Conversation

Notices

1. Embed this notice
   Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 05:21:11 JST Paul Cantrell

   A question:

   As Zeynep Tufekci remarked, we’ve built the machinery of an authoritarian dystopia to get more people to click on ads.

   An astonishing percentage of our technology, our industry, our energy, our human inventiveness, our •lives• pour into this. Targeted advertising is now a dominant force in our world.

   The question: Does this pay off for the advertisers?

   Leaving aside all the social costs and the moral questions, did it •work•? Is advertising radically more effective than it used to be?

   1/

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 05:25:27 JST Paul Cantrell

     in reply to

     I’m not even sure how to properly formulate this question. What would it look like for targeted advertising to be radically more effective? Well…

     Do companies spent a much smaller fraction of their budget on advertising than they did 20 years ago, because it works so much better now?

     Or do consumers spend a larger percentage of their discretionary income on things they’d never have bought without advertising?

     Or are people vastly happier with things they buy, because advertising is making their purchases so much better informed?

     (Reply speculation is not useful here; I’m looking for things we could actually gauge from empirical data.)

     2/

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 05:27:14 JST Paul Cantrell

     in reply to

     I’m much too far out of my depth to know how to approach this, but surely there’s some way to study this question. Has anybody done that?

     /end

   • Embed this notice
     Franceska Mann (franceskamann@freeradical.zone)'s status on Wednesday, 08-Jan-2025 05:37:06 JST Franceska Mann

     in reply to

     @inthehands

     Cloud capitalism can't be judged by the same parameters of past advertising. What AI does is learn about us, then sell ourselves back to us. it learns our reactions, and how to make us click. At the same time, it's feeding us tiny doses of dopamine. It's drugging us into compliance. This is not our grandfather's advertising. This is something else, and it's far more dangerous.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 05:37:06 JST Paul Cantrell

     in reply to

     • Franceska Mann

     @FranceskaMann
     Yeah, “advertising” may be the wrong word entirely.

     And yet…if, say, a social media site is creating hypertargeted addiction for us, isn’t advertising ultimately the income source in most cases? Doesn’t most of this money ultimately trace back to advertising?

     Or maybe the creation of this new kind of attention-capturing media isn’t about making marketing work at all; it’s about tech giants making themselves an unavoidable middle player in the whole damned economy.

   • Embed this notice
     Jay Stephens (jaystephens@mastodon.social)'s status on Wednesday, 08-Jan-2025 06:07:55 JST Jay Stephens

     in reply to

     @inthehands
     I'm pretty ignorant about this, but it strikes me that the same industry that built all this big data infrastructure, has already used it to answer the question, and the answer is "mostly no', which is why we hear nothing empirical, when all the tools to provide detailed empirical proof are already built.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 06:07:55 JST Paul Cantrell

     in reply to

     • Jay Stephens

     @jaystephens
     I of course suspect there’s a lot of truth to this, and suspect that the real answer is messy and disappointing — but it’s the sort of thing one would want to see really worked out through careful study.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 06:11:03 JST Paul Cantrell

     in reply to

     • Marko Karppinen

     @karppinen
     I can believe this: not any better for the advertisers, just an enshittification play from new tech+media giants making themselves the inescapable conduits for everything.

   • Embed this notice
     Marko Karppinen (karppinen@mastodon.online)'s status on Wednesday, 08-Jan-2025 06:11:04 JST Marko Karppinen

     in reply to

     @inthehands podcast ads seem to work great. That’s an example of the old contextual targeting model being used online. You don’t know the audience but mattresses still get bought.

     But the contextual model isn’t available at scale because traditional media was killed and you can’t target contextually on social media hellsites where all the eyeballs are.

     Hard for big advertisers to find a way back. P&G spent $10B on advertising last year. TV is the only contextual scale play left.

