods of affiliate marketers (like facade Android apps) to mislead the user into purchase.
If the product has high conversion rates and is highly demanded by a user, say, it grows organically or generates $100 in profit for $10 in advertising spend, thanks to your genius marketing—there is no need for affiliate partners, it doesn’t make sense for you to give away this margin to anyone.
Take an example we saw in our practice:
An advertiser offers to pay $5 per specific action
If a publisher or an ad network runs a test and sees that it costs $10 to generate an action, they will abandon this offer right there
If, oppositely, a publisher can produce an action for $4, the campaign will run generating profit for the publisher until the channel saturates
Our question here is—if it’s a given that the CPA reward should be higher than the cost of traffic for a campaign to work, why don’t you, an advertiser, run it yourself by buying impressions at a lower cost and pocketing the extra profit?
At the end of the day:
No one knows how to sell your product better than you
It’s naive to expect that publishers or ad networks that serve tens or hundreds of advertisers simultaneously will study your audience and write a marketing strategy for your product. You and your marketing team are the best people to advertise your product and no one in the world knows your audience better than you do.
“Oh, but we can produce pretty creatives with our messaging and give it to the advertising partners!”
No, Karen, performance comes with data and a continuous feedback loop, and not just ideas from your head. You may think that illustrating a dog poo on your ad for Uber for Dogs is catchy because you read it in a blog, but if the ad doesn’t perform, the publishers will lose their money, not you. On the contrary, when you bet with your money, you will be much more thoughtful about which ads work and which don’t and will optimize them proactively.
This idea of putting the responsibility on someone else without giving them any control, and hoping they will be okay with it, is a second fundamental problem which explains why the idea of CPA as it’s commonly imagined is dead on arrival.
Is it possible to make CPA work in crypto?
We may sound sceptical, but as with any business model, it can well live and thrive if it’s commercially viable. What makes a CPA offer commercially viable? A product that is easy to sell based on a proven track record! For that:
A distributor should be able to choose the offer they want to advertise based on the profit potential.
To estimate this potential the affiliate/referral network should provide the historical performance of the offers across different steps of the funnel: it should be clear without tests if the product can generate profit for a publisher if they plug it into their traffic source, and what parameters can be improved further.
A distributor should play the role of a marketing agency, not a publisher, as their job narrows down to buying traffic for cheap and optimizing the hell out of it to generate conversions.
Data on conversion events should be synced with a distributor’s analytics system to be able to optimize the strategy based on the results.
Without that, lifting a rock from the ground, saying I’ll pay you $5 if you sell it for $50, and expecting that there will be a crowd fighting for this opportunity is just absurd. And that’s exactly how CPA deals look in crypto today.
Peace ✌️