How to Design the Perfect Lifestreaming Content Reader

Over the last 3 years I have spent the majority of my time covering the tools and services that provide ways for us to record and distribute our Lifestreams across the web. Over that timeframe we’ve seen the adoption rate of such services increase exponentially. The problem that began to surface pretty quickly is how can we find good ways to consume this mountain of data being generated by the people we follow. Two recurring themes that you’ve probably heard often over the last several years is how do we reduce noise and how can we filter the relevant data for us.

While I may visit a few Lifestreaming sites every day to track my friends content daily I also still find myself visiting many other social destinations to find the content my friends are creating and recommending across those services. This is an inefficient and time consuming process and often I may not discover items that have fallen off the timeline. I’ve been giving quite a bit of thought to this issue and have started to develop thoughts on a web service that may be able to help provide a solution to this problem.

Several services have come along aiming to tackle this problem on either a single service or media type basis. They also use many different methods to analyze and personalize data for you. What I want to see built is a way to analyze all the data being generated by my social graph (the people I follow) on multiple social media services (Facebook, Twitter, Flickr, YouTube, Last.fm, etc.) and prioritize the content and display it to me in a meaningful way while also allowing me to tweak the initial logic that automates this.

Ok so that concept may sound a bit confusing and I’ve encountered quite a few “deer in the headlights” looks when trying to explain it to the many developers and bloggers and other technologists that I’ve spoken to about it lately. So I am going to provide a breakdown with details on a way to build this type of service.

A service that I’ve grown to really love which takes a similar approach to what I’m proposing is Twitter Times. The premise behind their site is that you provide them with your Twitter account and they in turn build a real-time newspaper styled page customized to provide content based on the links within the Tweets & Retweets being sent out by the people you follow. It also sorts the content based on the number of followers linking to the content. I have days where I don’t visit twitter for long periods of time and I always have the fear of having missed something being discussed that’s important to me. With Twitter Times I don’t have that anxiety because I can visit the page at any time and it’s archived those important items my friends have identified.

So I love the base logic behind Twitter Times but it’s limited to one service and just a slice of my social graph. I would love to include the people I follow and the content they’re sharing from Facebook, Google Reader, Delicious, Digg and other sites to also be incorporated into this page. And then to take it a step further I also want different media types to be presented separately on the page so that I have videos, books, music, and photos also displayed on the page by my followers using related services.

So what are my thoughts on the logic that would power such a site? Well I’ll provide some examples that can explain methods to do this. Let’s start by displaying how Twitter Times provides a personalized page for you.

Ok, as I stated before I’d like to see more services (Google Reader, Delicious, Digg, etc.) incorporated adding a larger sample size based on my followers on those services. So I’ll demonstrate that by moving on to how this could be achieved by also applying this additional logic for videos. But also allow for not just links but also gestures made by those I follow on the source services. So as an example a tweet might link to a YouTube video, but also a friend may have favorited the video directly on YouTube, and yet another commented on it over at Digg.

Yet another method to aggregate data could be used in cases where you have to pull data from multiple sources that don’t link to a distinct url / content type. My example here could be used for music where we may have either scrobbled or favorited tracks on Last.fm or perhaps we used Blip.fm and sent a Tweet out to a song or perhaps we made a purchase on iTunes or Amazon (friends and data available via Blippy). All of these data points with multiple gestures to associated content on different services. Here another option would be to normalize the data at the band / artist level. Pretty challenging but yet another great way to find out what our friends are paying attention to with regards to music.

So these are just some examples that could be taken to many other content types. You could add books based on friends and gestures across GoodReads, Amazon Wish Lists, or Librarything. You could add restaurants based on data pulled from Yelp, Foursquare, and Foodspotting. Apply similar methods to gadgets, photos, tv shows, etc. I think you get the point.

So while this provides a high level view of ways to create automated logic to pull the relevant data specifically tailored by our social graph, no automation will be perfect. In fact depending on how you use each service, there’s a chance that the data may not be as relevant as you would like. For instance if you auto-follow everyone on Twitter then there’s a good chance that the data from there may offer some chaotic patterns for you. These are some issues to keep in mind.

So how do you correct a situation like that or even provide some better ways to allow someone to tweak the data better. Why by adding some knobs, buttons and levers of course. Providing a way to add or subtract weight from individual users or even block them could be very useful. It would be nice to offer that at the content type level though and not globally. Adding the ability to create categories based on the meta data from the content and then offering ways to tweak that would also be great. There are plenty of other ways that could offer methods to filter and prioritize the data.

