Product & Startup Builder

Filtering by Category: "personal relevancy"

Why does CNN not get it?

Added on by Chris Saad.

Why is this poll on's home page?
  1. I don't have a pet
  2. I don't want to own a pet
  3. There are plenty of other more important things for CNN to cover than pets
  4. Argg!

Come on people... please. How hard is it to learn my interests and serve up relevant content (I don't even dare asking for APML support). Even without tracking user interests, I can almost guarantee you that people visiting do not care about Dog food. Not on the front page!

An even broader question - do they not watch Jon Stewart? Do they not get it? The world is begging for real questions and real answers to real problems. How hard is it to stick to real news in this day and age. Surely they can leave Pet food to the Lifestyle channel?

Why do they waste our time with O.J Simpson? Ratings? Imagine the ratings they would get if they actually picked a fight with Washington - if they actually spelled out the truth of things for everyone to hear and see.

This is why Media 2.0 will win. We can use tools to find the real content and skip the garbage.

Showing pets love... buh.

Facebook is using your data to target ads at you

Added on by Chris Saad.
According to the Wall Street Journal online Facebook is designing an ad system to use their extensive knowledge of its users to target advertising to them.

This move is hardly unexpected. Chances are many sites across many usage models are considering and implementing the same thing.

The WSJ writes:

Next year, Facebook hopes to expand on the service, one person says, using algorithms to learn how receptive a person might be to an ad based on readily available information about activities and interests of not just a user but also his friends -- even if the user hasn't explicitly expressed interest in a given topic. Facebook could then target ads accordingly.

The question, however, is how long users are going to accept having their information harvested and leveraged in this way - the very heart of the Attention Economy - without demanding portability and transparency.

The WSJ article continues:

While Facebook plans to protect its users' privacy and possibly give them an option to keep certain information completely private, some Facebook users might rebel against the use of their personal information for the company's gain.

And the perceptions that targeted ads create can be as much of a problem as the reality. "Most people don't realize how targeting works; it becomes so good that even though it's anonymous, you feel like they know you," says Rishad Tobaccowala, CEO of Publicis Groupe-owned consulting firm Denuo Group. However, he says Facebook needs to be careful in implementing any targeted ad system, lest loyal users "find it creepy."

This is key. Maintaining privacy is just a subset of giving users control. Control must include portability and transparency.

Using export/import formats like APML would soften the impact of privacy/control concerns. The problem is that walled gardens like Facebook (and yes - it is a walled garden) think that they need to lock users in in order to maintain their unique value.

The truth is, however, that unless Facebook begins to adopt more standards and open up its platform for export, it will be usurped by the first medium-scale network to do so. Don't believe me? Remember that little network Facebook that blew Myspace & Linkedin up by opening up just a little?

Let's hope that Facebook considers taking some measures before rolling out their new ad system.

Michael Arrington doesn't get Personal Relevancy

Added on by Chris Saad.
Mark Lewis has written a piece over on Cnet about the need to flip the information delivery model. He writes:

"Web 2.0 flips the information delivery model upside down--it's now about global access, and information at your fingertips, aggregated from sources that you don't even necessarily know about, or care where they exist. Based on a set of search criteria, information in all its rich forms--media, video, audio, images, documents, text--all will be assembled together in context and delivered to users and applications for real-time experience."

That's a very poetic way of saying that in an age of hyper-choice, the most important challenge is to move beyond 'What's popular' toward what's 'Personally Relevant'.

I happen to also agree with Mark's suggested implementation - Source agnostic aggregation filtered by persistent search (and Attention Profiling) and delivered in real-time.

We call it Particls.

With the announcement of Streamy and Thoof, however, Michael Arrington over on Techcrunch has declared that Personalized news is pointless and will never work.

He's felt that way for a long time. I know... because he told me so while we were playing poker. A number of other people have suggested the same thing to me as well.

However, there are two things those people don't understand.
  1. Particls is not about Personalized News, it is about Personalized Alerting. We use the personalization part to rank content and determine how urgent the alert is for each user on an individualized basis.

    Thoof, Streamy and others are doing a very different (and worthwhile) job - and they are all potential partners of ours. We wish them the best of luck.

  2. Just because something has not worked before does not mean it is not worth doing again and again until it's done right. There is a place for popular, social news experiences (as Digg's popularity has proved) and there is a place for targeted, personal and solitary news experiences (as Digg's trolls and pop-culture content has proved).

Context and Aggregation are king

Added on by Chris Saad.
Daniela recently pointed me to this Bear Stearns report via her blog post.

