Generating This Much Revenue From This Ad Placement?

Ad Placement Strategy: They May Look the Same, But the Revenue Isn't

1

Even if an ad placement looks completely ordinary, the actual revenue can vary dramatically depending on how it's implemented in the product and how the ad mediation is set up on the backend. Today, I'd like to share a case where we achieved a 10x increase in revenue simply by changing how the ad placement was implemented in the product, without any changes to our mediation setup.

01. Changing the Ad Format

In the weather section of Alarmy, we have two Native ads. Native ads have the advantage of preserving the user experience to some extent because we can freely select and arrange the components of each ad (icon, image, title, description, etc.) exactly how we want them. However, unlike standard banners, the downside is a disappointing fill rate (since there are relatively few ad networks—ad demand—capable of serving these separated components).

On the other hand, the MREC (Medium Rectangle, a 300*250 banner ad) format has a great fill rate and, above all, a high ad rate (eCPM). This is because it grabs the user's attention with its massive (well, slightly less than massive) size, which easily leads to a good click-through rate (CTR). But naturally, it ends up hurting the user experience just as much.

We wanted to use an MREC on our weather screen without ruining the user experience. Thankfully(?), the weather screen is long enough to scroll, so we were able to place the MREC right around the middle, making it only halfway visible on the initial screen. Positioned like that, it didn't feel overwhelmingly huge at all. Although there's a clear UI difference, it felt like we could capture both the pros of Native (user experience) and the pros of MREC (ad metrics).

How did it turn out?

Revenue from this placement quadrupled. Ad impressions went up 1.3x, and the CPM tripled. We tried this expecting some level of improvement, but we had no idea that the power of a single MREC would be this much stronger compared to two Native ads combined.

This is how we smoothly boosted this placement's revenue by 4x without receiving any negative feedback from users.

2) Increasing Impressions with Push Messages

We decided to get a bit greedier here. We figured that if we could drive more user visits to the weather screen, and thereby increase ad impressions, we could maximize the revenue boost from this placement even further. We decided to try sending a weather-related push message at the right time after users woke up in the morning. We figured nobody would be too upset about getting weather info right after waking up to their alarm, so we went ahead and implemented it immediately.

Of course, even this simple product spec was incredibly difficult to actually implement. Branching logic based on whether users had granted notification and location permissions was tricky across different OSs. And since this was an experience almost every user would encounter, we struggled quite a bit to include some conservative safety nets—like showing a dialog to let users turn off the push feature on their first visit to the weather screen, or automatically disabling it if they ignored the push message three times in a row.

How did it turn out?

Revenue from the placement tripled yet again. This time, ad impressions skyrocketed by 6x (+500%), but the ad rate (eCPM) dropped by half (-50%).

When impressions increase dramatically, you tend to lose leverage in the supply-and-demand bidding war, which naturally drives down the ad rate. The click-through rate is also likely to drop compared to before, which further drags down the actual rate in practice. We can cushion this drop to some extent through additional backend mediation improvements. If we can limit the rate drop to around -30% instead of -50%, we could potentially drive total revenue up to 4.5x instead of 3x.

Also, impressions themselves would increase even more if we could just get more users to grant location permissions upfront. (We didn't send pushes to users who hadn't granted location permissions.) There are still plenty of areas we can refine in our product spec. For instance, we could let users control exactly when they receive the weather push, or we could make the messages themselves more sophisticated.

Through these two rounds of placement improvements, revenue from our weather placement grew by a total of 12x.

Ad placements might all look the same, but the amount of revenue you can generate from a completely ordinary placement can vary wildly.

Let's keep sharing great learnings and make lots of money together :D

If you're interested in ad monetization, feel free to reach out to DelightRoom!