Today, I’m going to be answering the question that was emailed in.
This is an issue that everyone has to deal with, especially right now with iOS 14 changes.
Everyone has limited data, right now, so if you don’t have the same specific situation that the person asking this question has don’t worry, the answer is going to be helpful to you anyway, so the question comes from Jim and he writes in saying,
“I took on a client with a $10k/day Google Ads budget – split up between YouTube, Search, and GDN. This is a unique use case for me. They are driving traffic to their mobile app’s Steam store page where users can DL my client’s app for free. Since they do not own or operate that page, we are unable to implement traditional tracking pixels. Steam just introduced UTM in beta and I’ve created all the links with the proper parameters in place to track “activations”. The problem is the most recent data UTM analytics can show me is yesterday. Even then, the activation number isn’t totally reliable because a lot of users don’t sign up right away. They’ll sign up 1-2 days later and the Steam analytics will give that conversion credit to the day the user clicked on the ad.
I’m not used to optimizing without at least “near” real-time data. I won’t have reliable data on a day’s performance until 2 days later. How would you go about optimizing under these conditions?”
Okay, so great question and a question that really everybody should be thinking about because nobody has real-time data right now.
Nobody has reliable data, and even before these iOS 14 changes, the data was not reliable. We were still missing at least 10% of conversions on a good day, usually more.
Now, on mobile, we’re missing about 40% of conversions, and on tablets, we were missing almost 70% of conversions. So no, we don’t have reliable data. How do we deal with that? How do we deal with not having reliable data?
We still have it pretty good. There’s a book everybody should read called Scientific Advertising that was written by Claude Hopkins. This book is all about advertising, specifically based on data and using data to drive your decisions in advertising in a scientific way.
Claude Hopkins was talking about scientific advertising, he wasn’t worried about real-time data or iOS 14 changes.
He was able to figure out ways to track advertising without all of the bells and whistles that we have, that we’ve gotten used to, and are comfortable with. Now, I’m not saying that we should ignore the capabilities that we have to track things sometimes in real-time. This is great and it’s obviously a huge advantage to us advertisers. But it’s not the be-all-end-all of advertising and there are still a lot of things that you can do if you try to think about really optimizing a campaign.
It’s not just about optimizing for a conversion that shows up in the Google Ads account. It’s about optimizing for the user behavior for the sales, so let me throw a couple of things at you to help you think about this. Now, I am not exactly sure what some of the capabilities are and what types of people are downloading games for steam, but without real-time data, you optimize based on what you know.
So, if you’re looking at making a change, you might need to look at the data leading up to that change. Make the change and look at the data after the change. If all you know is your ad spend, your number of clicks, and the number of downloads or activations, you can still optimize with that information. It means you’re not going to be making a bunch of little tweaks and changes.
Instead, you’re going to be thinking about big changes that would potentially have a big impact on the results. Those are changes that you can see on a bigger scale and decide, “Okay, did this change have the effect I wanted – yes or no. I’ll keep it and then I’ll try something else” or “No, things got worse. Let me change it back to how it was and then I’ll think about testing another big change.”
If that’s all you can do, I would also think about this in terms of possibly setting up a bridge page. So rather than sending traffic directly to the store for the mobile app, it’s reasonable to set up your own landing page for this, and then that landing page links to the steam store.
Of course, if that causes downloads to be completely obliterated that would not be a good move.
It might be something you have to test, but there are some definite advantages to this if you can do that. So, instead of people clicking right to this steam page and downloading this app, they click to your bridge page. From there, they would click to the steam page. You can track that click as a conversion. So, we can assume that if they make it through the bridge page and go to the steam page, they’re highly likely to download the app at that point.
That’s a high-quality click because they clicked on you. Count that as a conversion and now you have a lot more data in the cloud account to optimize. The other thing that this bridge page could do for you is really allow you to test the messaging.
Let’s say you have a couple of different ideas for the main screenshot to show, the main description of the app, the reviews, or whatever is being used to sell this app. You can test those on a couple of versions of the bridge page. From there, once you find out what wins and what is getting the most traction from the bridge page to the download page, you then update the steam download page so you make these tests on different versions of your own page.
As you find winners, you can then update the app download page because now you just have to change it and see what happened before you revert it back or you keep it. If you’re testing multiple pages, you can find things that work and update this steam page as you find those things.
The other thing you can do is actually make different pages for different traffic sources. Maybe you have different mobile pages versus tablet pages. You can get a much better idea that way, even if you’re not able to see all the conversion data in Google Ads. Again, you aren’t able to see all that data in Google Ads, but now you can at least see it on the bridge page and you can calculate cost per download and ROI in a much more scientific way than if you’re simply sending all the traffic to one page, especially if you’re sending all the traffic to one page that you aren’t able to track very well.
Everyone is struggling with getting reliable data right now, but we’re going to be fine. Claude Hopkins was able to figure this stuff out in the early 1900s. He was able to figure out how to scientifically test advertisements and make them work. We have a lot more at our disposal than he did, it’s just a matter of thinking about it differently.