paying attention

Max made an interesting comment today, citing HP's Social Computing Lab's study, reinforcing his longstanding belief that attention is the currency of user contribution. It made me think about all the ways we pay-- and can be paid attention, and how we can convey it with varying degrees of depth.

We can visit, view, and look; or, we can interact, participate, engage... 

One of the most notable complements to any web technology is the ability to quantify it transparently. We have analytics for our every action, and as this study demonstrates, we're powerfully affected by our awareness of those metrics/ our performance. But what's interesting too is that most metrics consumed at face value lack depth (i.e. visits and views). 

What is the role of engagement in perceptions of attention? Does it matter? In the HP study, they used a Granger causality test to show that attention (as measured by YouTube views) CAUSED productivity (in the form of increased video uploads); and lack thereof actually led people to stop producing content. 

This is a pretty unbelievable testament to social perception. But more to the point, that level of behavioral change-- to produce more content or discontinue producing content-- was all a function of views. Not views and comments, dispersed views, return views, sentiment associated with views, influence of views, etc. Does engagement matter as much as we think?

As bloggers, which analytics are you most affected by? Mere quantity of visits? Furthermore how much does it affect you and your productivity?

sins of listening

I've always thought I had good hearing, but lately I've noticed (or have been told), my hearing is dwindling. FYI - my theory is I'm highly sensitive to noise (aka a superhearer) and thus have a lot of interference when it comes to discerning the signal. So naturally, I compensate for it by pretending I hear, doing some fuzzy (desperate) interpretation to make sense out of a blur of words and tone, and impulsively respond. I'm usually pretty off-the-mark...

I come from a brand monitoring world, a space where the potential for many variations of the above can easily happen. Importantly, I'm speaking from an analyst-as-consumer-of-data perspective-- could be a marketer, researcher, client, whomever it is attempting to listen to the data. Point being: ineffective listening is a sin. You should always query and re-query-- as many times as necessary-- so that the data can fully express itself. Data is, in some ways, more responsive than other conversational partners. 

As with consumer generated media, when the data universe is vast, participation is unregulated, and the best questions answered are unanticipated, the most frequent sins are probably acts of omission rather than commission. That is, it's not that you go wrong by acting on the data, but in not benefitting from what you could know if you had more effectively listened. 

I think clever interrogation is the formula for effective listening: a loop of listening, structuring, iterating and eventually analyzing. Which other effective strategies have you developed to better listen?

the new beast of editing

Yesterday I was obsessed with the idea of editing. Who are the new editors with democratized publication? What is their role? What are the varied ways editing manifests online, beyond the literal editing of a Wikipedia entry? A colleague offered up the idea of Twittering links to popular news as an example, be it an annoying one. 

I was thinking of it more in terms of collaborative filtering as a new mode of editing, a transformed mode. 

Is editing -- in this evolved state-- how the wisdom of the crowds prevails over the known dysfunction of groups (think: groupthink, loafing, etc.)? Is this how we achieve quality outcomes from crowdsourcing? If so, is the editor's need to 'get things right' self-serving? Is it even a need to 'get things right' or is more about filtering and mastery-- making sense of the world? 

I was ready to let it go, but then my mother called my attention to Tina Brown's newly launched Daily Beast a pastiche of traditional journalism and modern day online dialogue. David Carr's piece, aptly titled Editor of Note, Perched Online, says this:

With a slogan splashed across its home page promising rigorous editing of the culture for complicated times — “Read This Skip That” — the Beast is aiming to be a smaller, less chaotic version of the World Wide Web itself.

This style of 'rigorous editing' doesn't really seem to embrace the complicated times we are in (Note: purely technologically speaking). It seems like editing, online. I'd like to see Tina Brown achieve an evolved state of editing. But I think it will require more than 'sensibility'...

chasing engagement

Someone accused me of being like the Music Man when I suggested there are meaningful metrics that tap into the value in a conversation. Sadly, there is a lot of black magic out there in the form of widely variant algorithms for things like influence, engagement, and even less buzzy constructs like sentiment. But the constructs we're chasing exist; it's the methods that are misaligned. You don't necessarily need to add, or even use every variable available simply because you can.

People have deep interactions and make strong emotional connections. 

