The Evolution from Numbers to Relevance

Social media and Businesses on the web today are driven by the numbers game – of traffic, page views, and follower numbers. But the trend I foresee is:

The web is evolving from a numbers model to a relevance model.

Paradigm Shift: What is the Relevance Model?

Historically, monetization driven by CPC/CPM based advertising has led to websites and marketers focusing on page views and traffic. This is partly the cause of social media being spammed by internet marketers, ranking algorithms being gamed for traffic, and so on.

Numbers Model

Relevance Model

# of Followers Context-driven Lists
# of Clicks # of Interactions
# of Page Views # of Returning Visitors
# of Ads Displayed Time spent on site
# of Ads Clicked # of Subscriptions Gained
Obnoxious Ads Relevant Ads
Influence Management Dynamic Social Graph
Sharing Orgy & Noise Curation
Information Overload Filtered, Relevant Information
Traffic Economy Attention Economy
SEO and SMO Personalization


The above table lists different attributes of this paradigm shift. The “Influence Management” entry links to a post by Mia Dand who describes how leveraging social media is often about using a handful of influencers (read: with large follower numbers) to spread your message. Contrast that with Dynamic Social Graphs as described by Robert Scoble, where influence is dynamically determined based on relevance and not just numbers.

The Facebook Kingdom was built on Relevance

The king of the social web, Facebook, was not built on numbers, but relevance.

The success of Facebook and why it has garnered over 400 million users is because it grew on a base of real-life friends who were relevant in the users’ social circle. Other networks have failed to challenge Facebook partly because they have tried to go the other way around – from numbers to relevance.Bullseye

Prioritizing numbers over relevance is putting the cart in front of the horse.

Even as its explosive growth continues unabated, Facebook has not compromised on relevance. It knows that its success depends on users finding relevant content on Facebook and is willing to sacrifice advertising revenue to avoid becoming irrelevant.

I’ve touched upon various aspects of this ongoing theme while tracking the Google vs. Facebook race towards a relevant real-time. It’s becoming increasingly apparent that relevance wins over real-time.

While Facebook has never been in the numbers game, other networks like Digg are now moving from the numbers model to the relevance model.

Relevance vs. Real-Time in Location Check-ins

Consider the hottest trend of check-ins via location services, such as Foursquare or Gowalla.

When I check-in at a restaurant, the real-time checkins of my friends in other places is irrelevant. What is more important and relevant to me is the tips from my friends who have checked-in at the same place as I am right now.

In all cases, my friends are relevant in real-time only if they are at the same location as me. My other friends NOT at the same location become irrelevant.

Relevance wins over real-time.

The Mobile View

While mobile internet access grows, the screen of mobile devices remains constrained by its form factor. This is a major factor driving this evolution. If the content on your screen is constrained by its display, it had better be relevant.

Lifestreaming and Aggregation

As I discussed extensively in my post on why Google Buzz should not simply be yet-another-aggregator, lifestreaming and aggregation have failed to take off and gain mainstream adoption. The reason is simple – lack of relevance.

Which is why, it is personally heartening to see the champions of lifestreaming and aggregation turn their focus towards relevance and disaggregation.

Startups focusing on Relevance

Quite a few startups are hoping to capitalize on this trend:

  • my6sense – recently introduced an ‘Attention API’ allowing publishers to deliver relevant content to users
  • Cadmus – auto-filters Twitter/RSS streams by relevance
  • Knowmore – surfaces relevant stuff from Twitter/Facebook
  • TwitterTimes – personalized aggregation from Twitter
  • FeedTrace – personalized aggregation from Twitter
  • VictusMedia – ‘Intelligent Media Manager’
  • MixPanel – tracking what I’ll term “Relevance Analytics” for publishers
  • Cascaad – personalized news stream based on social graph from Twitter/Facebook

From Around the Web

Here are related posts that further elaborate on this evolution:

This entry was posted in Social Web and tagged , , , , . Bookmark the permalink.
  • Pingback: Relevancy, Curation & Transparency = Your Monday Signal()

  • Pingback: Come On Nick, You Can Do Better Than That()

  • Alright, big shot. It seems you're participating in the same circles I am, seeing the same things and talking about the same developments. Yes, I have some small bias, thanks to working with my6sense on relevance, but I am seeing tools for this coming from everywhere.

