Skip to content

Posts tagged ‘algorithm’

Getting the Most from #Summify: How It Works

This is the first part of an ongoing series that will help you better understand how Summify works and how you can get the most out of your summaries. First things first. In order to improve something you need to understand it.

You all know that Summify automatically identifies the most important news stories for you across all of your networks, but how does all that actually happen? Does it have something to do with extra-terrestrial beings from Area 51? Well that’s confidential, but here’s a few things we can say.

How do the algorithms work?

Plug in your social accounts > links are collected > links are filtered > summary arrives!

Accounts

You can plug-in three types of social accounts to Summify: Twitter, Facebook and Google Reader. How many accounts can you plug-in from each service? There is no limit for Twitter, Facebook and Google Reader, add as many as you have!

Collecting links

Once you’ve chosen which accounts to connect, Summify begins analyzing them, collecting all the links found in your news streams. Currently only links to articles (text-based links) are collected. Summify does not consider any other link types: e.g. youtube links, audio links, download links etc. But hey, things change pretty fast around here, so you never know what’s next.

Filtering links: cut the S***

The Summify algorithms use dozens of signals, but in general news that has been shared, liked and retweeted a lot by your friends is considered more important. However, not all of your friends are equal and we pay special attention to those who you interact with more. Global popularity signals such as the total number of times a link is tweeted, or liked and shared on Facebook are also considered, but are less important than the sharing activity of the people closest to you.

Filtering links: learning

The Summify algorithms are also fans of learning. As you continue using Summify, we analyze your click history and understand your interests to further improve your story recommendations.

We understand your interests indirectly by looking at what you click on. We notice if you tend to click more on stories:

  • from a particular user on Twitter, Facebook, or a certain RSS Feed;
  • from a particular domain or with common keywords in the story title; and
  • if you prefer Twitter stories more than Facebook stories.

We hope that this break down has helped you understand how stories make the cut into your summary. On average, for each story that makes it into your summary we reject 150 other less relevant stories, so with Summify you’re reading only the good stuff. Not only is this a huge time saver, it also makes your Twitter and Facebook experience more useful.

To gain a greater perspective on the volume of news being shared and the amount of stories being filtered from your summaries, check out our recent Social Sharing Infographic.

Finally, if you haven’t seen it already, take a peek at our “What is Summify?” video to get a helpful visual representation of some of the things we just covered. Next up: optimizing your summaries.

Follow

Get every new post delivered to your Inbox.

Join 89 other followers