So anyways these are just some of the rough ideas I have that have bouncing around in my head. I, and I assume many of you have enjoyed the ease by which we can now share and aggregate data but this has gotten very messy. If anything things will continue to get worse with the proliferation of mobile and the additional ways we can now share from anywhere we go. So we need to now put the focus back on finding ways to make sense and add relevance to the information. I’m really hoping that someone out there is thinking somewhat along the same lines and is starting to build the perfect Lifestream reader so that we can all enjoy the fruits of our over-sharing.

So if you or someone you know is currently working or thinking about building such a service, I’m ready and willing to provide any help and support I can. I’ll be waiting and hope you come a knocking.

About The Author

37 thoughts on “How to Design the Perfect Lifestreaming Content Reader”

  1. It sounds like we have a split on the word lifestream, which could be better defined. If I assume my lifestream is a collection of activity that I create around the Web, you seem to be focusing as much on the aggregation and prioritization of social content shared by others. Is that a lifestream, or an aggregation service?

    I do like your references to the Twitter Times and that UI, especially as you highlight the aggregation of social gestures on the service. It's similar to what we were always intrigued by with RSSmeme, ReadBurner, LinkRiver, etc., as we leveraged our social connections to drive top content by total sharing counts – but none of those sites emerged as winners, despite my preference.

    I also have concern around the entire field of social aggregation services, even as you know I am a huge user of them (See: FriendFeed and Buzz). I am finding a lot of pushback from people in terms of distributing updates horizontally across services, with people being rewarded for native posts originating on those services. FriendFeed, for example, won with us, but failed to achieve a mass market. Buzz looks also to be starting slow. Arktan and Cliqset and others are interesting technology examples, but even smaller than both of the previous.

    It could be you and I are focused on making something amazing that a small group wants. Your proposal would be great, and I would use it, but after we get the same 500 people to connect with us there… then what?

  2. This sounds like a fantastic potential service, Mark. When something like this gets built, I'll be clamoring for a beta invite right off the bat! I really like the idea of being able to weight different friends, too – even more powerful would be the ability to weight them depending on the type of activity (which I think you allude to as well). I may like one friend's taste in movies and video, so I definitely want to see their YouTube and Vimeo favorites and weight them more heavily than other people on that, but I don't care for the same music as they do, so I'd want to block their last.fm and Pandora activity.

  3. Smart take. I agree that no one has decisively figured it out so far, and Google Buzz has been a huge disappointment up to this point. Really not sure what Google was testing internally there…but it wasn't the kind of social media usage you and the rest of us are used to and are talking about here.

    Wrote a post recently about some granular filtering strategies that could be used to pop items to the top of your inbound stream, see here:

    http://3on.us/smart-filtering

  4. So I think you've misunderstood some of my points about this service. Don't worry it's probably me and others have had issues understanding as well.

    First off I agree on the term Lifestream having canned connotations by many people but I'm married to the term on this site. I actually have grown to really like the term Activity Stream but since that's an open standard initiative I don't want to confuse people using that.

    I think you may be missing the point of this service as it doesn't require anyone to join it to offer the usefulness I describe. I simply connect all of my external social media accounts and it in turn feeds me content based on the actions of those I follow on those external services. For example most of the folks I have feeding me content via Twitter Times have probably never even heard of the service. They simply have to be users that I follow tweeting links.

    I agree that the problem with this type of service is that it could be very difficult to build and for it to be effective needs a user base that has friends that are active on a pretty large variety of services. That target audience may not be a large enough fish to fry. Also, creating a simple compelling UI would be challenging as well. I wish I was rich so I could fund this pet project of mine 🙂

  5. Pingback: Den »Lebensstrom« ordnen (2) – Der Schockwellenreiter

  6. Very interesting post Mark… a topic i think a lot about. Though it's not a particular passion of mine, so i probably wouldn't be inclined to work on it, i have often contemplated this concept as it pertains to amplify. 3 things have consistently stumped me:
    1. i don't innately feel smart enough to figure out how to do it right.
    2. i don't think that most people would be willing and/or able to manage the “knobs, buttons and levers” that would be needed to set it up right
    3. people aren't always in the mood for the same type of experience…sometimes they want news, sometimes humor, sometimes just trivial banter. I can't see how a service such as this could effectively take that into account without again causing the user to work.

    I appreciate your thoughts on this and hope that someone can work it out. Hopefully if they do, amplify will be one of the services that people pipe into their activity stream.

  7. So to address your issues…

    with regards to #1 I think you may be right and it would be best to try and put as much automated logic in as possible for the system to do its best to generate relevant content. That could be done by adding weight to certain gestures such as favoriting over simply linking to content for instance.

    For #3 there wouldn't be a way to automate that unless perhaps if all you wanted was humor then you would only follow comedians on Twitter. A couple ways to solve that would be to either allow grouping of users into lists (either globally or per service) but that would be complex. Or generate categories based on meta data as I describe (think tag cloud) and then allow people to filter the content by those categories. But the bottom line is an automated system will only be as good as the content fed into it.

