Explaining “Stolen” Conversions

Most readers who read my blog come for obvious reasons. They want to read what they believe is going to have the most impact on their business, and for their time. So I try to write about the topics that matter most to advertisers as long as I can stay original and bring something new that has added value. I enjoy more writing about things that haven’t been widely discussed (at least not mainstream). In this post, I am dedicating it to explaining the concept of “stolen” conversions. To wrap your head around this concept, let me start with a question. If an advertising platform is reporting that they generated you a sale or a conversion, how do you know that they drove or generated that conversion for you?

Ever since pixel tracking was introduced to online advertising, most of the control over reporting and analyzing moved to the ads’ platforms. Strangely enough, we’re buying advertisements from all these giant advertising platforms like Google and Facebook, and we’re relying on them to report to us the actual outcome of the ads. In return, we then use that info to make extremely meaningful decisions about the future of our ads. If you’re reading this and don’t understand the big deal here, then consider this analogy… It’s a little longer than my usual analogies so bear with me:

Imagine you own a business, and you hire a freelancer or sub-contractor to do some promotional work for your business such as standing out in front of your store and trying to bring new people in. Then, he reports to you on how many people that he promoted ended up purchasing. You’re not exposed to when he advertised to these people and whether or not they were on the way to the store anyway, you just get a report that says how many people he spoke to and ended up purchasing. Then, you decide to hire another subcontractor to promote your store, the same rules apply, and the only difference is the results of his reports. When you do the math, you notice that some of these purchases appear on both reports and you don’t know who to credit those specific purchases to. It’s impossible to identify whether or not they were responsible for bringing that person into the store, or maybe they just briefly spoke to them on the way in? Remember how back in the days, when you would purchase something in a large retail store and they would ask you who assisted you with your purchase, and that salesperson would get credit (an attribution) for the sale? Does that mean that you purchased it because of that specific salesperson? And would you think it would be a good idea to rely on the salesperson’s reports? 

I know I am digging deep here and some critics might say that these advertising platforms are very transparent with the data that they collect and you can view and identify each one of these confirmations to validate that it’s real. I’m not arguing here that these conversions happened, but I am arguing and urging advertisers to verify whether or not these conversions should be credited to the platform in the first place. What’s the next thing that’s likely to happen after you receive a positive report through the advertising platform that shows that you got a great return on your ad spends? You guessed it right, you’re going to spend more money on the platform to generate more of that great performance. Advertising platforms understand the strong correlation between positive ROAS reporting and the advertiser’s spend, that’s why they do everything they possibly can to attribute as many conversions as possible even if it leads to less efficient advertising. Remember, their goal is different from yours. Your goal is to maximize profit or growth in sales, while their goal is for you to spend more money on their platform. 

So what’s a “stolen” conversion?

A “stolen” conversion is a conversion that happened through one traffic source and was claimed (attributed) by another. Think of, for example, your repeat customers who on average come back and purchase several times regardless of any paid advertising. “Interacting” with them through a paid ad will result in that paid advertising platform taking credit for the conversion and essentially say, hey- we generated that sale for you. Or think about the customers who are already in the process of buying something from you, and just before they purchased they viewed your ad on their newsfeed (perhaps without noticing) or clicked on an ad to get back to your website and purchase. If the advertising platform takes credit and attributes that conversion, then should it be categorized under return on investment? Should they be eligible to claim that return? This could become especially lethal when you are advertising on more than one advertising platform. Advertisers who are big corporations and have large budgets and don’t face the same obstacles as small to mid-sized advertisers. If they, for example, agree to run a pilot of ads on a new advertising platform which they haven’t advertised at a large scale previously, then they will ask for a statistically significant study showing the added lift or added contribution of this new advertising platform and they will decide based on the study, what’s the true value & contribution, and whether or not they think it’s effective for them. These tests require massive amounts of data and spend and are not available to most advertisers. Instead, small size advertisers are left having to figure it out on their own without having access to all these tools. Think about it from the advertising platform’s perspective, why would you want to offer that kind of tool to the public? If they offer access to a tool that can accurately calculate the true contribution of those ads (also known as a lift), then it’s almost always likely to be less than what they initially reported to you in their ads manager. The less positive feedback they give advertisers (decreased in reported return on ad spend) the fewer advertisers are going to spend on ads.

