This was an extremely interesting read. Your summary of the benefits of the APML attention profile are dead on! Great work man!
Attention Profiling: APML Beginner's Guide
Attention profiling is quite possibly the next wave of smarter, more relevant web browsing. APML is an emerging standard that promises to make it much easier for websites and services to cater to your exact tastes, reducing the information overload of seemingly endless web content.
Photo credit: Faraday Media
When you browse the web and sign up for various social media services, information is being collected about your interests and tastes. If you shop at Amazon.com, for instance, each time you look at a different book or DVD, Amazon makes a note of what you were checking out and makes suggestions as to other titles you might be interested in.
And when you sign up for social networking websites such as MySpace or Facebook, the explicit information you give - your gender, age, favorite recording artists and so on - is gathered, in addition to the more implicit information about which profiles you visit, or Facebook applications you install.
For the companies involved this is big business - as there are marketing firms willing to pay a lot of money for this sort of information.
For the end-user, though, there are certain benefits too.
We have reached the point of information hyper-saturation. It can become quite a chore to find relevant content online, when there is so much other information competing for your attention. But by implementing attention profiling, it becomes possible to have the services and websites you visit begin to make suggestions for content that you might be interested in.
APML is a proposed standard that gives you greater control over your own attention data, and in principle will allow you to selectively record your attention profile - the sites you visit, the search terms that interest you most, the content you most commonly link to - and share it with your favorite websites and services.
It is already being supported by several prominent online destinations, and shows no signs of slowing down.
In this beginner's guide to APML I talk you through the basics, how APML fits into a wider trend as we move towards a smarter "semantic web", and how you can make use of it today in your day to day web surfing and information seeking.
Attention Data - What It Is and Why It Matters
Photo credit: Faraday Media
In her excellent introduction to attention profiling through APML, Marjolein Hoekstra describes attention profiles as:
"...consolidated, structured descriptions of people's interests and dislikes. The information about your interests and how much each means to you (ranking) is stored in a way so that computers and web-based services can easily read it, interpret it, process it and pass it on should you request and permit them to do so."
APML provides a standardized way to collect and rank your attention data, then, in such a way that the information gathered is useful to the sites and services you use on the web, that they might better provide you with tailor-made information and content that suits your particular needs.
Your attention data is comprised of the websites you visit, the things you might write about in your blog, the music you listen to through such services as Last.fm, the websites you bookmark using social bookmarking tools like del.icio.us, the photos and videos you share with services like Flickr and YouTube, and so on.
As you surf the web and interact with these various social media destinations you leave behind you a vapor trail of information that, when pieced together, describes the content you enjoy and find interesting.
Of course you aren't alone, and whether you like it or not your search habits on Google, the applications you install on Facebook, and the content you write about in your blog are ripe for being "data mined" by those with a vested interest in compiling such information.
APML provides a means that you could personally take control of your information in a consolidated file, to use where you see fit to your own ends and benefit. And at the same time advertisers and content providers are given a much clearer picture of what might interest you, saving you the hassle of wading through irrelevant interruptive marketing, and being presented with information and offers that might very well be worth your taking a look at.
Web Standards - The Basics
Web standards matter.
Standards are essentially a commonly agreed set of protocols, or way of arranging data for maximum interoperability. In layman's terms, what that means is that if everyone uses the same system for processing information, that information will behave in a predictable way wherever it crops up.
Take this website, for instance. You can view the content in any modern browser without any great difficulty, because the content has been formatted using common standards, such as HTML and CSS, as standardized by the World Wide Web Consortium, or WC3. Were it not for standards, every website you visited might act and display in a remarkably different way depending on the browser you used, or the means you used to access the site.
If you want to access the information from Master New Media from your feed reader, that's also possible thanks to standards. The RSS standard has been adopted by the vast majority of modern websites and is all but ubiquitous in those websites published using blogging platforms. This means that whichever RSS reader I use, I can access the posts written at Master New Media via one of the several RSS feeds on offer.
Now let's say that I want to share my collection of RSS feeds with you in the form of a reading list, so that you can automatically subscribe to all of the websites I follow with a single click. Then I'd use the OPML standard, pass you the OPML file of my reading list, and you'd have access to my entire feed collection.
APML is an evolving standard that is gradually being adopted by more and more sites and services as a way of tracking and sharing attention data. So just as I can now share the contents of my blog with you via RSS, or my entire reading list with you via OPML, APML makes it possible for me to share a ranked list of my online browsing habits and interests.
