AllAdvantage

» see alladvantage.com as it lived, on the wayback machine

~ the last words ~

in 1999 i was fifteen and on a 33.6k modem in ankara that the family had to ration. i was already the kind of person who read tech news in english, because turkish coverage of the internet at that point was a paragraph at the bottom of the technology page. alladvantage came up in that context. you could not actually use it from turkey, payouts went to US bank accounts and the user agreement was very clear about that. but the pitch traveled.

the pitch was: install our viewbar program, which puts a small ad-strip at the top or bottom of your browser, and we will pay you fifty cents per hour while you browse. you also get a cut of every hour your referrals browse. they pulled in nearly two hundred million dollars in venture capital on this idea between 1999 and 2000. they actually paid out about one hundred sixty million of it before the music stopped.

then the music stopped. the dot-com bust took online ad spend down forty percent in eighteen months, and a company whose entire business was paying retail users with the proceeds of selling them as eyeballs to advertisers stopped working as soon as the eyeballs stopped clearing for fifty cents an hour. they shut in february 2001.

~ the record ~

~ how it landed ~

the founders called it the "infomediary" model, an idea floated in a 1997 john hagel book that argued users would eventually demand a cut of the data they were generating about themselves. alladvantage hired the first corporate chief privacy officer in tech (ray everett-church, august 1999) to make this story credible. for a few months in 2000 they were ranked among the top twenty most-trafficked websites in the world.

~ the fall ~

the math never quite worked. cpm rates were collapsing in real time, and the real per-user advertising value was below the per-user payout long before alladvantage admitted it. on top of that, the system was rampantly gamed. users wrote scripts to fake browsing activity ("autosurfers"), so the per-hour ad value was even lower than the headline numbers suggested. the company spent its second year writing fraud-detection code and clipping payout rates rather than building a business.

~ what got buried with it ~

the broader idea (you should be paid for the data you produce) just kept living. it came back as data-dividends discourse around facebook in the 2010s, then as crypto pitches like brave's BAT token. by 2024 it was back as "you should be compensated for your training-data contribution to LLMs." every five years someone reinvents alladvantage, raises a hundred million, then runs into the same cpm math. one of these decades it might work. it has not yet.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »