JP Rangaswami’s thoughtful series of blog posts on the why and how of filtering online info-flows is a fundamental infotention text. Instead of Scooping all seven, I’ve Scooped this blog post by Jon Reed that summarizes and links to all seven parts.
diginomica, 7 February 2014
I liked JP Rangaswami‘s series on filtering so much, I decided to filter it.
The Chief Scientist at Salesforce.com, Rangaswami has a personal blog site, confused of calcutta: a blog about information, where he blogs on far-ranging enterprise topics on behalf of himself, not his employer.
The filtering series has been a very good read, but quickly became a monster series. The initial post laid out seven filtering principles; there are now five follow up posts to chew on.
Why do filters matter?
As a semi-professional enterprise news curator, I have a fascination with filters that results in obsessive tweaks to my Newsblur reader – always trying to make sure I am getting the feeds I need, versus the information overload I don’t.
Like Rangaswami, I’m very interested in pulling the best content and recommends from my friends social postings (though I’m more cynical about ‘friends as filters’ than he is). Most of my enterprise colleagues are in the same boat. There is a select slice of info and updates they need every day, but a heap of distractions they don’t have time for.
On the personal side, this boils down to a serious productivity challenge I have written about a lot. On the enterprise side, filters (of internal and external data) have big implications for enterprise decision making. Filters shape the informational context from which decisions are made. And, As Vijay Vijayasankar has pointed out, if you get the context wrong, bad decisions will follow.
As the filtering stakes high enough then? I’d say yes. Here’s my rundown of the series highlights.
Breaking down the Rangaswami filtering series
Gist: Rangaswami lays out direction for the series and defines all seven principles.
Quotage: Rangaswami makes his own case for why filters matter:
- soon, everything and everyone will be connected
- that includes people, devices, creatures, inanimate objects, even concepts (like a tweet or a theme)
- at the same time, the cost of sensors and actuators is dropping at least as fast as compute and storage
- so that means everything and everyone can now publish status and alerts of pretty much anything
- there’s the potential for a whole lotta publishing to happen
- which in turn means it’s firehose time
- so we need filters
- which is why the stream/filter/drain approach is becoming more common
myPOV: I like the separation of principal and secondary filters Rangaswami defines. The principle filter has do to with filtering content based on dynamic criteria, which could be in stages, such as a. filtering by demographic or geographic, b. filtered again by associated values (e.g. hotter than, older than, closer than), then c. filtered by inclusion or exclusion.
Once the information set is filtered, Rangaswami sees a need for a secondary filter that is about conditional filtering and routing (such as routing to the right form factor or device). The wonderful site IFTTT is one example of building or adapting secondary filtering based on routing parameters.
The network as filter?
Gist: Rangaswami is – get ready for this – fed up with email. Using the example of an email pile up after vacation, he argues that social networks are a superior filter to email. He also sees social networks as laying the groundwork for more collaborative and powerful filters.
Quotage: ‘Your friends will tell you what you missed, which conversations you need to be part of. What’s important. What’s urgent. What’s trending. What’s not.‘
myPOV: I’d have to do a full blog on this one, but this may be the area where I differ most with Rangaswami on filters. Email still has a core value I haven’t been able to shake (though I have streamlined it by rejecting newsletters and paring it down to essentials). With a virtual company like diginomica, we have not been able to find a better means of private virtual communication than email.
As much as I appreciate my friends, they are understandably preoccupied with everything from job changes to sick pets. And yes, posting baby pictures and foodie moments. They are a valuable network but I have not figured out how to filter their feeds based on the relevance of what is happening on their side versus my daily life as it constantly shifts, nor do I really want to try. They can do what they do.
Google search is (usually) more relevant in a given moment than my Facebook stream because I dictate my search priorities. I’m also wary that the main social networks are behind walled gardens of web sites I don’t fundamentally trust. But when Rangaswami speaks to a broader ability to navigate my world rather than daily information feeds, then yes, friends in a networked capacity are crucial and perhaps still untapped.
Thinking about routing
Gist: We need conditional routing, which means we needs information routing systems based on context. Wearable devices and smart appliances may need to ping us, but in appropriate ways with appropriate levels of urgency.
Quotage: ‘Context matters. We live in an age where we expect alerts to be delivered to us sensitive to the time and place of delivery, all the more because the cost of knowing where we are and what time it is there is trivial.’
myPOV: Complete agreement on both prioritization of routing and context. Otherwise pings will submerge us.
We are all publishers, publishers are all filters
Gist: We are all publishers. This comes with responsibilities. We should not censor at the publisher level, but there are still filtering responsibilities of digital publishing, including factual accuracy and context.
Quotage: ‘We should not design filters on the ‘publish’ side, we should only design them to be used ‘subscribe’ side. There are good reasons for this, mainly to do with avoiding ‘censorship by design’: we do not want to build structures that allow bad actors to dictate what everyone can see, read, hear.’
mPOV: I’m not as concerned with censorship by publishers as I am by state-controlled media and Internet blockages, but commerce can also act an unwanted filter. In theory, independent bloggers can serve as a corrective. As we all become publishers, we do need to understand what that means from an accuracy, trust, and spoil perspective (the temptation to share titillating backchannel info can be overwhelming and it so often backfires).
What does designing for serendipity mean?
Gist: One of the biggest dangers of filtering is that the results become insular. We need to design for serendipity to avoid the tunnel vision of our comfort zone.
Quotage: ‘It is important to ensure that the filters we set, as subscribers, are formed in such a way that the heretical consequences of extreme homogeneity are minimised.’
myPOV: One of the most important points in the series. I try to do this on a curation level by making sure that some of my secondary news streams have a deliberate level of randomness, which often leads to unexpected correlations and discoveries. Doing this on an enterprise level is much harder, but Rangaswami has ideas on how to go about it, starting with including externals on enterprise networks.
Filters should be dynamic and interchangeable
Gist: We’re not just searching static web pages anymore; we’re trying to filter and search firehoses of networked posts. Most of us have some mechanism for filtering this deluge; we benefit from ‘trying on’ someone else’s filters and seeing their informational plumbing.
Quotage: ‘A person’s collection of filters becomes some sort of toolchest, where the instruments that allow that person to do the job are kept and looked after. This is as true for knowledge workers as it was for artisans and craftsmen.’
myPOV: With today’s tools, it is hard to see someone else’s filters unless you swap passwords and log-ins. Trying on someone else’s filters like we share music playlists would be nifty. As Rangaswami says in his intro piece, he’d like to be able to consume news as if he was a 21 year old Iranian – not by reading the same web sites, but by exchanging filters. On an enterprise level, the implications are about creating filters that are less like templates and more like dynamic criteria – or, death to aggregates!
I hope you enjoyed this take on filtering. To be fair I only scratched the surface of Rangaswami’s ambitious series, but hopefully this gives you someplace to start – either by addressing context at the enterprise level or by fine-tuning your own filters.
I just talked to a friend who doesn’t really filter their content – they don’t have time. I’d say you don’t have time not to. Besides, we are always filtering. It comes down to how smart those filters are. Within reason, they are as smart as you teach them to be.