Obviously the more sites and services that implement APML, the more useful it will become. At the moment, for instance, Amazon can suggest books to me based on my previous selections, but if I leave Amazon and go over to Barnes and Noble, the process has to begin from scratch. Were both sites to adopt the APML standard, I could take my interests with me from one site to the other, and get recommendations from both without each first having to rely on its own closed system of analyzing my attention data.
APML - The Basics
Photo credit: Faraday Media
As I've highlighted APML is a way of collecting and aggregating all of your attention data together into a single "attention profile". This profile exists in the form of an XML file - which is the same open language behind RSS and OPML.
When you make use of an APML-enabled service or application, your attention profile is pieced together into this easy-to-share XML file, which you can then take away with you and import to other services that support the standard.
Not only does your APML attention profile collect data about what you are interested in - using the URLs you visit, the tags you attach to photos and videos, the songs you choose to listen to - but it also gathers information about how interested you are in the various content you visit.
So if you regularly navigate to the communication tools and reviews at Master New Media, but aren't so interested in the counter-information news, or vice versa, over time your attention profile will rank your interests accordingly.
In this way services using your attention profile to serve you with relevant content will be far more likely to fetch you information on your favorite subjects, rather than those you simply showed a short-term or passing interest in.
It's easy to see how, over time, this could be a very useful tool for working with a number of different services whilst maintaining a highly detailed and therefore far more accurate reflection of your needs and interests as an end-user than is commonly provided by the closed systems employed by specific websites.
Why Use APML?
Photo credit: Faraday Media
The biggest reason you might want to use APML is the sheer amount of information bombarding you every day, and the ratio of signal to noise evident in that information. Because while you may have very specific needs and interests that you want served, it is not always a simple task to locate the information you're looking for.
Furthermore, as advertising is here to stay on the web, it would certainly be a lot better news if the ads you saw were tailor-made to your interests.
Think of this as the next step forward from the contextual advertising you can see in this article. Contextual advertising services such as Google Adsense attempt to serve relevant advertising based on the content of the article that they appear in. As such, you have a greater chance of seeing ads that will appeal to your tastes than you might through mainstream, mass media advertising, which simply sends out the same message almost regardless of context, and hopes that some small percentage of viewers will be interested.
If the APML standard takes hold, however, content providers and advertisers will have a much better chance to serve you with relevant information, so that ads become useful rather than something that interrupts what you came to see in the first place.
Marketing aside, given that many of us have busy schedules and a lot of information to process in the average day, it would certainly be very useful if it were possible to take your attention profile shopping with you, or as you research your paper or novel. Rather than having to go out looking for information in a hit-and-miss way, the information comes to you, and learns over time what interests you most through your acceptance and rejection of the offers served up.
So rather than a generic news website front page where I have to navigate to the sections that interest me, the news site could automatically display a custom set of headlines based on my interests and needs. If I'm interested in following the latest stock prices for Google, gossip about Hollywood celebrities, and what's showing at my local cinema (so long as it isn't a romantic comedy, and it rates highly among other users), it's clear to see how the implementation of APML might help me to cut through the endless information I'm not interested in.
APML in Practice
Photo credit: Faraday Media
While AMPL is an evolving standard and is still in its infancy in terms of uptake, there has already been a considerable amount of interest in the standard, and it is now possible to make use of an attention profile through a number of popular web services and applications.
Given the rate at which APML uptake seems to be spreading, it would be a good idea to check out the growing list of implementations at the APML wiki. At the time of writing, you can make use of APML using the following sites and services:
- Particls is a great way of receiving alerts about the news that interests you most in a news ticker, sidebar and pop-up format on your Windows desktop (Mac version coming soon). Particls creates a hierarchy from the news sources you subscribe to, highlighting the more important news and slightly downplaying the news likely to be of less interest to you.
Particls creates an APML file using the data it gathers from your news selection, along with that collected from your IM conversations, emails, browsing habits and documents. This APML file is then used to create a hierarchy of alerts for the latest news likely to be of interest to you
- Engagd is a service tailor-made to creating your APML file through your personal selection of RSS feeds. It also allows you to filter your RSS feeds according to your evolving profile. You can see a great write up of how this works in practice in Emily Chang's write-up of her own use of Engagd to create a personal data stream.
- Cluztr is a social network based on sharing your clickstream (the things you click on as you browse the web) with friends. Cluztr creates an APML file based on your browsing habits, and uses it to automatically generate tags for other users to explore content through
- Dandelife is a service for creating a socially powered biography, gathering your notes on the people you've met, the places you've been, the events you've been to and more, and creating an APML file from the results
- Additionally the popular Newsgator and Bloglines RSS services have adopted or promised future adoption of APML to aid your feed reading.
