Facebook, Bing & Google: 10 Reasons Why Social Search is Flawed

18th May, 2011

Last October I posted on how Bing and Facebook were going to integrate their search offering, to bring social signals deeper into search results. This week Bing announced even greater integration with Facebook. This raises a number of concerns, which I will go into below. But first lets look at what new Bing features you can expect to see.

Trusted Friends

Bing mentions the Friend Effect as the reasoning behind its move towards greater social integration. According to research the Friend Effect states we are 90% more likely to make a decision once we get a seal of approval from friends. With this in mind the new Bing results will show us what our friends have ‘liked’ not only what stories, but also what pages and sites they appeared on. With personalised results, content liked by friends will now be pushed to page 1.

Collective IQ

You can’t rely on friends to provide the answers to everything, that’s where Collective IQ comes in. This call’s on the collective expertise of the web in order to find the best content for your search. Collective IQ highlights popular content from within websites. For example the image below displays the most ‘liked’ content within a recipe website.

Social Search: The Problems

There’s a number of issues with Social Search that can skew our results. I say our results because, each individuals results will differ due to these social signals. One of the more obvious problems with social search I highlighted some months back –  the race for likes.

1. People sometimes don’t really ‘like’ the Page

The race for Likes is where brands go all out to get new Likes, because now each Like can help boost a businesses ranking within search. People sometimes don’t really ‘like’ a Page, they just want to receive the incentive to Like it, then hide the page from their Newsfeed. I know people who Like brands they don’t really even know, just to enter competitions. Social Search depends on every person being accurate in their Likes, their tastes not changing and when they do change they remember to Unlike the page and not merely hide that brand from their Newsfeed.

2. Doesn’t allow for changes in taste or opinion

Some people may take the time to Unlike Pages and not just hide the content because they want to de-clutter their Newsfeed. Thinking to  de-clutter will only be possible when you see Brand X has posted again this week, and you have tried Brand X but you now you realise you don’t like it. So you click Unlike. But what about that web page, or website article that you liked. That won’t be posting content to your Newsfeed, so although you may have had very controversial views some time ago, and Liked such content, you will hardly think of going back to unlike this content. This liked content is building a history of you on the web, but it doesn’t allow for changes in taste and opinions.

One of the best articles I have read on Social Search is this one by Econsultancy. It looks at Google’s +1 and some of the next few points are from that article.

3. Apathy

Only motivated people will like or share content, what about people who aren’t motivated? Their opinions are simply ignored.

4. Value of Opinions

I have friends who Like just about everything. This reduces the value of that persons opinion. In social search these people who Like everything will feature much more prominently. Will searchers looking for results that exclude certain people have to look further down the page or even have to look at page 2?

5. Reluctance

Social signals only show the opinions of people willing to be associated with that content. In the Econsultancy article the author mentions a hair restoration treatment, would people overcome embarrassment to ‘like’ that enough to be associated  with it?

6. Lack of Knowledge

Social media only reflects the opinions of those who are competent enough to use it, which may skew results that favour younger, tech savvy and time rich individuals. Although, I feel, this will change overtime.

7. Choice Paralysis

Will you tweet it, like it, send it, share it, digg it, stumble it, recommend it, +1 it, share this? Or none of the above? Too many choices to make. Google pulls in data from Twitter, Bing from Facebook. But what about all the other social signals? Am I likely to miss out on an amazing article because someone recommended it on LinkedIn instead of Liking or Tweeting it?

8. Conformity

Bing’s argument for Social Search is based on The Friend Effect, which isn’t too different to the Bandwagon Effect. In the Friend Effect, from the Bing Blog, “90% of people seek advice from family and friends as part of the decision making process”. The bandwagon Effect “arises when people’s preference for a commodity increases as the number of people buying it increases”. Change commodity for search result, Facebook Page or whatever. They’re not exactly alike, but the more popular content amongst my friends, my connections and then the wider Facebook community will rank higher. That can’t be a good thing. Will this conformity make people Like content of a lesser quality because others have?

9. I Can’t ‘Like’ Bad Stuff

By this I mean bad reviews, or content that goes against the general flow of thinking. I like to read reviews especially when I’m going shopping for electrical items. It’s not the good reviews I look for, it’s the bad reviews. The bad reviews, even for items that have a good rating overall, are often better at highlighting the deficiencies of products. They might be extreme examples, or be one off cases, but they make you better informed about your purchase. Take for example Apple fanatics, who will probably Like a totally positive review for a new Apply product. This gets bumped up in search rankings by so many fanatics Liking it and any review that is more accurate, painting the highs and lows, gets pushed down the rankings.

10. Filter Bubble

The filter bubble only gives us what Facebook and Google think we want to see, things that fit neatly in with our world view. The type of stuff that refuses to change our opinion. Think of it like this, your only exposed to Fox News in the USA as your one source of news. Fox is sensationalist, right wing and often mis-informed on issues or certainly decides what to ignore in order to fit its own agenda. How would that affect your opinion? It’s these filter bubbles that spurred Mark Little to set up Storyful, to find content we may not otherwise get to see. Perhaps its one silver lining that a new industry in identifying this type of content can work successfully. This next video by Eli Pariser better explains the concept of Filter Bubbles and why they could be bad for democracy.

And Finally
Social search could prove to be a better search product, but its over reliance on people may work against it. Where results are not more accurate but more diluted because its taking opinions, rather than facts to determine what the results are. How social signals will inform search is going to be a hot topic for a while yet. Not least because Google’s inability to get behind the walled garden of Facebook, which is Bings major competitive advantage.