Google’s 300 Million User Olympic Experiment

Google’s Enormous Olympic Experiment

Google have launched lots and lots of Olympic content within their main search results. They have also created a neat user interface to join all of this together and, in doing so, have essentially created a giant Olympics ‘website’ with tens of thousands of pages, all integrated directly among search results themselves.

Other search engines have included medal tables, and basic information around The Games. This goes a little further, removing much of the reason for searchers to visit the official London 2012 site and broadcaster sites. It also points to the future of Google’s “Knowledge Graph” project, and how this may affect other website owners.

This post covers the Olympic ‘website’, pulls out some specific examples from the UI (with notes), and then talks about the implications for the future. It’s split into 4 sections:

  1. The Background.
  2. Google’s Olympic “Website”.
  3. An explanation of the scale of this experiment.
  4. Some thoughts about future implications.


1. The Background

When Google first launched AdWords it was easy to spot the tweak itself (“they’re showing ads!”), but it was also very easy to miss the implications. Nobody would have said “that’s going to make Google the most important company in the world”. Similarly, when Google started talking about “The Knowledge Graph”, it was met with a similar shrug (“they’re showing facts??”). But the way they’ve used this to present Olympics data may have implications for every website owner.

Here’s a search for ‘London 2012’ at the moment:

The right-hand side over there is the ‘homepage’ in their Olympic website.

And here were 4 quick thoughts about that from Google+:

More interesting than that though: All of those links over on the right are clickable, taking you to subpages within a gigantic, constantly updated Olympic ‘website’ that overlays itself around search results.

2. Google’s Olympic ‘Website’?

Google introduced “The Knowledge Graph” earlier this year. The explanation was that they were now going to show facts & figures in the right-hand column of search results. For example like the following:

Where that differs to what they’ve done around the Olympics is that their Olympic results have A) a set of user interface elements specifically for the event and B) a fully self-contained navigation structure C) appear globally (whereas ‘knowledge graph’ content is largely reserved for the States).

In other words, Google have created a complete ‘Olympic Website’ within the right-hand panel of search results. Rather than linking off to other sites, 50% of the page on any search closely related to the Olympics is now Google’s own content (or, rather, other content presented as Google’s).

To contextualise that ‘50% of the page’ bit further: The Olympics is the biggest event in the world, taking the biggest chunk of ‘event advertising’ budget from the world’s biggest brands, and Google have chosen to ditch their right-hand ads area (and in some cases the ‘top’ ads area) in favour of showing ‘their own’ content.

This happens on core olympic phrases (‘London 2012’, ‘Olympics’, ‘Medal Table’, etc), but also on much wider terms. Here’s an example of a ‘long-tail’ Olympic phrase, followed by a comparison of the ‘Google’ areas vs other areas:

Here’s the comparison of the ‘Google data’ there vs actual search data leading to other websites:

(Why visit the official site (in pink), when you can stay inside Google’s hedged garden? (in green)).

Here’s a breakdown of the different elements contained within this Olympic content:

Google’s Olympic ‘Site’ contents

This Google embedded olympic ‘website’ includes the following:

  1. Every participating country. (205 of them I think, including the independents)
  2. Every sport (eg. ‘swimming’; 36 of them in total according to the official site; 26 according to Wikipedia (!))
  3. Every ‘event’ (eg. ‘men’s 50m freestyle’. There are 34 of these within ‘swimming’ alone)
  4. Every medal. (personalised to pull out local medals based on your location)
  5. Every competitor position throughout both heats & finals, in every event. (the UK alone has 541 athletes, including team sports)
  6. Details of world records. (from wikipedia)
  7. Links to the appropriate place on the official website. (no doubt negotiated to allow them to do all of this)
  8. Links to watch online (personalised depending on the country you view from)

Here are some examples of how this appears.

Examples: Core Olympic Terms

Here’s what shows when you search ‘Olympics’:

The Medal Table

And here’s what shows when you search ‘Medal Table’:

You’ll note that appears in the central search pane, not on the right-hand side.

