In information retrieval some words are powerful / potent. They are really descriptive and get right to the point of what someone is looking for. Other words have little to no value. The reason the concept of stop words came about is that you really couldn’t tell much about a document by it including words like a, an, the, and, are, etc. The flip side of stop words are words which have a high discrimination value. Recently I was searching to see if there was a FexEx office in the town where my mom lives, and in spite of there not being one, Google still returned multiple pages (the home page and the store locator page) from the FedEx.com website in the search results. That was a great search result, and Google was smart to place more weight on the core concept word in the search (FedEx) while placing less weight on the location.

Words which have a low discrimination value may have a higher discrimination value when combined with neighboring words. Hot and dog might have a different meaning when they are next to each other. As explained in this Wired article:

Take, for instance, the way Google’s engine learns which words are synonyms. “We discovered a nifty thing very early on,” Singhal says. “People change words in their queries. So someone would say, ‘pictures of dogs,’ and then they’d say, ‘pictures of puppies.’ So that told us that maybe ‘dogs’ and ‘puppies’ were interchangeable. We also learned that when you boil water, it’s hot water. We were relearning semantics from humans, and that was a great advance.”

But there were obstacles. Google’s synonym system understood that a dog was similar to a puppy and that boiling water was hot. But it also concluded that a hot dog was the same as a boiling puppy. The problem was fixed in late 2002 by a breakthrough based on philosopher Ludwig Wittgenstein’s theories about how words are defined by context. As Google crawled and archived billions of documents and Web pages, it analyzed what words were close to each other. “Hot dog” would be found in searches that also contained “bread” and “mustard” and “baseball games” — not poached pooches. That helped the algorithm understand what “hot dog” — and millions of other terms — meant. “Today, if you type ‘Gandhi bio,’ we know that bio means biography,” Singhal says. “And if you type ‘bio warfare,’ it means biological.”

The concept of discrimination value also has value outside of search. If you get feedback from an anonymous person on a third party site it gets so much weight (maybe none). If you get feedback from someone who is not anonymous it gets more weight. If you get feedback from a paying customer it gets much more weight. One of the most powerful levels of discrimination is indeed payment. If a person pays you the (typically) you know who they are & they have expressed significant interest beyond what most people will do.

I think online business models which require payment from the typical user are not hyped and are not considered sexy because those sorts of models are often slow growth due to the penny gap and the requirement of greater trust to convert. Whereas a programming marketer can hear of a new network (say Pippers) and create 40,000 bogus accounts in an hour. The owners of Pippers can then talk about their explosive growth rate in the media, which earns them media coverage. In turn this increases their ability to raise capital and continue their “growth.”

But many of the social networks end up being a bag of smoke that will fade because they aim to bucket people as beings in a database and are so broad as to have little discrimination value. I have been reading You Are Not a Gadget and he compared the depersonalization on the broad social networks to the beauty of an oud forum he is a member of. Much like charging for admission, obscurity is a filter which improves the level of discourse.

Compare the comments on *any* niche topic site to what you find on Youtube. If you can show me a site which is consistently worse than Youtube (outside of site like 4Chan which specialize in creating campaigns to try to make epileptic people have a seizure) I will buy you a beer then next time we meet. 😀

My wife deleted her FaceBook account because she was annoyed at some people’s behavior on it. Part of the problem with the social networks is that they are so broad and so frictionless that your activities on them really don’t matter. As a marketer there are a couple ways to play such networks

  • largely ignore them

  • be friends with everyone
  • use bots

As a marketer the first of those options means you are saving your time for higher paying areas, and the second of those options means more people seeing more distribution of whatever content you create. But many of the helpful aids are at best dubious short term opportunistic ploys. The third option means you are one of the people who is going out of their way to make the web worse, but many will. 😉

Generally any given month I haven’t been on Facebook for more than 5 minutes outside of writing & targeting ads, or approving a few real “friends” and hundreds to thousands of other people who claim to be my friend. But if you message me on FaceBook there is a precisely 0% chance of getting a reply. 🙂

When FaceBook launched Beacon a few years ago they wanted to sell peer pressure as an ad unit. If brands can show that your friends did something then maybe that can help lead to a cumulative advantage sort of environment which has you follow along. Beacon was such a flagrant violation of user privacy that it was quickly shot down by the market. But with the new FaceBook like button, they are trying to use like button clicks to put your name on ads:

“Marketers have always known that the best way to sell something is to get your friends to sell it,” says Sheryl Sandberg, Facebook’s chief operating officer. “That is what people do all day on Facebook. We enable effective word-of-mouth advertising at scale for the first time.”

In the short run it may work, but in the longrun I don’t like the concept. The reasons are many.

  • You can agree with one particular thing a person says and like it while being nearly diametrically opposed to their general philosophy on life. For example, when we launched that “How Google Works” infographic last week one of the reporters who wrote about it also mentioned how sleazy and nefarious the SEO industry is, and yet he was willing to promote the efforts of an SEO because it was published on a blog with a sister acronym in the domain name. 😀 … Of the 3,000+ people who voted for us likely less than half of them know anything about me, or even my association with the site.

  • You can like one product from a company, but not like their other products. I have worked with GoDaddy as a registrar for years. And I have had no complaints on that front. But they also sell some search engine submission service that I would cringe to see my name promoting.
  • You click the like button once on one page. Years later the business you liked is trading in another area…they moved from remnant inventory to spyware, and you recommend them. 😉
  • An individual can have multiple lines of work. You might like Thom Yorke’s role in Radiohead, but you might not like his political views or his solo work.
  • Imagine when someone buys a car that you passively recommended which has a manufacturer defect. One of their loved ones gets killed and you eat the blame.
  • Just like businesses, people change over time. This is especially true in the area of business, where a former partner or friend goes out of their way to betray your trust and screw you.
  • How do likes work with 301 redirects? How do they work when the content of the page shifts from genuinely useful to hawking trash with a hyped up sales letter?

A like doesn’t have much discrimination value. And it shouldn’t last very long. Why did you like something? When did you like it? Who knows.

Did you like Toyota right up until the breaks didn’t work? After you get out of the hospital, how do you feel when your friend asks you why you are still promoting their products? Did you work for a digital sharecropper overlord like Jason Calacanas who required you to push their junk elsewhere? How did you feel when your friend asks you why you are promoting his trash after he canned you with 1 week notice while boasting how they are nearing break-even and have over 8 years of cash in the bank?

Once people experience that will they become jaded and stop recommending things?

And if there isn’t a backlash against the like button then given enough time one of your friends will like almost anything. It doesn’t matter the product/service/offer … if your pool of “friends” is wide enough then one of them is receiving an affiliate commission for pushing something, one of them owed a favor to the merchant, and one of them liked the merchant because they picked up a tab in the bar last month.

A wave of 100 million blond hair 18 year old girls who are lonely have joined FaceBook friending up with the desperate and then promoting scammy wares to them via automated clicks of the like button. And then of course there will be services like SpikeTheVote.

Sure a fad might work in the short run, but given enough time and there will be friend recommendations for almost anything. Once the novelty wears of does any of it matter?

In time any database record can be an ad targeting mechanism. Will I be promoting some of the products my thousands of “friends” create or endorse by a click of the mouse which changes purpose after the fact?

At first online petitions were powerful because they seemed to have mobilized swaths of people. But then people realized that a vote represented nothing more than an automated form submission and clicking send. 2 clicks of the mouse. Not much discrimination value.

More: continued here