In it they make the same observations that I and others have been talking about for more than a year.

"User-Generated Content (UGC) Is Not a Fad...
Some investors remain skeptical that UGC is more than a passing fad. However, in our recent online video survey, UGC is the No. 1 and No. 2 most popular content category among men aged 18-34 (M18-34) and among all respondents, respectively. Moreover, if we define UGC as page views only from sites such as,,,,, and (which is quite conservative), we estimate that UGC now accounts for 13% of total U.S. Internet traffic, up from 0%-1% in 2004. Based on these statistics, we submit that UGC is here to stay."

Although using the term UGC is not great, their conclusion sounds very familiar to anyone reading this blog.

"apparent to us that as supply of video content rises, value will shift from content producers to aggregators and packagers of content that can best aid users in finding content that fits their specific interests".

Of course, APML as a way of describing user interests, and Particls as a way of filtering and alerting users about new, personally relevant content, are both key technology pieces to this new media 2.0 reality.

The APML Business Imperative

Added on by Chris Saad.
Ian has a great write-up about why he loves APML.

He writes:

This got me thinking too, what if other more established places like Trustedplaces, Last.FM, etc also gave away a APML file as part of the profile of each user?

One of the things I loved about APML is the Implicit Data (U-AR) and Explicit Data (I-AM) elements. You can just imagine how simple it would be to output APML from something Last.FM. (whats below isn't true APML markup, just my lazy json like writing)

Implicit (U-AR) {
concept{ Ferry Corsten = 0.87 }
concept{ Armin Van Buuren = 0.90 }
concept{ Sugar Babes = 0.1 }
concept{ Lemonhead = 0.00001 }

He also mentions being asked "What is the business imperative to support such a thing".

In other words - if companies make their money from data lockin - then why would they want to give that data away.

I would suggest that anyone asking that question consider that publishers used to think like that. Now they all support RSS.

If feed readers thought like that, then OPML support and the rise of proper and continued innovation in the space may not have occurred.

If you are a smaller guy, supporting APML means that users can jump in and get started quickly. The barrier to entry is lowered.

If you are a bigger player it means an increasingly savvy user base will continue to trust your data mining activities. Also it means you can get a more complete picture of your users if they choose to share their APML from other services. It also means you become part of an ecosystem instead of a data silo - data silos are dead.

In the era of user empowerment, the business imperative is: play nice or users will move to other services that respect their rights. Just watch the mad rush to Facebook.

Web Analytics Demystified - Getting there at least

Added on by Chris Saad.
Jermiah, Web strategy guru at PodTech and fellow Media 2.0 Workgroup memeber has a great video interview with Eric Peterson posted on his blog.

They discuss Engagement, Attention an Analytics. Great chat - well worth a watch.

I have previously proposed some new concepts for syndicating Attention analytics information called Audient and Attent Streams which will go a long way in helping us move beyond pageviews and server logs.

I also appreciate Eric's approach of multiple factors for measuring Engagement. Very much like our Personal Relevancy engine, we don't think that any single factor or method is enough for determining the importance of an item to a user - so rather we create a complex algorithm that takes many things into account proportionally.

Check out the interview on Jeremiah's Blog and check out Eric Peterson's consulting company Web Analytics Demystified.

Collaborative Recommendation 3.0

Added on by Chris Saad.

Every now and then someone asks 'Why don't you build Collaborative Recommendation into Particls'.

In case you don't know, Collaborative Recommendation is when a system uses the recommendation of many people to 'decide' that a piece of content is worth seeing. So, like Digg for example, if 100 people vote that something is great (vote is a word I use loosely here), then it is probably worth seeing.

There are a few answers to that question.

  1. Particls is really about filtering noise out - not discovering new recommended content (even though we provide some of that functionality just to get novice users started).
  2. There are plenty of other great collaborative recommendation services out there, stick the RSS feed into Particls and there you go.
  3. Google's PageRank and Technorati's Authority is already a form of Collaborative Recommendation - it isn't really very new.
  4. The next generation of Collaborative Recommendation is actually something different. Let's call it Collaborative Filtering 3.0 or... Peer LifeStreams + Personal Relevancy

You have friends (hopefully); they have lifestreams (or at the very least RSS feeds from all their social/sharing sites) - plug their feeds into Particls and we filter out the stuff you don't care about. What's left? Stuff your friends 'recommended' that you actually care about. Collaborative Recommendation done right.