  8. Hey Alex, I read the post and agree with your thoughts. My thought here is just one way to try and create a simple solution that can be embraced and understood by the widest user base. This is just a high level first stab which could surely benefit from crowdsourced iterations. But I think it's key to make the default automation as compelling as possible for users to understand the benefits and gain adoption. Then by offering easy ways to filter, categorize, and tweak the results you can get maximum value.

    I also saw in your post the link to Louis' excellent slideshow on Finding Signal in Real-Time Noise which I'll link here for people to see as well which offers some great tips on using the tools and services to try and achieve some of these principles today.

  9. Pingback: Lesenswerte Artikel – 26. March 2010

  10. Wow, the ultimate Lifestream Reader – my dream! I've thought about this a lot myself.

    Having it simply pull in my streams from my contacts I already follow on each of my social networks is a good way to solve the problem Louis suggested. Especially to get started, but I'd like it to be more like subscribing to Lifestreams (taking advantage of the open standard activity streams). This way it's pulling in posts from services I'm not necessarily on (eg. say Louis wrote a great post on Buzz, but I'm not on Buzz) but may still be interested in seeing.

    I'd also like to see different views, one like the Twitter Times, one in more of chronological order, etc. Maybe something like how Feedly does it for RSS feeds.

    I love the idea of giving priority to active posts/topics (eg. several contacts posted about a product release or commented on it), the service should bring that to my attention, maybe through featuring it much like Twitter Times would, or how Friendfeed brings it to the top.

    I agree with your response to Goldstein below, there needs to be some awesome logic in place to figure out which posts get priority, automagically by default (for simplicity of new users). However, it should all be customizable in the settings; I don't want some magic algorithm like on Buzz that I have no clue how it works and no control over it. This may include such things as the ability to weigh priority of specific keywords, tags, users, and services.

  11. Pingback: links for 2010-03-26 « links and tweets

  12. Pingback: Strandh.DigitalPR » Blog Archive » DigitalPR omvärldsanalyserar [2010/03/28]

  13. We do some of this in FriendBinder. On the front page of our search feature we show the most popular links from your friends and who linked to it. This includes tweets, delicious, Diggs and Facebook shares from your friends.
    There are some screenshots at http://friendbinder.com/tour/ (3rd one down)

    You can also drill down to see what your friends said about the links and search for what your friends said about any topic recently e.g. “iPad” and of course that searches their tweets, Facebook, Diggs, delicious etc.

    I'm working on an improved version of this feature which will do more with filtering down your incoming stream to filter out some of the noise.

    You can find at http://friendbinder.com and I'd be very interested to get people's feedback on what we've done.

  14. Pingback: The Evolution from Numbers to Relevance by @ScepticGeek

  15. We are looking at this within business content domains.
    The key thing is to establish within a given topic what the level of authority someone has; and also to understand whether they are a leader or a follower in ideas. You can start to glean this for people with public profiles (many of us) and what they write, share and recommend.

    We'll bring the bulk of this out in PeerIndex (peerindex.net) in the next few months.

    There is an early Alpha of Digg-style business curator/aggregator called Viewsflow (viewsflow.com). Viewsflow is a testbed for us to see how well our undelrying rankings are working in the field.

    One issue is the manage the gamability of rankings. In the same way that SEO has murdered the quality of search results in certain areas (e.g. consumer products), any kind of person ranking needs to impose a cost for whoreishness, and we hope we've got a way around that. Wd be interested in your thoughts.

  16. Adding a bit more to your idea about allowing people to tweak things, I think that this is essential – what I like isn't the average of what people I follow like. Feedscrub has a pretty good mechanism for learning what you like and don't, as does my6sense (although it doesn't let you actively tweak, it just observes you). Adding something like these onto what you have above makes it a killer service, delivering to me what it is I want to see.

    The other thing I really need in the ultimate content reader is off-line capability in the mobile apps, which usually gets overlooked. I use Byline a lot, mainly because I can wade through stuff on my commute to and from work even when the signal is patchy (or I'm underground), and just star things I need to look at on-line later.

    Plus of course sharing and commenting within the reader which get redistributed back via Salmon/whatever.

    It's not been cracked completely by anyone yet so hopefully someone will take up your ideas and really build the killer content reader.

  17. Yes, the tweaking is key as many complain when they don't have control and are fed by a “black box” algorithm such as the current state of the Facebook news reader. I agree that a mobile app is essential as a tie in. I currently find myself using Google reader much more lately along with the same functionality you describe using with Byline.

  18. Well Mark… I have been following your blog for quite some time in the hopes that no one had created what I think is a idea to address some of these concerns and more. Many here are correct and all I have to say is that as with anything.. it's all in how you present it and learn from there.

