As websites become more complex and interwoven with marketing partners, customer tracking partners, statistical compilers, Business Intelligence warehouse services, and whatever, we run the risk of slowing down the overall experience . We are playing havoc with a visitor’s perception of privacy and intention that someone thinks they are only dealing with one company, one website.

Each of these partners invariably wants to place a 1-pixel image somewhere at the bottom of the page. What could one little pixel hurt, right? Along with this pixel are passed a whole host of variables about the origin of the browser, the last page they were on, how long they’ve been there, shoe size, whatever. Well, some sites have two or three– others have a dozen. To me, these things are suckerfish– small little animals along for the ride that eat up the leftover bits from whatever the Great White Shark of a website had for lunch that day. Sure, a suckerfish doesn’t seem so bad– but if a dozen are stuck to the belly– they’ll start to slow things down a bit, and probably cramp the shark’s style (to carry the metaphor):

1. Suckerfish usually come from an external site, so there is no load on the host website’s infrastructure. But this remote call means an additional DNS lookup for each suckerish, along with whatever network problems may be between the browser and the suckerfish’s home. This means that every page of every site is making a half-dozen DNS lookups. (Note to self: Buy stock in Foundry and Cisco)
2. If a suckerfish is late or slow, that wheel in the corner keeps on spinning– cueing the browsing customer that the website “isn’t finished downloading yet”. Well, guess who gets the blame for that? That’s right– the shark, not the suckerfish.
3. Most of these suckerfish are asking for the same data points 80% of the time: search engine source, browser type, time on page, cookie ID. Why can’t someone invent a ‘universal suckerfish‘ as a service, and pass that along to all the other fish? I wouldn’t mind so much if there was just one or two on my site.

I fear the issue will only grow in size– as more and more marketing schemes are invented (not unlike financial derivatives and options), the need for more and more data will increase, and site owners will be tempted to tap in by filling their footers with these suckerfish. The problem may soon come, however, that as these fish return some marketing data (upsell, crosssell, recommendations, marketing-spooge) in real time to be displayed, we may see race conditions– suckerfish A returns some data to the site for display, which in turn sets off new data from suckerfish B, which in turn sets off C, or worse, resets something back to A, and we have a race condition/runaway train. The problem here is this would be difficult to detect, as the site owner may or may not be tracking all of these remote connections, nor have access to the data structures and algorythms that are driving these fish. (tracking would require more suckerfish, right?)