    If 2008 and 2009 were all about creating the noise, 2010 is about finding ways and tools to break through the noise and find what is personally relevant.

    It was a pleasure to talk with The Cadmus guys tonight and learn more about their work to find personalization in your social streams. They get it. Google Reader also gets it and is working on making their Magic better. (Which you missed in this post) 🙂

    I don't care about page views, except as small clues as to trends and engagement. I do care about on-topic discussions and discovery. I would rather talk about something new and relevant that a small focused group cares about than a mediocre one that many people glance at.

    Thanks for the discussion here and watching and curating so well.

  • Mahendra, most users miss out on the relevant part – for those who understand the sphere that's what it's all about and in my own little way have too gone down that route.

    Another great post, and a great eye you have for trends. 12 hours a day on your desktop, i'd say that the fruits of your watchful eyes are steadily materializing…

  • Agreed on Mahendra watching and curating so well..

  • Louis, I am no big shot, live on the opposite side of the world from Silicon Valley, but merely try to observe, gain, and share insight. 🙂

    I am excited about developments in this sphere, so am happy to promote startups like The Cadmus guys. I've already covered enough Google on my blog, and have always maintained that they have the best resources to achieve algorithmic relevance.

    Appreciate your sharing and the feedback!

  • Thanks, Addy. Appreciate your feedback!

  • Well weaved thoughts. Appreciate the kind words Mahendra, thank you. Relevance is personally defined but shares patterns between many users, so it has been difficult for one social tool to capitalize on. Facebook succeeded maintaining small/social while growing global and this served them well. But it's little surprise that there are many who haven't found Facebook as relevant to our social web needs.

    Choice must be part of every satisfying social engagement. We need to experience the freedom of opting in, as well as opting out. But thus far services like Facebook are difficult to avoid, when many of our friends shares can only be observed within its walls. The variety of tools that improve the soc web will provid this choice. My goal is to help construct the “WordPress” of the social streaming web. An open source feed based utility that enables powerful filtering, grouping, and curating with a simple default setup that's highly programmable. This will be done leveraging the great work of folks developing open social standards. We're working now on a functional framework that others can build on.

    Our business Victus Media, has made a number of directional changes as we've worked on tools for the social/information web space.

  • Pingback: News Update – Best of the Day()

  • gregdelima

    A phenomenal post with excellent relevance. I thing from my perspective, that the most important part is the paradigm shift you talk about. By being able to shift from a generalized advertising model to one that is much more focused as marketers and advertisers, the availability of information, demographics, and even psychological insights due to keywords in streams gives an ability to hone in on who we look to.
    Also, as Louis says, we're looking to make things “personally relevant” we're looking into the long tail, we're finding what we want and we're exploiting the availability of the information!
    Focusing on what we can find and how we can use the vast information available depends entirely on curating real-time information in a way that's digestible for everyone.

  • Good post. But I'd actually dwell a little more on relevance and real-time. While 'relevance wins over real time' works in the scenario you've outlined (and the way many services are working now), I'd actually like it better if they worked in a complimentary fashion. Taking the example of FS/Gowalla, and looking at it from a different perspective, lets assume I like a particular brand of beer and have been tweeting about it regularly. If I check in at a pub, and they're able to give me say, an offer on that brand, because they've been monitoring the tweets about that brand (relevance) and have mapped it with real time activity on location based services (Foursquare is rolling it out for businesses), doesn't that work better than just making an offer on Twitter (even if its @ me) at some random time?
    I do admit that's quite some mining, but i also think we aren't too far away from that either.
    Again, thanks for a good set of thoughts 🙂

  • Mark – thank you. When citing Facebook, I didn't mean to imply that it has perfected the art of relevance, but it has been better than other networks. I think Facebook has lacked appeal to users who found the app/game notifications annoying precisely because they were irrelevant. And they're learning from that mistake, with the changes introduced for app notifications.