If you want to add me to your Collaborative Recommendation lineup, you can find me on Jaiku

AOL will update their feed reader soon

Added on by Chris Saad.
Techcrunch reports that AOL will be upgrading their feed reader with new AJAX tricks and better OPML support. Check out the report on Techcrunch.

Also Nick of Feeddemon has released version 2.5 of his great desktop feed reader including item sharing through News Bins and a great popular topics feature.

It's more than redundant to say now, but it's clear that aggregation (in all forms) is here to stay and will be one of the primary ways people recieve and manage their information. From the mainstream AOL offerings to the power-user ready Feeddemon.

The next frontier, of course, is adding in Personal Relevancy.

Major media outlets are starting to understand the zero sum game of Attention

Added on by Chris Saad.
Scott Karp has written a great piece summarizing what we here at Touchstone has been alluding to for quite some time.

Individuals can now make a good living as content creators, without ever creating or becoming part of a scale content business. What’s more disruptive, however, is that in the market for original content, the attention economy is draining dollars out of the cash economy. There remains a zero sum game for consumer attention, so for every minute a consumer spends with content created by an entity whose compensation is in form of attention, there’s a minute not being spend on content created by a for-profit entity.

In contrast, the content aggregation and distribution side of the divided media industry has all the advantages of scale, with the technology-enabled platform (e.g. MySpace, Facebook, YouTube, search) serving as the organizing principle for the new scalable media businesses. Content creation is asymptotically approaching commodity status, while platforms that can effectively aggregate content and allocate scarce consumer attention can unlock immense value in the new media marketplace.


This does not mean, however, that commercial content creators will lose out while Aggregators destroy their businesses. It means that content creators need to understand and respect and value the role of aggregators to help them find an audience. Further, they need to understand how Personalized Aggregation (based on Attention) changes the publisher/audience dynamic.

First, as we all know, there is no more audience, only participants. But more importantly for this discussion - the participants have different expectations. They want highly tailored content experiences that meet their tastes exactly. And they have no shortage of places to find that conent - in fact too many sources. We sometimes call this hyper choice or information overload.

This means that publishers need to:
  1. Start thinking niche.
  2. Start finding ways to cut through the noise to reach niche audiences.
This is where Attention comes in. By measuring one's Attention you can learn what they are interested in. By learning what they are interested in you can learn what content they want to see more of. From there, it's a hop, skip and a jump to connecting content creators with participants.

It also means that Aggregators will have a growing responsibility to content creators. A responsibility to report statistics, create transparency in their platforms and find some way to help the eco-system of Content Creators become successful.

Read Scott's full post for some more great insights.

Touchstone in your referrer stats - Audiences of One

Added on by Chris Saad.
People have started to notice Touchstone in their referrer logs. So I thought I would write a little about it.

I don't think anyone will ever see a 'Digg Effect' style mad rush from Touchstone. So we probably wont make headlines that way.

So what does a referrer from Touchstone mean?

I think it means something significant. Maybe even more significant than the Digg effect. It means that your article got through the Touchstone Personal Relevancy filter of our Attention Management Platform and connected with at least one person.

One person might not sound like much, but consider that one person after another might turn into hundreds and thousands. Consider also that each of those people are intimately interested in what they came to see.

Not only that - but the user clicked through (despite seeing your headline and synopsis) from inside the Touchstone UI.

With this in mind, Touchstone traffic could become a great measure of your sites ability to intimately connect with audiences of one - people just like you. People that might want to buy what you are selling.

The hardest thing I have to do every day is to decide what to ignore

Added on by Chris Saad.
What a great line:

The hardest thing I have to do every day is to decide what to ignore.

This comes from Jeremy Zawodny.

He goes on to say:

I need to invert my thinking. I should be starting most days with a strong idea in mind of what I want to spent the majority of the day focusing on. If there's time left, I'll tend to the other distractions.

This has implications for both business and media consumption:


Jeremy is correct. We must define our scope of interest first, and then make intelligent decisions about what to pay attention to.

That's what Touchstone does with APML. Your APML file (generated by Touchstone or any other APML compatible service) describes your scope of interest. Toucstone then ranks and filers incoming information for you against that profile.

Jeremy I'd be happy to give you a Beta Invite - drop me a line.

Some might say that this approach limits spontaneity or serendipidy. I'd argue that if you want spontaneity check Tailrank or Techmeme or Digg - they are fantastic Popularity/Meme Engines.

If you want a productive awareness of what you do all day, you need an Attention Management Engine.

