Building network effects is hard . . . but once a startup has gathered traction, network effects can also reverse itself pretty quickly. Its not all rosey, sit back, and collect the money. . . Case in point. . . Miva networks.

The company generated 219 million paid clickthroughs during the quarter. That’s a 13% reduction from what it produced a year earlier, despite growing its base of advertisers by 20%. This isn’t a mixed bag — it’s just bad. Revenues falling faster than clicks means that the company is generating less money per click.

Found it on Motley Fool this morning while checking on my portfolio. . .

These guys was one of the pioneers of the PPC ad network model. However somewhere along the way, “bad” sites got into their network and they are caught in a situation where they are unable to sacrifice critical mass for quality. And thus, the downward spiral begins. . .

1) bad sites join network
2) low quality traffic click through ads
3) low conversion for advertiser
4) bids for words goes lower
5) high quality sites leaves cause they can get higher monetization somewhere else
.. . 1) more bad sites join network that got kicked out of google/overture

So even though partners are increased at 20% word pricing droped significantly more . . . this is a cautionary tale for all the web 2.0 plays out there. . . if you attract the wrong kind of community initially, you are building the wrong kind of network effects that could quickly deterioate. . . . for example, if Digg attracted spammers when it was initially launched, they would not have become what it is today. . . if craigslist was over-ran by best-buy when it first launched, no one would go there . . . chose carefully where you launch your site. . . not all traffic is good. . .