    I'm enthusiastic about what VictusMedia is doing and look forward to your tool. Good luck!

  • Gregdelima, thank you. I think the metrics and tools for targeted advertising may be lacking today, but the market need for such data is very real and exists. I think it's just a matter of time that we begin to see such advertising tools being developed. In fact, Mark from VictusMedia had a prototype for keyword-in-streams based real-time advertising, if I'm not mistaken.

  • Manu, absolutely. When I say relevancy wins over real-time, I am considering scenarios where you can only achieve one and not the other. Achieving both is the ultimate goal in tech right now.

    In your example, the beer offer becomes relevant because you are at that location, and is irrelevant if tweeted to you 'in real-time' within seconds of that offer being introduced by the beer manufacturer.

  • alliefisher

    Great article Mahendra, I certainly agree with you that the web is rapidly moving towards a relevance model. With the infinite choice now offered by the web, smarter tools are needed to aid user discovery and engagement. Working at The Filter, we are passionate about equipping companies with the recommendation and relevance tools that will match users to the web content they care about.

    We've witnessed a significant shift in the content space of late, which has meant an increased demand for filtering technology. Content owners, publishers and platforms are keen to make sense of their growing consumption and catalogue data, and use the information in a way, that provides users with relevant content suggestions, promoting consumer loyalty. This in turn, has meant the old hierarchy of monetization metrics has changed. As your model suggests, engagement and customer loyalty metrics such as dwell time and user interactions have become key.

  • asuph

    This is easily one of the best posts I've read about Social Media, Mahendra. And you know I won't say this because I know you. It's almost a lesson on how to write a non-fiction article. Coherent, articulate, organized, and informative.

    And I hope you're right 🙂

  • Great post! You might want to add Cascaad to your list of companies focusing personal relevance. Cascaad leverages your social graph and interest profile to help filter what might be of personal interest on the real-time web. We have also released the SuperTweet API that lets developers add valuable semantic context, in-text affiliate services and relevance scores to Twitter posts.

  • I'm keen on relevance but the ways that you get there can be very different. Most of what I'm seeing are ways in which to translate the noise into signal.

    But isn't that really attacking the symptom and not the cause? The real cause of this information overload (noise) is brought on by oversubscribing and follow abuse.

    Many of the tools coming out (Cadmus seems interesting) identify these dupes but … shouldn't that be extended further? If I constantly see dupes from a person, do I really need to follow them? Sure, if I have a personal relationship then keep them in your social graph, but otherwise, why are you REALLY following that person?

    In the scheme of things I don't have to follow a lot of people, I just need to follow the 'right' people. And by right I mean those who are great human filters of information. That's Louis Gray, that's Rob Diana, that's Atul Aurora and you.

    I've been sketching out a tool that identifies people who you probably could drop from your social graph. THAT is what I believe would go a long way to solving relevance.

    Addition by subtraction.

  • Allie, thanks for the feedback and I'm glad to see you are seeing these trends as well.

  • Thanks, Erik…done! 🙂

  • AJ,

    Absolutely – relevance is a very personal thing and I'm highlighting the shift of focus from traffic to relevance. There is a long way to go. The beast of traffic has to migrate from the oceans to land, after which we will see intelligence evolve. 🙂

    The tool you describe seems to help curate one's social graph – part of the Dynamic Social Graph concept – which is one defining element for relevance.

  • Mahendra,

    Love the oceans to land analogy. And yes, the tool I'm describing could be part of the Dynamic Social Graph but it's less about dynamism and more about … scale.

    I simply believe that reducing the number of inputs could greatly increase the ability to provide relevance. The GIGO adage springs to mind. Now, it's a bit harsh to refer to some in your social graph as garbage but … the key is to ensure the inputs are solid, or else getting any output becomes difficult if not nearly impossible.

