To wrap up, Charlotte Rommerskirchen asked, "Why are Podcasts Ads so bad?" in her look into how advertising's past, present, and future haunt the podcast ad landscape.

Then, Koray Caliskan and @addiemcgowan gave a double paper on Google's superplatform as a demarketised supply chain, exploring the complexities of "adtech spaghetti" in "Prodexchange: Digital-Industrial Economisation in Google Search Supply Chain."

@DonaldMacKen kicked us off by asking "What Your TV Knows About You," bringing a consideration of #adtech's embodiment in domestic settings (think: on the sofa, do you lean in, or lean back, for certain ads?).

Our project team recently presented some work-in-progress at our Platforms, Advertising, and Power event, captured by a live illustrator. Check out the visual summary here, and special thanks to Illustrated Live for capturing our talks!

Seems like this is quite a genre. One of our friends told us about soothing music apps that help you get to sleep. There, the ad *could* appear, but by the time it's loaded the screen is probably dark and if the app has worked, the user is off in the land of dreams ...

Puzzled by "fireplace app inventory" we asked a couple of our friends in AdTech. It turns out to be ads that aren't visible on screen but loaded in the background of apps that simulate fireplaces, and sold as good quality video ad inventory.

Is your cosy fireplace "showing" you ads? We've been researching digital advertising for 4 years, and keep getting surprises. The latest was triggered by a reference in AdExchanger by John Gentry of OpenX to "fireplace app inventory".

Donald MacKenzie discusses our project's work in his da Vinci Medal Address, thanking @SocHistTech for this honour. Check out his talk which has just been published in the Journal Technology and Culture: muse.jhu.edu/article/933102

As both the IDFA saga and the controversy over third-party cookies show, preserving privacy is often in tension with other priorities such as fostering competition. But a productive balance between them is possible, and astute public policy can help find it.

Apple could further obscure user identity by encrypting IP addresses and using a system of relays, but this could face government opposition and increase energy consumption.

Some are turning to IP addresses for user identification, though this is less precise and more contentious, existing in a gray area of Apple's policies.

Post-ATT, only about 20% of iOS users opt-in to tracking. This has made user acquisition more difficult and expensive for game developers, keeping those match-three whales elusive to advertisers.

IDFAs allowed for "deterministic targeting," enabling advertisers to build comprehensive user profiles across multiple apps and behaviors. But Apple's App Tracking Transparency (ATT) policy in 2021 severely restricted IDFA use, disrupting the mobile advertising ecosystem.

The submerged data iceberg is enormous and opaque, and requires "identity resolution" to connect up behavioural data. Historically, unique device identifiers (IDFA for iOS, GAID for Android) enabled you to tell when the same phone was involved in two or more actions.

But below the water line is the vast, deep data that can make machine learning optimized advertising much more effective than human targeting. Optimization happens based on conversions, leveraging behavioural data, not starting from sociodemographic characteristics.

In our latest @LRB article, we share how @eric_seufert describes ad platforms as data icebergs. The tip above the water line includes the characteristics of audiences advertisers focus on, like age, gender, and interests. lrb.co.uk/the-paper/v46/n16/…

As #adtech peeps know, game advertisers therefore use sophisticated machine learning algorithms to target potential players. Platforms like FB and Google use extensive amounts of user data to optimize ad targeting and delivery.

This also indicates how hard it is to target match-three whales: only 2 or 3 players in 1,000 are whales, we've been told. Targeting this group based on their behaviour, as opposed to their demographics, is an example of a common challenge faced in online advertising.