The challenges facing publishers are formidable.
For all the advancements in ad tech, publishers are still more rewarded for squeezing another ad on a page or lighting up a duplicate supply path than creating a monetization experience that extends a user session.
While no legitimate publisher aspires to work with myriad partners or overwhelm their pages with ad placements, if the economic realities and outcomes of our industry continue to reward these practices, it raises the question:
What would a catalyst for meaningful change look like? Judging by the standing-room-only crowd at Sharethrough’s recent Green Media Summit, sustainability in ad tech may be a formidable candidate.
Though I remain a climate neophyte, it doesn’t require seasoned expertise to understand that decisions I make about ad layout, SSP partnerships and bidstream efficiency make a difference on my carbon footprint as a publisher.
Conceptually, this makes sense, but it can be a divisive conversation for publishers that reasonably equate removal or reduction as detrimental to business outcomes. Whether packaged as an environmental benefit or demand path optimization, elimination may be the purest form of supply-path optimization. But, unfortunately, it’s not always the most realistic.
What about improvements that come at no cost to revenue? Is it possible to improve your efficiency as a publisher while also reducing your carbon footprint?
Focusing on requests
Ad requests begin the ad-serving process. They are publisher signals sent to enabled SSPs that define a user and ask for an ad. SSPs take those signals and send “bid requests” to enabled DSPs. This happens per user, per ad unit across every active exchange and supply path that a publisher runs. As a result, the compounding effects can be dramatic, especially in instances of multihop reselling.
While SSPs and DSPs are built to handle large volumes of “queries per second,” there is a financial and environmental cost associated with being inefficient – one that will be increasingly important to optimize as advertisers, DSPs and SSPs fight against increased supply-chain costs and output.
With the goal of reducing total ad requests to SSPs, we’ve been running tests over the last 10 months. So far, two results stand out as having significant impact. These features are integrated via Prebid and require a small amount of code to monitor and manage bid requests and auctions. They have significantly reduced requests while showing no measurable impact to monetization.
Test 1: No ads if idle
Like most publishers, we refresh ad placements only when in view. We wondered what would happen if we added an additional requirement that the user also must be actively engaged for an ad to qualify for refresh.
To test this, we implemented a change on page that started counting defined events like click, scroll or touchstart. If the counter extends beyond 60 seconds, the user is deemed idle, and we stop requesting ads.
Prior to testing, I was confident we would reduce requests with this change. What gave me pause was that the potential reduction of total monetized impressions could lead to a negative impact on revenue.
The actual results were encouraging. Desktop bid requests were reduced by 20%, while average desktop viewability improved by 14%. Without any changes to variables like flooring, unpaid impressions per game (impressions with bids that failed to clear our price floors) decreased by 11%, while paid impressions increased by 5%.
Test 2: Smarter request
For this test, we wanted to understand what would happen if we used SSP participation signals to throttle our ad requests.
The concept was simple. By ad unit, if Prebid SSPs fail to respond with a bid above our dynamic price floor on three consecutive auctions, we stop requesting that SSP partner. We check again a few minutes later – if they return a qualifying bid, they resume ad serving.
Since this initiative applies to all sessions, we saw an even greater increase in efficiency. Total ad requests decreased by 30% while SSP fill rates (note, this is not the same as win rate) improved by over 50%.
A step toward greater efficiency
In aggregate, these features reduced total ad requests to our SSP partners by more than 50% without any measurable revenue impact. Although that headline lacks the punchiness of a case study touting an immediate revenue increase, these changes have made our inventory more compelling due to our increased efficiency.
As publishers, we tend to prioritize changes that show immediate revenue lift. But it seems reasonable that the benefit of increased efficiency wouldn’t immediately appear in the top line, even if it could have positive impacts down the road.
PubMatic, for example, cites publisher supply efficiency as one of their evaluation criteria for inclusion in curated marketplace deals.
And Rocky Moss, CEO of ad quality firm DeepSee.io, points out that DSP SPO efforts would benefit from a function that removes inefficient supply paths, “and publishers can also benefit from policies that reduce bid volumes via inefficient supply paths.”
There’s also a case for user-facing benefits. With fewer ad requests being made, a website is more responsive, especially deep in a session.
Obviously these changes are not a panacea to ad tech’s carbon emission challenges. But they can represent a step forward and are worth exploring for publishers looking to improve overall ad serving efficiency.
Whether you consider these features the development of more conscious supply or demand path optimization under the veil of sustainability, the label matters far less than the impact their adoption represents.
“The Sell Sider” is a column written by the sell side of the digital media community.
Follow Emry Downinghall and AdExchanger on LinkedIn.
For more articles featuring Emry Downinghall, click here.