Now some might say this sounds all academic and very 'Business Productivity' focused. But the reality is that this applies to media consumption as well. With a growing underbelly of great niche content, it is becoming very difficult for content creators to find an audience and audiences (or should we say participants) are finding it increasingly hard to pick the right entertainment experiences from a huge range of possible choices.

Thanks to Paul for pointing me to this post

I'm falling in love with my APML file

Added on by Chris Saad.
When you are intimately involved with developing a piece of software you grow to become unattached to your application configuration. At any given time the next experimental build might blow up and kill your settings or you might have to delete your config files to see what a 'fresh install' might look like for a new user.

Over the last year I have destroyed many an installation of Touchstone - each time thinking nothing of it... Deltree *.* (don't you remember DOS?)

Lately though, I have noticed a change. As the app has switched from a Swiss cheese set of features into a complete product my APML file - the file that contains my Personal Attention Profile - has started to become precious to me.

I can no longer just delete it and start again. It has grown to identify me. It produces content results that I like. I want to protect and nurture it. In fact now I back it up and carefully ensure that I never let it die in the process of trying the next experimental build from the dev team.

My APML file is becoming "the digital representation of my physical self" (Morpheus - Matrix 1).

It is obvious that APML is going to become something quite personal for people, and I would like to publicly recommit ourselves to protecting and nurturing it with all our might.

We are the user's ally in the fight against information overload and the search for great, personally relevant content. And APML is our BFG.

The Wizards of Buzz - The influencers deciding what's cool on our behalf

Added on by Chris Saad.
The Wall Street Journal has an interest post about "The Wizards of Buzz".

From the article:

"Most sites are based on a voting model. Members look around the Web for interesting items, such as video clips, blog entries or news articles. A member then writes a catchy description and posts it, along with a link to the material, on the site, in hopes that other members find it just as interesting and show their approval with an electronic thumbs-up vote. Items that receive enough votes rise in the rankings and appear on the front page, which can be seen by hundreds of thousands of people. When an item is submitted by a popular or influential member -- one whose postings are closely followed by fellow members -- it can have a much better shot at making the front page."
It's a little scary. They imply that services like Digg, Reddit and Netscape have made influencers out of little-known everyday people. Why is that scary? Because we don't know these people. They have not been vetted by public opinion and to many users they are an opaque part of the process. It's not democracy if there is a small group of people pulling the strings.

No one diggs around digg looking for the 'Top Diggers List' - in fact now they CAN'T dig around Digg for it - because Digg has taken it offline. Check out the article to get a list of the top Digg, Reddit, Netscape and StumbleUpn user they found. It's not a list of people I want deciding my news for me.

So these popularity platforms are giving rise to micro-influencers who are actually having a huge affect on our news and information choices and most of us have very little idea who they are. That doesn't sound very social, transparent or desirable to me.

As I have said in previous posts - while popularity engines are fine for working out "what's cool" the real question should be "what is personally relevant: - finding news that affects my life and aligns with my interests.

Then, the only influencer in my media consumption is me and my Attention Profile.

Jason Calacanis thinks having top influencers is great. I guess he would because he also thinks paying the top contributors is great too. I'd invite him consider the Personal Relevance angle (he seems to be taking up challenges this month so why not).

Thanks to Marianne for pointing out the WSJ post to me.

Brad Burnham asks "What's Next?"

Added on by Chris Saad.
Brad writes a great post on the Union Square Ventures blog asking "What's next?". In it he explains that innovation moves up the stack and that we are now at the 'data' layer.

"In the early days, the central value proposition in the computer business was hardware. Later, it shifted to systems software, then applications software, and then networks. As more software functionality was delivered to a browser over the Internet, the basis of competition shifted from features to service level metrics like reliability, accessibility and security. I believe that today, at least in the area of consumer web services, we have already moved on to a new focus of competitive differentiation based on data."

I am glad VCs are starting to ask these questions. It shows they are thinking beyond 'me too' services.

He goes on to write:

"One way to look at that question is to argue that we have arrived at the end of history. The progression to date has been up the stack in a classic architecture diagram, data is on top of that stack, and nothing sits on top of the data. I disagree."

I disagree too Brad. Here are a few more layers to think about.

Once data is structured and syndicated, the next job is to:

  1. Aggregate - And I don't mean like a feed reader. I mean like Edgeio and Vast

  2. Personalize - Using Attention Data

  3. Visualize - Using all sorts of techniques. With Touchstone we are using escalating alerts. The more personally relevant the content (from step 2) the bigger the alert.

