The important role social media content plays in search engine marketing shifted again most recently when Bloomberg published a news article about a deal between Google and Twitter in February 2015. This deal would allow 140 character tweets to be indexed and searchable on Google instantly. Twitter previously had a deal with Google where tweets could be within its organic search results for trending topics, allowing real time content to appear for certain search queries – so called Google Real-time Search, which was retired in July 2011.
So, what’s in it for Twitter?
The growth rate of Twitter users has slowed down alarmingly over the last few years (the decline from Q1 2012 is evident in the below chart) which means Twitter has an uphill task to keep its investors happy.
This deal seems to be a part of series of initiatives taken by Twitter under Dick Costolo in order to get Twitter exposed to more eyeballs. Twitter is also trying to convince developers to use MoPub (ad serving solution) within games and applications in a bid to generate more revenue through mobile
In a similar move, Twitter has also opened up access to non logged-in users so they can access content on different content topics – from tech blogs and to pop artists and cute animals – all in order to get more exposure in search. This presents rich real-time content, just like the Twitter experience for users who log in. No only allowing influencers to reach a wider audience, but also enable Google to index conversations around rising trends easily.
Current Homepage
New Homepage
Why start with mobile?
The tweets are currently available in Google app (IOS and Android) and web search results on mobile web. The decision to go mobile first seems to be in line with Twitter’s commercial objectives when you consider that the platform receives 88 per cent of its ad revenue from mobile.
Why does Google care about Tweets?
The value this social content has to Google is in helping the search engine giant to better understand content on the web, particular entities (people, place, organisation or an event). With more than 236 million active users, Twitter can be a rich source of real time information about an entity. The end goal is to offer users with a rich experience, particularly for those topics where both users of Google and Twitter are searching online for same information (especially surrounding events or TV programmes) but on separate platforms.
Take an example of Eurovision, there were 500,000+ searches on Google UK about Eurovision in the month of May 2015. Similarly on Twitter there were over 6 million tweets on the topic, of which 70 per cent of conversations were from users in the UK.
Google trends
Conversations on Twitter about Eurovision
Access to Twitter Firehose
Twitter content has remained part of Google search results since the last deal ended in 2011, however, when this deal did end, Google had to crawl Twitter to index relevant tweets which is an ineffective way to discover relevant content quickly. The challenge is evident in the volume of user generated content on Twitter: there are around 6,000 tweets per second on Twitter, 500 million tweets per day and around 200 billion tweets per year – all which Google has to crawl, process and index.
Under this the new deal Google will get access to Twitter’s Firehose, its aptly named data stream, that will give Google unlimited access to its content. Unlike Twitter’s streaming API, which only gives you a sample of tweets against a particular query, Firehose guarantees 100 per cent delivery of tweets on a particular topic in real time giving Google a wide range of options when it comes to displaying the most relevant tweet in search results.
With Firehose, Google will be able to see over 150 attributes that provide information about the user that posted a tweet. The entire JSON associated with the tweet can be found in this example, but some typical attributes could include:
- Entity: Person, place
- Link to the original tweet
- Friend count
- Posted time
Entity extraction from tweets
As mentioned earlier that this social content is value in helping Google to understand entities. This is an important step in the further development of semantic and conversational search, which reflects the change in the way people search online. Whereas Google has previously been reliant on the likes of Freebase (an open, Creative Commons licensed repository of structured data) and Wikipedia to understand/disambiguate concepts and entities, this deal with Twitter acts as another valuable source for Google to validate information on entities. Twitter content on a trending topic maybe highly relevant and timely, but a popular subject will inevitably carry a lot of noise. The challenge here is how Google is able to extract meaningful entities from a huge volume of 140 character tweets.
However, search engines have developed sophisticated entity extraction mechanisms which allow them to draw the most relevant information from these data streams. To give an example of this process in action, the event calendar below shows the most prominent events mentioned on Twitter in association with each day. This calendar has been created by annotating the entities and resolving temporal expressions (e.g. “next Friday”) along with mentions of events that will occur in the near future.
Possible expansion of Twitter results from entities to Q&A source
Essential to Google improving its search results is understanding the user intent behind a search query and how fast they return the most relevant information to the user. Google’s aim to present the most useful and relevant information to users in the fastest time possible.
With that in mind, of its no surprise that question and answer based results started appearing in Google SERPs (see example below).
Twitter can be a great source of questions based content
It’s already a well establish behaviour, people using social media platforms for researching queries and asking their peers for information. So how can Google best access this content? In fact, some research has already been done to identify and detect questions with in tweets. A notable example of such work is by researchers Efron and Winget, who have built a taxonomy of questions on Twitter and found that people asked questions on Twitter to both seek information (facts and opinions) and recommend information by linking out to external sources.
Although it’s too early to say if Google will go down a similar route but identifying these qweets (tweets with questions in them) and displaying the best answer within search results is another potential example of how Google could use Twitter and access to its Firehose effectively.
Value of a link in a tweet
From a search engine marketing perspective, Google indexing Twitter’s content presents an interesting opportunities for business and brands. What would be the SEO value of a tweet? StoneTemple ran such an experiment, to see if tweeting a link can influence Google to crawl and index the page linked from the tweet. The crux of the study involved monitoring the online behaviour of volunteers but a key finding here was that of the six people who tweeted a page, two managed to get their tweets indexed. More can be read about the study here but this does imply there’s an assigned value to tweets, although more analysis would be necessary to get a clearer picture on the value of this content.
Implications for organic search
There are several implications for organic search, which have been referenced throughout this post.
The main points to consider are:
- Deeper integration between SEO and social: With tweets and a possibility of seeing more Qweets in Google search results – SEO and social content strategy should be much more aligned.
- Content curation, particularly Q&A content: Pulling question based tweets (Qweets) and creating event calendars are the examples of how Twitter can be used for content curation.
- Tweeted content isn’t guaranteed to be indexed: A mere inclusion of links in Twitter will not guarantee inclusion of a page in Google’s index. There are more than 150 attributes that will determine whether a tweet can be indexed or not.
- Entities are increasingly important to Google: It’s even more important to send a clear signal to Google about who you are as a brand through owned properties and sources such as Wiki Data.
Summary
The deal is beneficial for both Google and Twitter and it will be interesting to see if Twitter’s average user growth improves once this deal is implemented outside of the US market. But the real winner here is Google: it’s all about getting access to the real time information that they can use as a lever to bring people back to search.
