How We're Thinking About Raising the Value of Existing Ad Slots Without Adding More

I want to write about what someone who has made a living in advertising is thinking right now.

As a means of growing ad revenue, the long-standing standard has been "add more slots." Put another banner above the article. Slip an in-feed ad into the body. Drop a rectangle in mid-scroll. Because the revenue projections for these are basically accurate, they have been the quickest way for the person on the ground to hit the monthly target. I myself have made this proposal many times.

However, facing the structural plateau in ad revenue, we are forced to consider paths other than "adding slots." What I am thinking about now is raising the value of existing ads without adding more slots. From the standpoint of someone who built a system to turn articles into audio, I want to lay out how far the logic holds.

Adding slots and raising slot value are different things

First, let me draw a clear line between the two.

Adding slots is the act of taking screen space away from the reader. As a result, the user experience is squeezed, dwell time shrinks, and you enter the loop of falling CPM. I laid this out in the previous article.

Raising the value of existing ad slots, on the other hand, takes nothing from the reader. It is not about adding a new ad to the screen; it is about making the slots that already exist be seen by the reader's eyes for longer and more reliably. These two are completely different acts. The only thing I want to do is the latter.

How does ad viewing move when dwell time grows?

As a premise, the MRC viewability standard judges a display ad as "viewable" when "50% or more of its area is displayed for 1 second or more." In other words, viewability is a threshold judgment, not a continuous quantity that gets better the longer the time. This is a common misunderstanding, so let me state it up front.

That said, when dwell time grows, there are three things that can be accurately claimed.

1. The probability that ad slots outside the first view (below the article, in the sidebar) are seen goes up. Readers staying on the page longer means a higher probability that they scroll down and reach those slots. As a result, the number of viewable impressions increases.
2. The viewed duration (time-in-view) per impression grows. While the reader remains on the page, that slot continues to be displayed.
3. For refresh-type slots, effective impressions increase. In a design where ads rotate at regular intervals, longer dwell means more displays.

These are not stories of "we added ads," but of "the same ads were seen more reliably and for longer."

Attention metrics as a bridge to CPM

In recent years, the concept of "attention metrics" has been spreading among ad buyers. Adelaide's AU and the average display time in Active View are representative examples. It is a movement to evaluate inventory not only by "whether the ad was displayed," but by "how long the reader's attention was directed at the ad."

In my observation, more and more buyers are adopting these attention metrics. If so, inventory with long time-in-view may be valued by such buyers. This is the core of the logic that bridges "dwell time grows" and "ad CPM rises."

However, this is a hypothesis, not a proven fact. How far the adoption of attention metrics progresses, and how directly the dwell-time lift from audio translates into improved attention metrics, depends on future measurement. At this point I am writing it as an observation that "things may move in this direction."

I deliberately do not claim CTR

Here is something I will intentionally not write: an improvement in CTR (click-through rate).

A reader who is reading an article while listening to audio is immersed in the content. The more immersed a reader is, the less they click on ads. In fact, in the data I see from deployments, while dwell time and repeat rate grow, I have not heard clear reports that ad CTR goes up.

So I keep CTR out of this concept's claims. Writing "both dwell time and CTR get better" sounds nicer, but the risk of being torn down by someone on the ground saying "that's off" is greater. Reducing claims makes the argument stronger in the end.

"If they're not looking at the screen, isn't it padding?" — the sharpest counterargument

This is something I want to bring up myself, first.

A reader who is listening to audio may not be looking at the screen. A "listen-while-doing-other-things" reader may have their phone in their pocket with only the audio playing. If viewability is being counted in that state, the numbers grow, but the ad is not being seen. That is just another version of "the numbers went up but the reality didn't match" that I myself contributed to in the past.

This is something that troubles me too.

That is exactly why PUBVOICE is designed on the following premise: measure "read while listening" sessions and "listen while doing other things" sessions separately. Clearly distinguish sessions where the reader is looking at the screen from sessions where playback is in the background. The viewability argument holds logically only in "read while listening" sessions. Counting background-playback time as dwell time and using it as the basis for ad value is, honestly, a stretch.

To be candid, this measurement separation is not fully implemented at this point. What I am writing in this article is a concept — "we are thinking with this design as the premise" — not a report that "it's already done." But to discuss ad value in a way that is not padding, I believe this distinction is essential.

Placing this here as a concept

What I have written in this article is, in every case, a concept — not a results report of implemented features.

• As dwell time grows, the probability and viewed duration of ads outside the first view grow (hypothesis)
• Against the backdrop of spreading attention metrics, inventory that is viewed for longer is more likely to be valued (observation)
• On the other hand, without separating background playback from screen-attention, it looks like "padding" (self-awareness)

As I also wrote in Why we didn't go with an ad-supported model, I have no intention of rotating back to the side that adds more ads. However, if there is a path to honestly push up the value of existing ad slots, that is another meaning that lies beyond audio.

We are also considering the delivery of audio ads themselves (via VAST tags, etc.), but timing and specs are undecided. For deployment decisions at this point, I think the framing of "retention and measurement as the main purpose" is appropriate.

Related articles

• Why Web Media Ad Revenue Stagnates: Dissecting 3 Structural Problems — The structural limits of ad revenue, with the latest data on ad blocking and platform oligopoly
• What changes when web media articles are turned into audio — How dwell time and repeat rate moved with audio
• Why we built an AI audio SaaS for web media — Why I built an audio product coming from the ad side
• A field comparison of web media article audio services, by people on the ground — A comparison of BeyondWords, PUBVOICE, ReadSpeaker, and others
• Imitation in the AI era and delivery beyond text — What it means to hold a delivery route other than "reading" in the AI search era