   • Embed this notice
     J Miller (jmmaok@mastodon.online)'s status on Wednesday, 08-Jan-2025 06:23:07 JST J Miller

     in reply to

     @inthehands

     You may find this lit review of interest https://www.tandfonline.com/doi/epdf/10.1080/00913367.2017.1339368?src=getftr&utm_source=wiley&getft_integrator=wiley

     Something to keep in mind is that we can’t really compare with the past, because things like magazines, TV, newspapers and their audiences are not what they were before. So an auto dealer that used to advertise heavily in the Sunday paper needs to try something different.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 06:25:00 JST Paul Cantrell

     in reply to

     • J Miller

     @JMMaok
     Yes, very much of interest!

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 06:51:49 JST Paul Cantrell

     in reply to

     Refining the question somewhat based on replies:

     What is the •difference• in effectiveness between purely contextual ads (i.e. served based on adjacent content, search term, etc.) vs. ads served based on individual user profiling (i.e. info sites collect at the user & data brokers sell)?

     Does the surveillance pay? For whom?

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 07:17:15 JST Paul Cantrell

     in reply to

     A different refinement of the question, in a different direction:

     Who’s buying from data brokers? What do they think they’re buying? Do they have any idea whether it’s effective for them?

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 07:18:11 JST Paul Cantrell

     in reply to

     • Jen

     @JetlagJen
     I have to imagine they’ve done some form of this, and also have to imagine that the results are (a) confidential and (b) methodologically near-useless.

   • Embed this notice
     Jen (jetlagjen@geekdom.social)'s status on Wednesday, 08-Jan-2025 07:18:13 JST Jen

     in reply to

     @inthehands the people who own platforms could do a/b testing on that last question.

     Randomly select a sample of users. Make half of their ads based on their current profile, the other half based on a null profile. Count the clicks.

     Of course, those who own the platforms also have a vested interest in selling their super-effective micro-targeted ad space, so they never would in case they were wrong.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 08-Jan-2025 12:00:12 JST Paul Cantrell

     in reply to

     I realize the first of my two refinements above slightly changes the question, to wit:

     It’s plausible that there’s a sort of “you have to yell in a noisy room” effect here. Maybe there’s a cap to how much advertising can do, to how much any given person can be influenced, but the arms race of advertising means that achieving the same results now requires far more effort.

     IOW, it’s possible that individually targeted advertising works better than context-targeted advertising •today•, but no better than context-targeted ads •used to•. How would one determine that, I wonder?!

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Wednesday, 08-Jan-2025 12:14:38 JST Rich Felker

     in reply to

     @inthehands Spoiler: no. The advertisers are getting scammed but no one has the guts to tell the boss he has no clothes.

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Wednesday, 08-Jan-2025 12:16:38 JST Rich Felker

     in reply to

     @inthehands The real lucrative market is selling to stalkers, parties who want to engage in illegal discrimination, law enforcement, etc.

     "Targeted ads" is just the laundering of intent.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Monday, 03-Feb-2025 05:06:24 JST Paul Cantrell

     in reply to

     • Jonas

     @adoranten
     FOMO drives so much of business. I've been known to quip that executives basically spend their professional lives ping-ponging between fear of change and fear of missing out.

   • Embed this notice
     Jonas (adoranten@tldr.nettime.org)'s status on Monday, 03-Feb-2025 05:06:30 JST Jonas

     in reply to

     @inthehands IMHO, so much of online marketing spend comes down to “irrational exuberance”: Companies see that their competitors are spending X amounts of $ on programmatic advertising and they don’t want the online space to be full of the competitor’s ads so they make sure to spend equal amounts. This despite everyone hates advertising and the efficacy of it is extremely unsure and extremely hard to verify.

   • Embed this notice
     Jonas (adoranten@tldr.nettime.org)'s status on Monday, 03-Feb-2025 05:06:31 JST Jonas

     in reply to

     @inthehands Your thread made me recall this article (and another one linked at the end of it) and then I forgot about your thread, but now, a couple of weeks later, I remembered it again… Also, Augustine Fou is a critical researcher in the field. https://thecorrespondent.com/125/the-non-sense-of-online-advertising-when-the-numbers-dont-add-up