Of the services mentioned, perhaps the best introduction to creating your own APML file might be through trying it for yourself through Engagd. It's clear to see, however, that with news of web giants like Google and Digg backing APML, you can expect to hear a lot more about it in the coming months and years.
Photo credit: Dem10
Perhaps one of the biggest objections that people have to trying out APML is the intrusion into their privacy that it potentially presents. Sure it's a nice idea to be served with relevant content, but what if there are certain parts of my web browsing and social media use I'd rather not have made public?
In his coverage of APML for Mashable Mark 'Rizzn' Hopkins voices the concern that should APML become standardized:
"various vendors and APML consuming software now knows exactly what sort of porn sites I may be paying the most attention to, for instance, or about research I may have done on militant Islamic websites for a political piece for my blog - something considered dangerous information these days. I’m just not comfortable with that sort of information sitting out there in the public’s hands."
This is an understandable concern, but as Mark goes on to explain, after a conversation with the APML workgroup's Chris Saad, APML is selective in the information it gathers, so that a passing interest in terrorism, as might come about when writing a paper on the subject, will barely register when held against the things that you show interest in on a more consistent basis.
Likewise, as for "private browsing activities", such as pornography, the services that implement APML can be set to ignore certain browsing habits. Particls and Cluztr, for instance, will ignore pornographic content as they put together your profile, so there is no fear of your APML attention profile showing as much interest in a certain genre of porn as it does your professional proclivities.
In fact privacy is a central factor of APML, which aims to put the attention data you generate in your control, rather than in the control of the different sites and services you use, as is now the case.
In a section of the APML website dedicated to "your rights", the APML workgroup explains that your fundamental rights include:
- "Property - You own your attention and can store it wherever you wish. You have CONTROL.
- Mobility - You can securely move your attention wherever you want whenever you want to. You have the ability to TRANSFER your attention.
- Economy - You can pay attention to whomever you wish and receive value in return. Your attention has WORTH.
- Transparency - You can see exactly how your attention is being used. You can DECIDE who you trust."
These points are extremely important in understanding that APML is not so much a big-brother like device for monitoring every breath you take as it is a means of creating a useful, open relationship between content producers and consumers based on ownership and sharing of your own attention data.
The APML Workgroup
Photo credit: Faraday Media
The APML Workgroup are the numerous individuals gathered together with the task of "turning the current specification into an agreed standard", as the APML website explains.
This group is made up of a number of developers, CEOs and advocates working within the "attention economy".
The workgroup has both an open wiki, a Facebook group and a public Google group, so should you wish to become involved in the development of the standard, or to follow its progress, there are several means of doing so.
Data Portability - The Bigger Picture
While APML stands up as its own standard, it is also possible to see it as part of a bigger picture.
As the web evolves we are seeing a great shift towards smart information filtering - the evolving notion of a "semantic web", but also, just as significantly, a move in the direction of expanding data portability.
Open standards make for an effective way of allowing data to freely flow from one web destination to another, rather than keeping different sets of data in closed silos and walled gardens. At the moment if I want to have a profile on MySpace and Facebook, for instance, I have to create them separately, and any information I enter on each remains locked into that particular destination.
Open standards and data portability are all about allowing me to take my information and use it across a number of services.
If you pay a visit to DataPortability.org it becomes clear that a number of open standards used in tandem would allow for a much freer web, a web that allows you to take control of your data and freely remix and share it as you see fit. APML might be joined by the following complimentary open standards, which are in various stages of current implementation:
- OpenID - OpenID is a standard for the creation of a single digitial identity that you can use across the web. Effectively this will allow you to leave comments on blogs, sign in to payment services, and access a number of social networks using a single sign-on
- OAuth - OAuth is an open protocol that would make it easy for application developers to use a common standard for authentication, allowing different services to authenticate your secure data with ease while maintaining your privacy
- RSS - RSS makes it easy for users to syndicate content from around the web and receive new information every time their chosen news sources are updated
- Yadis - Yadis is, in their own words "a service discovery system allowing relying parties (aka identity consumers or membersites) to determine automatically, without end-user intervention, the most appropriate protocol to use"
- OPML - OPML is an easy way for people to share their reading list of RSS feeds with other users
- hCard - hCard is a microformat that makes it easy to share and download formatted contact details across a number of applications - rather like a business card for the web
- XFN - XFN is another microformat, and provides an open standard for gathering and sharing relationship information. XFN stands for XHTML Friends Network. Using XFN I can share the people I connect and network with across the various social networking platforms that I use, rather than having to add new "friends" for each service that I subscribe to
What each of these protocols shares in common with APML is a commitment to a more open web in which I can take control of my own data, and use it across a number of sites and services, rather than having to set up new accounts, friend lists, feed lists and so on at each service I choose to use.
