Andy Taylor

Andy Taylor

How Google Should Go Social

A rumour came out of SxSW that Google is working on a new social network called Circles. I have no idea what this is, but I know what I think it should and shouldn’t be.

It shouldn’t be another Twitter rip-off. And it certainly shouldn’t be a Facebook wannabe with the ability to put people in groups. Facebook already does lists.

What Google should do is add recommendations based on social influence and friendship circles to what they do best. Search.

Think about it, on Twitter you follow your friends, you follow leaders of your industry, you follow people you never met, but you know based on the little snippets of their lives, that you trust their opinions.

When someone posts a link on Twitter, you don’t just blindly click the link, you first see who posted it, you’re judging that link based on the person who’s posting it.

If your friend, who likes the same type of books as you, posts “Two chapters in and I’m loving this new book”. You’re far, far more likely to click that, than the 240th Google result for ‘great book’. But wouldn’t it be great if you could search for ‘great book’ and actually get books that you’d think were great…

If you follow Jeffrey Zeldman, and he tweets “Perfect example of semantic markup” your probably far more likely to click that link than if @webdezina3 posted it. But you’re probably also much more inclined to read A List Apart. Maybe your Google search for ‘semantic markup’ should return a relevant A List Apart article on the first page.

And maybe those results should indicate that they’ve been bumped up because of people in your social circle.

That’s all great in theory, but how would it work?

We certainly don’t need another button to put on our articles and websites. I’m already sick of seeing Like and Tweet buttons all over the web. And not too many people added Buzz buttons to their websites, did they?

Although some people use them, browser bars are pretty horrible. But discrete browser extensions have proved pretty successful in the past. And with clear instruction, many people can work out how to add a bookmarklet.

There’s also Googles own browser, Chrome. What better way to get uptake than to add two buttons, one heart button and one ‘other’ button next to the URL bar.

Hit the heart button

It would need to be super easy. Find a website you like, hit the heart button on the homepage. Find an article you like, hit the heart button on the permalink page. Find a product you’d recommend, hit the heart button on the product page. Find a YouTube video you enjoyed, hit the heart button.

I read plenty of of articles and find plenty of useful websites. But I only share a fraction of them on Twitter. Why? Because I’d be flooding my timeline with stuff lots of people aren’t interested in. But I would click one button if I liked something, if I knew that next time a friend was Googling that something, my recommendation ‘might’ rise to the top.

That ‘might’ is a big thing. Search results still need to be relevant. It’s not about replacing search results with social results (Twitter search already does them). It’s about enriching search results with social insight.

The other button

I’d love this idea to just be one button. But the ‘other’ button is necessary. The other button is the private heart. It’s a way of saying “this was useful, but don’t tell…”.

Imagine this service reaches the point of mainstream acceptance that Twitter has. Which is entirely viable if Google did it.

Your doctor’s on it. They might heart respected medical journals. But your friend Googling his rash probably doesn’t want you to know about it. They both find valuable information. One of them hearts it from an authoritative perspective, the other hearts it anonymously because it told them what cream to buy. You see that Dr Schwartz thinks this is a valuable and reputable resource, and that ‘1 other hearted this’. The latter clearly isn’t quite as valuable, but it still is.

Google could also extend this to a stand alone web service. A website you could visit, that had a live feed of all the things your friends were ‘hearting’. Maybe there’s a way to comment on the things you heart, but it can’t be required for the service to work. Fundamentally it needs to be as simple as clicking one button, and be centred around search.


The side note here is automation. Google’s pretty creepy. They track everything you search for. Presumably to provide better search results and serve more relevant ads. But this means they could develop algorithms for social recommendation, based on what you search for, without you needing to hit the heart button. It’s potentially powerful but also game-able. The main reason for social recommendations is those dodgy SEO people. It makes it much harder for them. Google have done incredibly well despite them. And if you maintain a good list of ‘friends’ that doesn’t include them, automating recommendations based on you visiting a site say 10 times, might be viable.

People ‘heart’ being asked their opinion.

People like recommending things. And if helping your friends and other people in your wider social circle find great content was that easy, why wouldn’t you. Why wouldn’t Google?

Google certainly don’t need another Buzz privacy fiasco. But it should somehow start with your Google contacts. But, there needs to be a way of following people who aren’t really friends, just like on Twitter (starting with importing Facebook and Twitter contacts). So you can find those few highly regarded people in your industry and follow them. You can find those people who are fanatical about the car your drive or the type of bike you ride and follow them. Not because you want to have a constant feed of information from them every day, but because when you’re looking for information on WAI-ARIA, or parts for your Golf, or a seat post for your bike. Their recommendations might be there.

If that’s what Google Circles is, I’d heart it.

Thanks to Jane, Googe and Jess for their feedback on this.

Have a comment? @me on Twitter.