Here are some components that I use to help me identify over-or inaccurate reporting that might include conversions or ROAS in question:

Too Good to be true – 

This is probably the easiest of them all to identify. If you have been running ads for more than a second, then you know that in almost all cases generating good or great results is complex. Competitive ad spaces apply to all industries, and when we normally come across exceptional reported results by an advertising platform, we want to dip deep to identify the reasons. This commonly happens when you have one main traffic source that you use for advertising (Facebook Ads, Google Ads, etc..) and then you introduce a secondary traffic source on a lower budget. Then in many cases, the secondary platform reports exceptional results in very little time. One might think that he is on to something and has a breakthrough in his marketing efforts until he realizes that these conversions are mostly counted on both of the ad platforms. I like to use the analogy of hyenas and the lions. The lions tend to do the hard lifting when it comes to hunting down their prey and their conversion rate (the hunting success rate is pretty low). Then, in many cases comes the pack of hyenas and either eat the left observers (low hanging fruits) or in some cases can outnumber the lions and get the remaining of the prey. But you wouldn’t want to be able to get better results if you wanted the hyenas to do all the heavy lifting right? 

Abnormal conversion rates – 

Another way to identify these conversions in question is to break down your ads/campaigns and look at their conversion rate. The conversion rate is a simple calculation of the total number of people who clicked on your ads (link clicks) divided by the total number of conversions you generated (purchases/sign up or whatever you are optimizing your ads for). Then, you analyze the total conversion rate coming from all ads and break it down to see how they perform individually. You can analyze the global conversion rate on your website that includes an average of all traffic sources combined and see if these numbers are close to each other. If your ads are reporting an abnormally high conversion rate, then you can assume that much of that traffic is existing traffic that will be reported by another traffic source as well. If for example your conversion rate on your website is 2%, and some of your ads are converting at 25%, then you can confidently assume that these conversions are over-reported by more than one traffic source. If specific ads are generating an extremely high conversion rate, then it might be because they are only targeting and delivering warm traffic that’s either in the process of purchasing, or currently has a high intent to purchase. 

Switch and Swap- 

Similar to the first point of results that are too good to be true, switch and swap refer to achieving results on a new advertising platform really quickly. If you have existing sales and conversions from one traffic source and you are testing out a new advertising platform for a new traffic source, consider rapid positive results might be due to a fact of overlap of reported conversions. When ad platforms tend to show ads to users for the first time, there is a learning curve that takes some time until the system knows how to accurately predict who to deliver the ads to. 

An increase in ad spend leads to ROAS drop.

Another commonly seen symptom that signals that you are swimming in the ocean of very warm traffic is the inability to increase your daily or monthly spend without taking a hit to your return on ad spend. I can’t tell you how many times I hear the story that an advertiser is getting an exceptional return on investment, but the only problem is that when they increase the daily spend, the return on investment tanks. This happens for the simple reason that your warm audience (especially the users with the high intent) are normally smaller audience groups. That’s why when you spend relatively small amounts per day you reach most of them, but when you try to spend more money, the algorithm will look for new opportunities, and then you will be swimming in cold traffic. Results will be nothing like the results with warm audiences, and together with warm traffic their average return on investment might look ok, but when separating the reporting, you might see a pretty big range in the results that they deliver.

IN SUMMARY

Stolen conversions, or conversions that were incorrectly claimed (attributed) by an advertising platform, can be detrimental to your advertising campaign if they go unrecognized. Platforms like Facebook attempt to claim as many conversions as they can in their ROAS reporting to trick the advertiser into thinking that the ad is more successful than it really is. This is especially troublesome for advertisers who run ads on multiple platforms. 

High-value companies have the resources to run their own statistically significant studies on the success of their ads. Unfortunately, mid to small-sized businesses can’t afford to run a statistically significant study. Instead, mid to small-sized businesses are encouraged to look for the telltale signs that conversions are being stolen: If your short-term success appears too good to be true and you recently introduced a new advertising platform, then it probably is. Abnormal conversion rates may imply that multiple advertising platforms are taking credit for the same conversion. If you find your return on investment tanking while increasing your ad spend on a seemingly successful campaign, then your warm audience target may be tapped out and the algorithm may be swimming in cold traffic.