How to get linked from the A-list

Added on by Chris Saad.
Robert Scoble is a genius. I will say this over and over. If there's one thing he knows how to do is to create a brand of his name as the A-list blogger of the people. His trademark 'Who are you' opening question, his disarming laugh, his simple 'everyman' questions (most of which he knows the answer to I'm sure) and his ability to stem the flow of negativity with brilliant stunts all contribute to his power.

But this post is not just to suck up to Robert - I'd like to ask a question.

His latest post (and stunt) is a thread where he asks the question "Do A-lister bloggers have a responsibility to link to others". In it, he asks that question and then opens the comments for everyone to spam a link to their own stuff.

One of the commenter’s, though, raises a very interesting point.

Krishna Kumar Writes:

The PageRank algorithm is probably one of the key factors in this whole argument about link sharing. While the initial search engines used the “content” of your web site or page, nowadays (because of content spammers) authority (determined by incoming links) matters more.

The problem is that if a newbie or Z-lister has something really important to say or has some great idea, he or she will not get the necessary audience to propagate that idea.

I am not sure how this can be resolved because the commercialization of the Internet along with SEO businesses have changed the rules of the game that unfortunately now negatively affects new ideas.

And yes, a tech-savvy person can get his or her idea spread, but what if the person (non-profit, medical field, etc.) has no clue about Google juice and stuff like that.

I know that back in my Z-list days (I am now on the Y list for those keeping track) it was/is hard to get a post you think is fantastic noticed by hardly anyone. But is that because the A-list is so hard to break into or because the tools for mining the long-tail are so poor?

Does Google Juice matter? Does being on the A-list matter? Whose A-list are we talking about?

I've said it before and I will say it again. Personal Relevance is more important than Popularity.

People who care about what I'm saying should find it - irrespective of how many incoming links I have.

Why? Maybe because I am not as popular as Robert but I still want to be heard. Don't we all? But more importantly because a local school does not need (or want) Robert's audience. They want an audience of locals. And locals should be able to discover that content without knowing what RSS is.

How relevant is relevant?

Added on by Chris Saad.
Chris Anderson has a great post called 'The vanishing Point Theory of News'. He posts a great little musing about the relativity of relevance.

"For instance, the news that my daughter got a scraped knee on the playground today means more to me than a car bombing in Kandahar. [...] Am I proud of this? No. But it's true."

He goes on to say:

"There's nothing new about this (it's a truism of the American newsroom that Paris, Texas counts for more than Paris, France), but it bears repeating. The future of media is to stop boring us with news that doesn't relate to our lives. I'll start reading my "local" newspaper again when it covers my block."

I often tell people (in regard to Touchstone) "Put in the name of your kid's school and your favorite golf course" - how much more local can you get.

We can't work out if your daughter got a scraped knee unfortunately - unless the school blogs about it maybe.

Book: The Paradox of Choice

Added on by Chris Saad.
I'm starting to sound like a broken record now - but here is yet another quote about hyperchoice from a lengthy interview with Barry Schwartz, the author of 'The Paradox of Choice'.

“The problem used to be, ‘how do we get information out to people?’ That problem has now been solved in spades. Now the problem is, ‘how do we filter the information so that people can actually use it?’”
He is concerned that filtering technologies may not be up to the task of helping us deal with the overabundance of choice and information we seem to find ourselves having at the moment.

We'll see what we can do.

Thanks to Marjolein of CleverCogs for this!

Channel ME

Added on by Chris Saad.
Mark Sigal's recently posted about Channel Me and the Rules of New Media.

He talks about concepts that we have long discussed here such as:

  1. If content was king, then aggregation is now the master of the universe
    He writes: Unlike "old media," where content was the star, in new media, it is about the users and giving them control of what they digest, how they digest it and with whom. This article attempts to provide a framework for thinking about the rules of new media and how to work them to your benefit.
  2. The audience has left the building
    He writes: Once upon a time, content was content, an ad was an ad and the audience was a passive consumer. No more. Increasingly, the lines between consumer and producer are getting blurry.
  3. Personal Relevancy is more important than What's Popular
    He writes: These tools will have built in recognition systems (like deep profiles) to systematically connect like minds together, and filters that provide transparency that highlights what’s new, popular, recently viewed, talked about or related content.

And he finishes with:

The evolution of the Web from text, pictures and links to video-powered social nets is as profound as the evolution of broadcast media from radio to television, and it is destined to be no less exciting.

I wholeheartedly agree Mark.

Thanks for pointing this post out Randal