Project Wonderful

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

~ the obit ~

you needed a website with an english-speaking readership before project wonderful was any use to you. in 2006 i had a signature on a turkish hardware forum listing my motherboard and how much ram was in it, which is not a property anybody buys advertising on. so this one belongs to a scene i read from the outside, the anglophone webcomic internet, where a few thousand people were quietly running small businesses off daily strips.

ryan north was one of them. he drew dinosaur comics, the strip that reuses the same six panels every day, and in early 2006 he was complaining to a friend about google adsense. you got paid per click or per impression, and neither the advertiser nor the site owner had much say in what actually appeared on the page. eight months later he had shipped an alternative out of toronto.

the model was an auction on time rather than on clicks. an ad box on a given site was permanently up for sale, you bid what a day of it was worth to you, and if your bid was on top the box was yours until somebody outbid you. no contract and no minimum spend. the site owner could see exactly who was buying and reject anything they did not want on their page.

at its height around ten thousand publishers ran project wonderful boxes, over five thousand of those webcomics, against roughly a hundred thousand advertisers. it carried the ad inventory for a large slice of the webcomic world at once, qwantz and questionable content among them, which meant that for about a decade a whole cottage industry ran its rent through one small canadian company.

~ the rap sheet ~

~ how the auction worked ~

the part that made it feel unlike advertising was the floor. the minimum bid was zero. if nobody wanted a box, it went to whoever asked for it at nothing, which meant small comics could advertise on other small comics for free during quiet weeks and the boxes were rarely empty. a lot of people's first ever ad campaign was a few cents a day on a strip they liked.

the mechanism was second-price, mostly. you paid one increment above the next bid down rather than your own number, so bidding your true value cost you nothing extra. it was not quite airtight, and people who thought about it noticed the same edge: when your bid stood alone above a much lower one, the increment rule could still charge you more than you needed to pay, which quietly rewarded bidding low and waiting. that is a small flaw in an auction that otherwise ran for twelve years without anybody gaming it into uselessness.

~ the rules it kept ~

project wonderful did not follow the rest of the industry anywhere. no behavioral tracking, no cookies following readers between sites, no auto-playing video, no ads that pretended to be a system dialog. every other network spent the 2010s adding exactly those things, because that was where the money went once facebook and google had built better targeting than anyone could match.

the company knew what that cost. the closing statement said it plainly: the ways left to compete were invasive tracking and tolerating the kind of ads they had spent twelve years refusing to carry. they had built the better option and the market had stopped paying for a better option. the underlying shift was simpler still, and they named it. large sites like facebook do everything they can to keep readers inside, rather than sending that traffic out to individual websites, and an ad network for individual websites needs individual websites to be where people go.

~ the last day ~

the shutdown ran on a published schedule, which was more courtesy than most closures manage. announcement on 11 june 2018, no new accounts from that day. ad serving stopped on 11 july. the deadline to withdraw your balance or do anything else with your account was 1 august, and the servers went dark on 6 august 2018. the closing post counted the run out to the day: twelve years and six months, then another twelve days on top.

what went with it was not really the ad network. the webcomic people mostly moved to patreon, where readers pay the artist directly and there is no ad box at all, which is a better deal for anybody with a loyal audience. what actually died was the idea that a small independent site could sell its own attention on its own terms to a small independent advertiser, without either party being surveilled or signing anything. nobody has rebuilt that, and after 2018 nobody has really tried.

@}-->--

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

« previous  ·  random  ·  next »