    For each topic area I'd guess you could get the information you need through fewer than 150 people.

    Mind you, in the scheme of things we might be looking at the 'bubble' view of this problem. I'm guessing most people (e.g – mainstream) don't have an information overload problem and instead still need help finding interesting and relevant people and content.

    Are we solving for our own, early adopter, issues or for the mainstream?

  • Excellent post Mahendra. Thank you. I caught it on Ahad's page.

    We talk alot about “relevance” and “recurrence” as the 2 big ideas that drive our work at HiveFire. We think it's less about numbers and gaming, as well. If you deliver high relevance, you'll get high recurrence. In the attention economy, high recurrence means you've acquired valuable attention.

    It would be a pleasure to connect and speak about HiveFire, if we've not yet connected.

  • Pingback: links for 2010-03-30()

  • Pingback: Medial Digital» Audio im Web Interviews mit Innovatoren Neu Social Web » Michael Prätorius über Audio in der Aufmerksamkeitsökonomie()

  • Pingback: Steady stream of information | Admin Toolbox()

  • The continuum between early adopters and mainstream users is long and has millions of users.

    If you consider 'mainstream users' as those who can't describe what a browser is, or can't distinguish between the Facebook Login page and the RWW blog, there are millions of non-mainstream, non-early-adopter users who are able to choose their browser and are web-savvy. These users face information overload and are caught up in the numbers game at present, IMO.

  • Thanks for the connect, Taariq. I'm following you on Twitter.

  • Pingback: How To Use Twitter To Promote Your Photography Business()

  • Pingback: A New Focus on Relevance « Reinventing the Newsroom()

  • Pingback: Link Love Monthly: March — SocialFish()

  • Agree completely. I prefer topical relevance to “people I know” type relevance.

    I follow hashtags now on twitter instead of people in order to maximize relevance. It is still very people oriented because periodic hashtag chats ( and regular posters forms a community.

  • Agree completely. I prefer topical relevance to “people I know” type relevance.

    I follow hashtags now on twitter instead of people in order to maximize relevance. It is still very people oriented because periodic hashtag chats ( and regular posters forms a community.

  • Brilliant post. (I'm a bit behind in my reading so pardon my tardiness). You exactly nailed the need for relevance as well as highlighting the trend.

    One of your best Mahendra!

  • Keith, thank you for the kind words. Your feedback is highly appreciated!

  • Pingback: What To Read This Weekend: The iPad Edition()

  • Pingback: #Tip of the day from – filtering social media information | Editors' Blog()

  • Very late to this post Mahendra but you've done a great job outlining this. Also thanks for the links. I recently spoke to a founder at Knowmore and based on the conversation I'm really looking forward to looking at their service. Many of these are limiting their relevance to our social graph at Twitter and Facebook and I feel that's very limiting. Many of us publish and interact on many other large services like YouTube, Flickr, Delicious, etc and to not include these and segregate by them as well is far to limiting.

    It's great to see so many services out there trying to solve this problem, I'm hoping but not so optimistic it will happen very soon though. I'd be more than happy to be proven wrong though.

  • Hey Mark, thanks for the comment. I agree – and Kevin Anderson from The Guardian also does – see last para at…. Looks like even Yahoo's in the game:….

    So yeah, this space is certainly getting a lot of momentum, and I too think we'll start to see some good tools soon.

  • Pingback: Relevance()

  • Pingback: DataSift Curation Engine Aims for Relevance in Real-time by @ScepticGeek()

  • Pingback: Filtering for Relevance with my6sense by @ScepticGeek()

  • Pingback: Comparing Approaches to Information Filtering for Relevance – Mahendra Palsule - MediaNama()

  • Pingback: The Year in Lifestreaming for 2010 | Lifestream Blog()

  • Pingback: and Newsmix Show Glimpses of the Social Content Reader Evolution | Lifestream Blog()

  • Pingback: Newspapers versus Netflix: Adventures in the attention economy()