I’m sure there are better ways to decompose social networks, but in my simple mind, there are only a small number of fundamental components:
- directories and address books (you need to find the person or group you’re looking for)
- profiles and CVs and suchlike (there has to be some way of describing the person or group)
- communications infrastructure (you need ways to talk and listen and exchange messages)
- scheduling infrastructure (you need ways to agree times and places to meet)
- event notifiers (you need ways to spread news and gossip)
All these then get wrapped into a larger infrastructure, which covers four other things:
- an ability to identify oneself
- an ability to personalise the experience
- an ability to have Four Pillars support (search, syndication, fulfilment; we already have conversation)
- an ability for developers to add applications
In a way that’s what Office and Exchange and Outlook was about. In a way that’s what Bloomberg was about. And in a way that’s what Facebook and even Twitter are about.
This is an emergent and evolving space, but only in a narrow respect: communities have been around a very very long time; communications processes have also been around for a very long time. What has changed is the following:
- communications tools are becoming ubiquitous, especially with the web and mobile devices
- communications themselves now persisted digitally (allowing efficient archival, retrieval, search and syndication)
- the tools and the modes of communication have become more affordable
As a result, there’s a lot of learning to be had. Some people are concentrating on the interactions within the communities, the social graph as it were. Some are concentrating on movements: take a look at this article, published in the latest issue of Nature, and building on a theme established by a number of essays written by Barabasi and Gonzalez et al over the last couple of years); some continue to focus on the ownership, privacy and portability of the information. Yes, there’s a lot of learning to be had.
From an enterprise context, the learning takes on one further dimension. Enterprises have always been about walls and perimeters; now, as the walls become more and more porous, as the enterprises extend beyond their traditional boundaries to their customers and supply chains, communities become alliances, they become ecosystems, they become groupings of what Venkat calls “network-based competition”.
Yes, there’s a lot of learning to be hard.
Me, I’m utterly fascinated by one small piece of this overall puzzle. Alerts, event notifications, status messages, whatever you want to call them. Maybe it’s the old journalist in me. That’s why I loved the mini feed in Facebook. That’s why I loved Twitter.
And now, as I see more and more tools that help scrape information to do with events, I find myself going off at a tangent. Realising that we’re going to get overloaded by such messages (remember what happened when people started connecting Twitter to their Facebook status messages?); realising that current tools are already being stretched; and realising that the historical response (aggregation and summarising) is inappropriate.
I think we’re going to see an explosion of activity in the status message related tool space, with two different sets of tools. One to do with personal “manual” input, one to do with automated input. In both cases, I think we’re going to see this explosion connect with a similar set of explosions in the visualisation space, so that we see more colour, more heatmaps, more timelines, more fractal representations, more radar diagrams, more tag-cloud-like diagrams …… but all to do with status messages.
Status messages with a difference. Not aggregated, not summarised, but built around a capillary-action publish-subscribe model. Truly personalised.
That’s provisional enough. Now I wait for the comments so that I can learn more about this.
How about recommendation system to cope with status stream? Like book recommendations.
A Status message is just the tip of the iceberg of deep context. Like social graph, working out the context graph of a person must be possible.
A learning recommendation system that is aware of my the social graph and my context graph will help me cope up with all the status overload.
I immediately imagined a Google maps style mash-up with the individual’s time-line/diary at the centre.
You commit yourself to being present somewhere for certain chunks of that timeline (be it a crucial meeting, a party, travel or alone time) and prioritise it in terms of how sacrosanct, immovable it is.
When alternative alerts seek to gain your presence for those timeslots, they might become viewable only if the existing prioritisation level were set low enough. If not, then they’d bounce away, you’d never see them and wouldn’t know you were missing anything.
Additionally, the source of alerts could be rated either via automation (based on how many you’d accepted from that source in the past) and via your personal rating (think family messages, summons from the boss etc) so that the barriers to entry to your timeline that you have set could become more rigid for certain sources of messages and less o for more trusted sources.
Overlay on this a map application, so that any alerts that would be geographically irrelevant given your location for that timeslot would be rejected and I can see that alert overload could be quite easily managed though no doubt I’ve overlooked something.
JP:
Can’t help but say that for a contrarian view on mashing everything up, do see Penelope Trunk’s post:
http://blog.penelopetrunk.com/2008/06/04/twitter-social-media-and-unmashing-the-mashable/
I think the more fundamental question is to be resolved first in our own (i.e. users’) minds. That is what should drive technological innovation. Anything else sounds like the cart before the horse to me (and more consulting bucks, of course, sorting the mess so may be I should not look that gift horse in the mouth, mixing metaphors, or should that be ‘mashing’ metaphors, a bit…)
Shefaly, Balaji, John, thanks for your comments. You’ve made sure that a follow-up post is needed, both by your comments and by the references you make. Which makes this all worthwhile, because I learn by reading what you say and then trying to articulate what comes to mind as a result. Lakoff-Searls snowballs.
I disagree with your statement “Enterprises have always been about walls and perimeters”. Enterprises are about groups of people getting together to achieve things they cannot do as individuals. The walls and perimeters are an emergent property of enterprises, not part of their essence. Indeed, if they were, the enterprises would start to collapse when “the walls become porous”.
Anyway, to comment on your thoughts about alerts. I think we will increasingly see people switch off alerts altogether. I find that it is rare that the information in an alert is worth more than the time taken to read the alert (let alone the time involved in the context switch associated with the alert). There are two exceptions – an alert relevant to the immediate task at hand – for example an update on a flight departure when I am waiting for that flight, and a “Class A” alert something of immediate personal relevance (eg an accident to someone in the family).
So, rather than an explosion of activity in the status message related tool space, I predict the opposite. I think people will tire of the alerts and start switching them off altogether, or at least start cutting them down to a small number of interesting (to them) sources.
You may be right, Martin; or it may be that we mean different things when we use the word “alert”. One way or the other, you’re just in time to have added grist to my mill for the next post on the subject, thank you very much.