    I have been working on a business plan for this however due to many other forces, I can't seem to find the time finish the foundation to begin to seek more help; since this is a massive undertaking as with many social media sites. You threw out the net and I say to you… I need help putting this invention complete with a marketing and branding strategy together. I have this idea however I don't have a clue of all that it takes to start a business of this nature. My other issue is I don't know who to trust so I walk cautiously. However I do have a passion to see this site come into being.

    What do ya think?

  19. My offering to help someone along the lines of building something such as this is limited mainly to that of a traditional product management role. I have ideas regarding functionality and UI but am not interested or have the time to pursue this as a business model. My limited time and interest is purely in helping anyone who is looking for ideas or feedback that is building something along the lines of what I'm describing. It appears that you are looking for quite a bit more.

  20. Hello Azeem. Yes the approach of personalizing the content based on users submitting content and gaining authority points to determine the algorithm is a very different approach. I've seen that done quite a bit and as you point out there are issues with gaming the system and others as well. Unfortunately that isn't an approach I'm interested in seeing. I want the filtering of content submissions based on my social graph and then putting the power in my hands to determine the level of authority individual users from my graph have by allowing me to add or subtract values that determine their weight. It's a far different approach that is tailor made specifically for me.

  21. Fair enough however I'm not sure what gave you the impression that what I'm talking about is so different from what you are describing. After all as I stated before I have followed you for some time and I choose this post as my first to comment on. It does seem as if you have a lot going on and I thought that given your stated passion for seeing the ultimate lifestream and your knowledge of the trend you would be willing to help.

    So I guess I will be contacting you in the future to offer you an invitation to try out the beta version as I respect your opinion, does that work for you?

  22. Well , the view of the passage is totally correct ,your nike air max details is really reasonable and you guy give us valuable visits-our-site informative post, I totally agree the standpoint of upstairs. I often surfing on this forum when I m free and I find there are so much good information we can learn in this forum!

  23. Well , the view of the passage is totally correct ,your nike air max details is really reasonable and you guy give us valuable visits-our-site informative post, I totally agree the standpoint of upstairs. I often surfing on this forum when I m free and I find there are so much good information we can learn in this forum!

  24. Well , the view of the passage is totally correct ,your nike air max details is really reasonable and you guy give us valuable visits-our-site informative post, I totally agree the standpoint of upstairs. I often surfing on this forum when I m free and I find there are so much good information we can learn in this forum!

  25. Pingback: The Evolution from Numbers to Relevance – Mahendra Palsule - MediaNama

  26. Looking on the face of it you wouldn’t think that hitting a golf ball into a hole would be that hard but it takes a great deal of skill to do this. To acquire these skills you first need to know the basic fundamentals.
    The basics are to hit the ball with a golf club into or as near to the hole as possible. Timberland Shoes As golf is a competitive sport each time you hit the ball or try to hit the ball is classed as a stroke. The golfer himself keeps a scorecard and records the score at the end of every hole. The game of golf is generally played around 18 holes and the winner is the person with the least amount of strokes at the end.
    When the players have teed off and had one shot each, the player whose ball is furthest from the hole is first to play the next shot and so on, this includes play on the green and the fairway. The player with the least amount of sots on the first hole then goes first on the second hole and co on.
    The golf tees are used when first teeing off .Generally most people have at least a couple of drivers, irons and a putter as well as balls, tees and a golf bag. On a par 4 or 5 drivers are generally used when teeing off for greater distance. The golf ball must weigh 1.62 ounces and be round in shape, Timberland Boots usually balls are marked with a name and number.
    Golfing etiquette allows players behind you to play through, as in the case where there are two players behind a group of four. The group of four should always be allowed to play through. This allows the smaller group not to be held up. Also you should always make sure that the party in front are not in danger of being hit.
    Make sure you have a good set of golf clubs and wear the right shoes so as not to slip when hitting the ball. Always get your stance right with good posture, get the grip right on the handle. Aim where the ball is going to go, Cheap Timberland boots keep your head still and take a full swing always trying to follow through. Keep practicing and try to get a little further each time and try not to tense up.

  27. five finger shoes to seize this market vacancy, the introduction of exclusive South Korean technology, launched an unprecedented vibram five fingers . mbt shoes  to come out and enter the market, same as in the current footwear market into a gravity bomb, rushed to ignite a new boiling point of care market, the design characteristics of the toe has become the new darling of the fashion world. five finger shoes , innovative ideas in the design, the Vibram Five Fingers apart, a perfect fit body motion characteristics, just like people’s “second skin”, mbt shoes  caused the emergence of an explosive sensation.

  28.  Choose the most suitable and elegant always means texting all the designer bridal showed at this century, and in recent years viewing all different and nice design is acceptable on the famous designers. Only after seeing all the designs that you will be able to make the final decision,

Comments are closed.