How do these manifest? Using a simple text analysis program, I did an experiment comparing 500 blogs in which people make a "definite recommendation" to a random sample of blogs (with a similar topical focus). Why? Recommendations have been hailed as the holy grail of word of mouth and even organizational performance (c.f. Net Promoter score).

If you look closely at the language people use around "definite recommendations" you see they are far more emotional, and as you might expect, more positive than negative. Interesting ecological validity there...

More interesting to me, recommendations are more intimate. People used more personal pronouns (I, me, my) indicating that they felt connected to the recommended items ("products") they were evangelizing.

And here's the clincher, they use more verbs than nouns: recommendations included more discussion of experiences with the products than discussion of objective attributes. Recommendations are experienced-based, not removed evaluations.

I don't interpret this as the precise litmus test that will give you the PH of a conversation, but it's a good example of the ability to tap into something valuable, relatively easily. Point being, I think something deeper usually lies beneath: subtle ways that people express deep emotional connections to brands, products, people, and the world in general. 

body choir: unexpected lessons for a healthy community

I was watching an old friend play some music last night and witnessed something that might offer some unexpected clues about what "health" could look like in an online community.

Scattered throughout the audience were friends of the musician who are also members of body choir. A friend told me that there's no talking allowed at body choir-- you speak, or rather sing with your body. This results in spontaneous, yet very fluid movements that you might have seen back in the days of EST groups. Eventually, they're dancing, completely in synch.

It used to amuse me, but last night I decided I was impressed by the symbiosis they achieve. 

There's this interesting process that sets off as various members start to internalize the music, make eye-contact with one another, exchange cult-like smiles, and eventually begin harmonizing their movements, seemingly rolling off one another. You can almost see a meme disseminate across the audience. 

How could an online community achieve this symbiosis-- each member internalizing the community's core values, generating that level of enthusiasm, and harnessing it into synchronized communications?

Can we study groups like this to derive an algorithm for the health of a community? A certain level of information, a vehicle to facilitate the transmission of that information, an optimal number of participants?

imposing structure; discovering structure

I was reading the Wikipedia entry on enterprise social software and was taken by an idea, which I think is actually a philosophy- my philosophy.  

"In contrast to traditional enterprise software, which imposes structure prior to use, this generation of software tends to encourage use prior to providing structure."

Use prior to structure... 

People always tell you this when you buy a house-- before putting in furniture, use it, walk around, figure out the traffic patterns, see which areas you and others naturally gravitate toward. Then, structure accordingly - make the furniture fit your lifestyle. 

This is my precise philosophy on how to approach new research constructs (e.g. Influence, Engagement, Authenticity), and data analysis in general. Map it out. Use your data and then divine structure. Down with Field of Dreams, 'If they come, it will build (grow)' seems a bit more apt today.

Surely I'm naive. When is it good to impose structure prior to use? People tell me schedules and plans work well in life. Is there anything in the enterprise that clearly calls for structure prior to use or can everything evolve to a place where structure emerges organically?

wwpjd? analyzing data, philip johnson style

I've been thinking a lot about what it would mean to analyze data from a glass house... All of a sudden reminds me of the time when Nick Denton opened up his blogging storefront on Crosby Street. 

Who can't relate to the situation when immersed in data, you yearn to channel various statisticians, or even your grad school advisor to figure out how best to proceed? Alternatively, have you ever been reviewing someone else's data, wanting desperately to point out something you think is interesting-- that you feel they should know, given your own forays in different data sets?

Several companies are getting into this idea of collaborative data analysis to help efficiently communicate important things as the amount of information proliferates. Early on, I reviewed one such impressive offering, Swivel.com. I was immediately taken with the idea of a community designed around data analysis-- not only the ability to annotate analyses, but access to wealths of metadata on who made the annotation, what else they've commented on, their background, expertise, etc. 

Today, Matt Hurst, a former colleague of mine blogged about DataDepot from Microsoft Research. Strangely, I can't access it anymore, but earlier I noticed there's a lot of room for engagement with the data here... commenting, annotating, easily adding new datasets. This could really transform the way we happen upon empirical insights and highlight complex relationships; and, in the end, happen upon emergent outcomes. No more ivory towers!

What, beyond privacy concerns, are the repercussions of analyzing data in a glass house? Seems like it could really motivate researchers to be productive...