Usenet, the newsgroup system that carried the internet's arguments before the web, kept a school calendar. Every September a wave of American college freshmen got their first network accounts, found the discussion groups, and asked the questions each group had answered a thousand times. The regulars rolled their eyes, corrected them, and a few months later the freshmen were regulars themselves. The Jargon File, the dictionary the early internet kept of itself, described the annual wave without any alarm: "These relatively small drafts of newbies could be assimilated within a few months."
September 1993 did not recede. Commercial providers had begun connecting their paying subscribers to Usenet, and the annual wave became a permanent tide. The next sentence in the Jargon File entry is the eulogy: "But in September 1993, AOL users became able to post to Usenet, nearly overwhelming the old-timers' capacity to acculturate them." The intake that had processed every previous September never caught up, because the incoming class never stopped arriving. In January 1994 a regular named David Fischer wrote that the month would be remembered as "the September that never ended." The name became the name of the era.
Here is the part I want to keep from that story: Usenet was not wrecked by the number of newcomers. It was wrecked by the ratio of newcomers to people who could teach them the place.
Assimilation is one to one
A community transmits its norms person to person. Someone has to read the clumsy first post, answer it, correct it, and model what good looks like. That work does not parallelize. Twice the arrivals means twice the teaching hours, and the teachers are volunteers with day jobs. Past a certain rate, the newcomers stop learning from the old-timers and start learning from slightly earlier newcomers, who pass on norms already diluted by one generation of photocopying. Nobody decides anything. The culture just gets resampled until the original is gone.
Notice where the blame does not belong: nothing about a 1993 AOL subscriber was incompatible with newsgroup culture. They were not worse people; they arrived faster than the culture could reproduce itself. And that is my blunt opinion about growth generally: every audience has an absorption rate, and nearly every growth plan is written as if that rate were zero. The plan counts the arrivals and never counts the teaching hours the arrivals consume.
The front page turns on its editors
If Usenet is the flood story, Digg is the mirror image: a site that broke itself by demoting its own assimilators. At its peak Digg drew more than 40 million unique visitors a month by Hitwise's count, and in 2008 it was reportedly in Google's sights at around $200 million. The site worked because a few hundred power users submitted most of what reached the front page, and the community's votes decided what millions saw each day. Those power users were Digg's teachers. They defined, every single day, what a Digg front page was.
In the last week of August 2010, Digg launched version 4. It removed the bury button, removed the queue where users vetted upcoming stories, and let established publishers pipe their feeds straight onto the page. The redesign was reportedly aimed at taking emphasis away from exactly those power users, and the power users got the message. On Monday, August 30, some of them declared "Abandon Digg Day" and filled Digg's own front page with links to Reddit - at one point, six of the top seven stories pointed at the rival. Hitwise put Digg's U.S. visits down 26 percent in the weeks after the redesign. In July 2012, after raising $45 million, Digg sold to Betaworks at a price The Wall Street Journal reported as $500,000.
Two sites, opposite errors, one mechanism. Usenet accepted arrivals it could not process. Digg dismissed the people who did its processing. In both cases the number that decided the outcome was not the size of the audience. It was the capacity of the culture to absorb, in any given month, either its newcomers or its own losses.
A small board on purpose
A club owner and a fire marshal agree on something most growth decks ignore: a room has a capacity, and past it, the people inside stop getting the experience the room exists to give.
I run Bid Spotlight, a paid-rank leaderboard, which means I sell the thing this piece warns about: attention, on demand. So the discipline has to live in the machinery rather than the sentiment. As I write this, the November board has no paid spots yet, the homepage counter reads 3,564 visitors since October 2, and the only brands on the board are our own house spots. The minimum bid is one dollar. Small is not a phase this site is stuck in. It is the setting.
The absorption machinery is in the rules, in plain sight. Every submission passes a review against our halal and safety policy before it can rank, which makes the intake rate something a human can actually enforce. The board resets on the 1st of every month with the minimum bid back to $1, so no amount of money compounds into permanent tenure; every month re-earns its ranks. A first bid must beat the current top by at least $1, which keeps the auction legible to the people already on the board. And this blog, twenty-four articles in twelve days, about two a day, is the other half of the same machine: teaching the room what the room is, one post at a time.
I will take that trade knowingly: growth at the speed the screen can process, over a board that swamps and stops meaning anything. If you are building an audience, count what Usenet and Digg both suggest counting - not how many arrived, but how many you can absorb - and if our screen sounds like yours, submit your site at bidspotlight.com/submit.