Country Terms

There’s a country page for every single country. You can reach these by searching for “London 2012 [Country Name]”, or other related phrases like “[Country Name] Olympic Team”, or in some cases things like “TeamGB”. You can also reach these by clicking any of the links within the Google ‘site’ itself.

Here’s the example for the USA team:

Clicking any of the links there takes you to another page within the ‘site’. Eg. the ‘Swimming’ link takes you to a hub page…


Every sport has its own ‘hub’ page, with links to all of the individual events within the sport. Here’s the ‘Gymnastics’ hub page:

So, for example, a Google search for ‘swimming’ no longer shows the ads it used to, it instead shows content similar to the above.

In most cases, the right-hand results appear fully expanded. In a few cases they appear like this:

And below these Sports pages are individual event pages, for example:


3. Just How Big is Google’s Experiment?

So Google have essentially created an entire, real-time Olympics website overlaid on top of search results. All of the content is readily available on the official site, and on a number of broadcaster sites. That’s quite a big experiment.

The Olympics is by far the biggest global event of the year; and by doing this they are foregoing lots of ad revenue. By doing this they are also training lots of users how to use the right-hand sidebar in this context. It’s worth taking a look at how many users might experience it. Here are some back of the envelope calculations:

There are 3 main entry points to all of this:

  1. Click on any Google Doodle  over the period of the Olympics. They all lead to a different ‘sport’ page.
  2. Search for any of the sports-related terms.
  3. Search for anything Olympics related.


1. How Much Doodle Traffic Sees This?

At the moment, every Google Doodle points to one of these Olympic pages. Here’s the ‘Fencing’ example:


To estimate how much traffic a Doodle generates, let’s look at an example: Amelia Erhart’s birthday was a week or so ago and she was featured on a Google Doodle. This is the graph of her Wikipedia traffic over the last month:

That’s 1.776621 million views in a single day. Her Wikipedia page is the 2nd result for her name so, if we take an optimistic view of how many people would click it (25%), that means her doodle generated roughly 7 million searches for her name in a single day. Across the 2 weeks of the Olympics, we can therefore roughly estimate the doodles will cause close to 100 million searches.

2. How much ‘Sports’ Traffic Sees Google’s Olympics Content?

The next traffic entry point is ‘sports’ related searches, many of which show these Olympic pages as results.

Monthly ‘phrase’  match for “Basketball” gets 20.4m monthly searches. So if we’re pessimistic there and estimate only 25% of those show the Google olympic page, that’s still 2.5 million searches over the 2 weeks for that sport alone (one of 36). Let’s assume that basketball is representative, that gives us 90 million searches across the 36 sports.

3. How much pure ‘Olympic’ Traffic Sees the Content?

Finally, we have people searching specifically for Olympics related traffic.

Here’s the brilliant Alex Balfour (do follow him on Twitter if you are not already), head of new media for London 2012, explaining just how much interest there is in the Olympics:

So, if we pessimistically assume another 100 million ‘Olympics’ specific related searches over the 2 weeks, we’re close to 300 million searches displaying Google’s olympic-specific results. And that excludes the Paralympics.

Experiment Size Summary

Summary: We can assume Google are training users across roughly 300 million searches how to use these types of results. We can also assume they’re foregoing much of the ad revenue across these 300 million searches.

Even if those numbers were out by a factor of 10, that’s still enormous. And as we’ve been pessimistic in all our numbers, it could just as easily be ‘300 million users’ rather than ‘300 million searches’.

4. Future Ramifications

It may be that this is all just a bit of fun. But rolling out these big new UI changes for The Olympics is quite a big bet. It may well be that there is strategy behind this.

Here’s an example of how 4 recent strands of Google activity could tie together nicely alongside this. The 4 strands are:

  1. The Knowledge Graph.
  2. The Walled Garden.
  3. Attribute-Based Searching & price comparison.
  4. Paid Inclusion.

If we add that all together, this type of UI may be a useful model for Google to replicate entire websites worth of content alongside search results, and to expand its ad revenue further in future.