APML is an increasingly important standard that makes it easy for me to keep an automated record (or "attention profile") of my interests and tastes as I browse the web and use a variety of social media services and applications.
So, as I surf the web, upload and tag videos and photos and listen to music, my attention profile is updated to mirror the things that interest me most.
The benefits to my using an APML "attention profile" are several:
- Targeted information - by sharing my APML file I can have the sites and services that I visit suggest content that fits my personal tastes and interests, rather than serving me generic, one-size-fits-all information
- Painless advertising - as such, any advertising that I am subject to will also be tailored to things that I would find useful or interesting
- Control - by giving me control over my own attention profile, APML gives me the choice of who I wish to share my data with, rather than being subject to unscrupulous data-mining
While the standard is still very much in its infancy, it is already being adopted by a number of important web developers and content providers, and you can - even now - create your own APML profile using services such as Engagd and Cluztr.
APML is a further step in the direction of a smarter web, a web that bring you the content that you are interested in, rather than sending you off to wade through reams of useless information to find what you're looking for. Together with a growing number of other open standards APML is creating an easy to use, personalized web browsing experience and holds a great deal of promise for the future.
If you want to get in early on what could well be a standard as significant as RSS or OPML, you might well want to get yourself set up with an APML attention profile and check it out for yourself.
If you'd like to learn more about APML you might want to check out these links:
- The APML Workgroup website
- Basics of Attention Profiling Through APML - Marjolein Hoekstra's excellent primer on the subject
- Explaining APML - What it is, Why You Want it - Another great introduction by Elias Bizannes
- APML - The Next Big Thing or The Next FOAF? by Mark 'Rizzn' Hopkins at Mashable
- Attention - NewsGator and Bloglines Join APML Workgroup - Marshall Kirkpatrick breaks the news of some recent uptake of the format over at Read/WriteWeb
Programmer Links - thanks for pointing that out. Fixed now.
Chris - Many thanks for your kind words, I'm looking forward to seeing the standard grow and reach increased adoption
Matthew - I can see your point, but I wouldn't say that the danger is any less there than with mainstream media. For example, if I choose to read the Guardian over the Sun newspaper (excuse my UK-centrism) I have already made a political decision. Likewise I know that if I watch Fox network I'm going to get a particular bias in the information I'm fed.
At least with APML I can select my own focus, much as I can with RSS, rather than being served that of a particular corporate-political media conglomerate. The variety of scarcity-based media doesn't necessarily make it any less biased IMHO.
With information at saturation point I think that it IS possible to enter a solipsistic world trying to filter signal from noise, but surely that isn't so different from channel surfing on cable TV or selectively throwing away supplements and entire sections of irrelevant dead tree media? I guess finally the choice of "reading wide" or focusing on a very narrow information stream will always be that of the individual.
Surely there's a danger of these tools helping us to become more isolated from new ideas and opinions? Is anyone thinking of a 'Tolerance for New ideas' element here so we don't stagnate in pools of our own politics?
This is a landmark post Michael - it will be looked back on as one of the key signposts in APML's long journey.
Thank you for taking the time.
Thanks to everyone has has embraced our vision of standardized Attention Profiles.
Chris, APML Workgroup Chairperson
One of your links to APML points to http://APML.com and it should point to APML.org
Thanks a lot for the comment! I'll have a word with Robin, and I'm sure we'd both like to hear more about Lifestrea.ms.
Thanks, and apologies for the error - I think I have nailed the offending mistakes now. A case of me being a "klutzer" I'm afraid.
Great post Michael!
Just a small correction...
it's Cluztr not Clutzr and
the url is http://www.cluztr.com/
this is honestly the best article on APML and "the bigger picture" behind it that I've read so far!
Chris Saad gave me the hint to your post and I would like to add one more APML application to your list:
http://lifestrea.ms has launched tonight and supports not only APML, but also the complete list of OpenID, Yadis, RSS, OPML, XFN, hCard and OAuth coming soon.
If you're interested I would like to invite you for our private beta.