« My Question for Twitter’s Founder | Main | Economics in Wired: Two Articles Worth Reading »
When Signaling sends the Wrong Signal
By Chris | June 22, 2009
Google became the world’s most popular search engine by radically increasing the relevance of search results. It was able to do so by rewarding sites that received a lot of links from other sites. Links from popular sites were worth more than links from obscure sites. The popularity, or rank, of a site became known as PageRank. This blog currently has a rank of 5. MarginalRevolution.com has a PageRank of 8. PageRank isn’t linear though. Supposedly, a PageRank of 8 is 1,000 times as popular as one with 5. Anyway, after inventing PageRank, Google ran into a problem. People with websites had an incentive to create lots of links back to their pages. Comments and wiki’s became infected with irrelevant links. To combat this problem, Google created a way for owners of websites to signal which links should count and which one shouldn’t for ranking purposes. By tagging a link as “nofollow,” a web site could tell Google to ignore the value of certain links. Today, almost all comments on blogs and all external links on Wikipedia are nofollow links. Spammers no longer get any credit for their backlinks. Problem solved. Or is it?
Nofollow has become increasingly popular. Linking to another site upright conveys upon them a positive externality by improving their reputation. The recipient is helped and the source gets nothing in return. Not surprisingly, it is underprovided. Some websites have a blanket “nofollow” policy. They only give conventional links to recipients that pay. Advertisers pay lots of money to get followed links from credible sites. When signals are bought and sold, they lose their value. Links that are followed are no longer the most relevant, they are the most profitable. On some sites, “followed” links are less relevant than organic “nofollowed” links. When everyone knows the signals, they lose their value.
This has many applications. Signals are constantly evolving. Fashion is a good example. Brands come and go, as a signal is recognized and copied. For example, The North Face is a popular retailer of outdoor clothing and gear. Initially, it was the purveyor of apparel and gear for an elite group of outdoor enthusiasts. If you wore The North Face, it signaled that you skied, hiked, and spent time outdoors. The signal was reliable. But, it was an easy signal to imitate. Americans liked being perceived as outdoor enthusiasts. The North Face become popular with the masses. Today, someone wearing The North Face has some extra money and brand awareness. The signal’s meaning has made a 180.
An acquaintance of mine recently joined the Peace Corps. Before he left, I asked him why he decided on the Peace Corps. “Simple,” he said, “It’s my ticket into a good law school.” The experience is probably beneficial nonetheless, but it’s overt attempt to game the system. Admissions committee are looking for compassionate, interesting people. In the past, the Peace Corps was a good signal about your character. Today, it says less about your character and more about your ambition.
I think this explains why, on a gut level, I don’t like some people who send all the right signals. The guy who is too friendly, dresses too perfectly, has impeccable manners, says all the right flattering things, and is a great student. If you are sending all the right signals — if they are all conscious choices — all I know is that you can play the game well. I don’t know if you are kind, or curious, or interesting. Everyone does this to a certain extent, but most people allow some of themselves out without conscious calculation.
Just because they can be copied, doesn’t make signals meaningless. It just means that certain signals are more important than others. Signals are most valuable when they are costly to send. Job market candidates in economics can send three signals to the schools they are most interested in. These are indeed valuable. A thoughtful gift is a valuable signal. Sticking with an unpopular idea or friend is a valuable signal. We’re constantly on the lookout for such signals.
Does Google need a more complicated signaling system? No, they’re doing pretty well on their own. My guess is they calculate the value of the sites to which a website links. If you link to quality sites (especially sites whose quality is not yet apparent) your signal is worth a lot. If you only link to advertisers, your signal is worth less. Ultimately, in life and on the internet, signals mean the most when we don’t know how to send them.
Topics: College, Economics, Internet, Web 2.0 | 10 Comments »
June 8th, 2012 at 3:11 am
Hi there, simply was aware of your blog through Google, and discovered that it is reallyinstructional. I am gonna watch out for brussels. I will appreciate when you proceed this in the long run. Lots of other people will probably be benefited from your writing. Cheers!
March 18th, 2015 at 11:58 pm
That’s a creative answer to a dilufcfit question
April 14th, 2015 at 1:17 pm
I am also writing to make you know of the fabulous experience my princess went through going through your webblog. She learned several details, most notably how it is like to have an amazing giving character to get certain people without difficulty have an understanding of a variety of problematic subject areas. You really did more than her desires. Many thanks for producing these interesting, trustworthy, explanatory and even easy guidance on that topic to Janet.
May 8th, 2015 at 2:03 am
Economies are in dire straits, but I can count on this!
May 27th, 2015 at 11:20 pm
I didn’t know where to find this info then kaboom it was here.
May 28th, 2015 at 1:19 am
It’s like you’re on a mission to save me time and money!
May 28th, 2015 at 1:20 am
Your answer lifts the intelligence of the debate.
May 28th, 2015 at 1:26 am
Just what the doctor ordered, thankity you!
August 13th, 2015 at 1:21 am
Help, I’ve been informed and I can’t become ignorant.
September 17th, 2015 at 10:55 am
A provocative insight! Just what we need!