1. The Knowledge Graph

Google officially announced The Knowledge Graph, and have been pushing this in the states. At the moment this is purely informational, and usually appears on terms with no ‘buying intent’ (ie. terms where you wouldn’t normally see ads).

The Olympics is the first time the UI elements from this have been used alongside search results in anything other than a ‘Wikipedia-lite’ type way. Google show here that – rather than simply individual collections of data points – they’re happy to recreate full, cohesive ‘websites’ of content if it suits their needs.

2. The Walled Garden

Facebook is very much a walled garden, trying to create its own fully self-contained web within the web. Facebook tries very hard to suck visitors in as often as it can, but does all that it can and incentivises advertisers to avoid giving them options to leave the garden. Twitter appears to be going down the same route, closing off APIs, bringing display of photos and videos within their own interface.

Google have also made several moves toward this over the last few years. To mention just 3:

  • Pushing Microformats and author information, which allow Google to understand content and give them the context needed to redisplay it themselves in any format they choose.
  • The almighty ‘Not Provided’ debacle, whereby Google now hold back masses and masses of important data from website owners, all of which used to be seen as ‘quid pro quo’ for allowing Google to use our content.
  • Google+ itself, which encourages both brands and users to create more content on Google’s own properties, rather than to try and suck visitors away from Google.

Moving visitors away from Google and on to their intended destination used to be Google’s core purpose. Now they would prefer that users’ ‘intended destination’ is Google itself.

This fits in very nicely with that. All of the information in Google’s Olympic ‘site’ is already out there on other websites, but they would now prefer you to find the information without leaving Google itself. These UI elements again give them a great way of doing that with all types of content, and to tie that together into a cohesive, navigable architecture.

3. Attribute Based Product Search

The entirety of the Knowledge Graph, and Google’s Olympics ‘site’ are based around a UI and data that lend themselves perfectly to attribute-based product search.

Google recently bought 2 different ‘attribute-based’ price comparison engines: ‘Beat That Quote’, a UK-based price comparison site and ‘SparkBuy’, a Seattle-based laptop comparison site.

Also along these lines, Google’s ‘Credit Card Comparison’ UI and entry point have been tweaked recently in the UK:

Google have also launched a few other attribute-based search tools themselves, with fairly neat UIs, including:

Just as the Olympic site lets you browse and narrow down between ‘Countries’, ‘Sports’ and ‘Medals’, it is not much of a stretch to see that this could just as easily be ‘Brands’, ‘Categories’ and ‘Products’. If you take a look at the ‘flights’ & ‘cars’ comparison sites above, you’ll see elements from each could fit nicely into the Olympic model.

4. Paid Inclusion

One final piece in this is the change of Google Shopping from a ‘free’ service to a paid model. The official story here was that this was to improve the quality of data by discouraging advertisers who submitted anything and everything with very little accuracy.

But fitting this with the type of UI used in the Olympic ‘site’, it does 2 things from Google’s point of view:

  1. Allows them to gather far more ‘attribute-based’, relational data from product advertisers, and incentivises them to ensure its accuracy.
  2. Introduces the ‘paid inclusion’ model, which means advertisers no longer bid based on keywords specifically, but based purely on their willingness to pay if a product is clicked.

Once this is fully working, it allows Google to choose how and where products are displayed, whereas previously those have been the choice of the advertiser. Again, that fits in perfectly with the UI elements here to allow Google to show ads that users have ‘navigated to’ rather than ‘searched for’.

As just one example of how this may benefit them: Whereas a search for ‘Amazon’ reaps Google just a few pennies if the user clicks on the Amazon brand PPC ad. If they were able to bring all of Amazon’s products into their own UI, and train users to browse through those on Google itself, they would reap far more than a few pennies each time the user jumped off to a product page.

Adding this all Together

Adding all of this together, and the new UI areas Google is testing would work equally well for both pure content, and for paid inclusion ads.
  • From a pure content point of view, the new UI elements work very nicely for recreating complete ‘content’ websites alongside search results.
  • From a product point of view, just as the Olympic site lets you browse and narrow down between ‘USA’ to ‘Tennis’, to ‘Medal Tables’, this could just as easily be ‘Samsung’, ‘Phones’ and ‘Galaxy S3’. It would not be such a stretch to extend this to recreate the full ‘product discovery’ phase of ecommerce sites.

And of course, by launching the UI elements for this tagged to such a huge event, they are able to gather data extremely quickly on its use, and to educate a huge mass of users in how to use these elements.

What do you think?

What do you think of Google’s Olympic content? Is this just a nice user interface piece Google have put together to help us all? Does it point at other changes? And what  would you do if Google did this to your niche?

ps. big thanks to the many people speaking about this on Twitter, including @danbellj, @StokedSEO @wilreynolds, @dannysullivan.

Increase the Chance Readers Will Share Your Content

How to Improve the Chance Readers Will Click Your ‘Like’, ‘Tweet’ & ‘Share’ Buttons

This blog post offers 5 simple steps to improve the likelihood that readers will tweet, share, like, or ‘plussify’ (is there a word for google+?) your content.

As ever, this will not work if your content has no natural value (or no perceived value) in the first place. This article assumes that you are putting some genuine effort into content.

There is an image accompanying each step: Look closely to spot the simple tweaks.

The Example

Below is an image taken from the foot of an article from The Guardian.

Looking just at the area at the bottom of the image, you’ll recognise a really common way of displaying ‘share links’. This is basically the ‘checkbox’ implementation that you would apply if someone asked you “could you add some share links to our articles?”

Below, we’ll make a few tweaks to this to increase the likelihood of shares.


Step 1: Remove the Ringfence

In the image above, you’ll notice that the ‘share’ buttons have been grouped together & ringfenced into a little block in the page. This looks neat aesthetically, and the reason that it looks neat aesthetically is that they are essentially rinfenced into a  block that is easy for your brain to understand “that’s the share button area”.

By ringfencing it like that, and easily identifying it as the ‘share button area’, that automatically tells people who habitually use these links “your sharing stuff is here. click us!” But, of course, it has the negative side effect of saying to everybody else (who does not habitually click sharing buttons) “feel free to ignore us – this entire block is just the sharing stuff you always ignore!”.

Here’s the same area with that ringfencing removed, and the share buttons placed directly into the ‘content’ area:

Not such a big difference, but the equivalent of placing them as ‘the next logical thing to look at’, rather than a dead end stop.


Step 2: Add ‘Social Proof’

‘Social Proof’ is a bit of a cliche, but – speaking incredibly generally – if you show someone that others have taken an action before, they are more likely to take it themselves.

There are many ways to do that, the most simple being to automatically include the number of tweets/likes/shares.


Step 3: Remove the ‘Choice Paralysis’ or ‘Choice Blindness’.

Again speaking very generally, if you offer someone a choice of 5 things, it is far harder for them to decide than  if you offer just 1 or 2.

If it is more important to your business to get ‘tweets’ than it is to get ‘shares’, prompt for the ones you want.

Above we’ve included just 1. It’s often better to make big changes first & then taper down, than to go the other way round.

Step 4: Add a Prompt

Many people are not in the ‘habit’ of sharing content directly from web pages. It is not always appropriate, but a simple, friendly prompt can sometimes help to temporarily break that ‘not sharing’ habit on your page:


Step 5: Give it a Go.

Test this and see what actually works for you.

Everything here is a rule of thumb that has worked on some sites and for some audiences, but may not work on others. If you like, it would be very easy to A/B test each of them (for example, it may work better for some audiences if you dramatically highlight your ‘social sharing’ block, rather than simply place it in the content area).

If A/B testing is too much effort, simply take the less scientific approach & try things out for a limited period of time to see if you can judge any change.


That’s it

Please do leave feedback in the comments below. And of course, please do share this article if you think others would find it useful!