echo_echo

I actually got sidetracked a bit at the end of my previous post, which is perhaps why I’ve had such trouble writing the continuation. What I mean to examine isn’t the community that develops around a given website, but the level of individual engagement (feedback) that the medium offers. But let me explain why I have an interest in this, before I go further.

Each day over the last few months, I find I have more and more mental acuity at my disposal. I keep having ideas, everywhere, all the time. I used to struggle for material to occupy my mind with, but I now find that half as many hours exist in my day as would be necessary to properly investigate all the concepts my mind is chasing after. Maybe I was always having lots of ideas and I just never took them seriously – I really haven’t a clue. Point is, I’ve been on a roll and it doesn’t appear to be slowing down. That’s great – so long as I’m headed in the right direction. Which is where it would be nice to have feedback. Facebook, G+, Twitter – these things may generate plenty of viewership for me, but none of that comes with feedback.

I’ve been pondering why, exactly, that is. Part of it, I believe, is that they’re dense sources of information. You can scroll down a handful of pages and encounter a hundred links to different places all across the internet. The more I thought about this, the more certain I felt that it was time to make a graph. Tremble and despair.

There’s a lot of things worth disagreeing with in this graph. I’m aware of its imperfections, but it’s an idea I’m driven to explore. I’ve been trying to write this damn post for over five hours now, because this has been driving me nuts all week. There’s something here, but finding how to express it has been a major challenge. I don’t even like either of the terms I’m using here. It’s all I’ve got.

There are places that efficiently distribute content. There are places that enable effective user-to-user or group-to-group communication. There aren’t many that do both.

In my search for feedback, I realized that I would never receive the kind of analysis and criticism I want through any of the mediums I’m currently engaging in. I would need to find a community of individuals that were capable of providing the feedback I want. It’s hard to search for something when you aren’t sure what it looks like or where it might be. That search goes on – but in the meantime, I remain interested in this relationship between content and feedback. Specifically, I want to know what something in the top-right corner of that graph would look like. I hope to return to this topic a number of times, especially once I’ve reached some more solid conclusions.

14 thoughts on “echo_echo”

  1. [Ah, a plot! Talk about a fly to honey!]

    Information density is always low on the web, that's why there are search engines.

    I think this gets back to signal theory: any increase in the signal (the feedback metric in your post and the axis in the plot) is typically accompanied by an increase in noise. Noise lowers the information density; so you could flip the abscissa of your plot and call it noise or signal-to-noise ratio (SNR).

    It seems to me you are looking for the equivalent of a search engine to filter through your feedback. That way you can cast your net as wide as possible, get as much total response (signal i.e. usable feedback plus noise i.e. trolling) as possible, and use the filter to get rid of the noise.

    For example, we do this with spam filters for email. But the “spam feedback” filter you are looking for is much more complex. For example, if there is a troll who only reads your writing to, well, be a troll, then you can easily block that. But there is a line between that and blocking people with whom you may sometimes disagree but to whom it is nonetheless worthwhile to listen to improve your narrative/case/argument/point/&c. Or maybe it's worth reading even the trolls from time to time, for entertainment or whatever.

  2. for example, look at this problem statement (it may require a registration):

    http://community.topcoder.com/longcontest/?module=ViewProblemStatement&compid=23864&rd=15034

    in summary, a client of TopCoder.com put up at least $10k for a bunch of coders to find an algorithm to improve their ability to distinguish between real DNA reads (signal) and fake DNA reads (noise) from ~92% on some metric to 95% or better.

    What's interesting is that the match to solve that problem ended yesterday and the highest-scoring solutions were in the 99.99% range, but they exploited a weakness in the test metric that it appears will make them not very useful to the client. If that is indeed the case, the their metric was no good and/or it needed a spam filter.

  3. Btw, in one sense I could claim that you have some nerve claiming you want to move to the upper-right of the plot. Why? Are you that special? Is your time more intrinsically valuable than the trolls'?

    (;-)

  4. Okay, I admit it, I'm a troll today.

    But I'm writing this in an airport, using “free” wifi, and you're whining about increasing your SNR instead of expressing wonderment at the capabilities that exist? “How do you know 'e's a king? 'e's not all covered in …”

    “This internet thing is wonderful. If only we could harness its power for good!”

    We often take what we have for granted and then lament that it's not X% better.

  5. I agree that this has something to do with signal-to-noise ratio, but I don't think that completely describes the problem at hand. When I engage in a 1-on-1 chat conversation, there isn't a lot of noise going on (unless you count small talk). There just isn't that much pure information exchange going on. I do, however, get a direct response to nearly everything I type. That's also why I put MMOs up there, as well. Even if you ignore the fact that video games are interfaces that provide a constant stream of real-time feedback, there's still a pervasive social element, wherein you can expect other players to very frequently respond to your actions.

    By contrast, if I post something to Facebook, the most I can reasonably expect is one or two likes (or +1s in the case of Google) and, if I'm lucky, a meager handful of one-line responses. Compared to what I might get in the previously mentioned mediums, these are useless towards the end of creating better content. Really, all these responses indicate is who is a part of the demographic that has an interest in what I'm creating. However, if I want people to view my content (which I do), Facebook is hands-down the most efficient vector for viewership. Achieving the same results via instant messaging or even a forum would be impossible.

    The question is, is it possible for a medium to achieve a high level of individual interaction while also efficiently dispersing large quantities of information? If so, such a place would be a major boon to my desire to enhance my creative ventures.

    Don't get me wrong; the internet is amazing as it stands. I just think its potential can go far beyond what we currently have, even without technological advancement. I think it's a challenge of creating effective incentivization systems for participation while making the process of distributing information as fluid as possible, and then allowing for adaptive categorization to facilitate the consumption of that information.

  6. How did you create the graph? Not to hate or anything, but you didn't specify where your data came from despite presenting it quantitatively. If you actually measure these things, this could be very interesting.

  7. That data are entirely made up. You'll notice there are no numbers nor even a scale indicated.

    My goal is to find a way to measure these qualities. I'll probably be making it my term project for my stats class.

  8. “The question is, is it possible for a medium to achieve a high level of individual interaction while also efficiently dispersing large quantities of information,” [to ]?

    You mean, the question is, could we pick a day (or make every day a day) where everyone listens, and responds to, . The optimum utilization of all people's available time would be achieved if everybody focused on and thought about what has said or written.

    “If so, such a place would be a major boon to desire to enhance creative ventures.”

    Here I am using the word to represent a completely generic hypothetical human being with internet access.

  9. “The question is, is it possible for a medium to achieve a high level of individual interaction while also efficiently dispersing large quantities of information,” [to (TIM)]?

    You mean, the question is, could we pick a day (or make every day a day) where everyone listens, and responds to, (TIM). The optimum utilization of all people's available time would be achieved if everybody focused on (TIM) and thought about what (TIM) has said or written.

    “If so, such a place would be a major boon to (TIM's) desire to enhance (TIM's) creative ventures.”

    Here I am using the word (TIM) to represent a completely generic hypothetical human being with internet access.

    P.S. Apologies. Got hit by the <> symbols the first time.

  10. “I think it's a challenge of creating effective incentivization systems for participation while making the process of distributing information as fluid as possible, and then allowing for adaptive categorization to facilitate the consumption of that information.”

    The numbers cannot support this for everybody. The best you can hope for is a ponzi scheme with (some random;-) content at the top.

    Isn't this similar to the difficulty of art? You generate art (creative content, if you prefer), and traditionally artists starve or thrive as a function of the disposable resources of their community. Find a patron, build a following, get a showing in a gallery, find some people willing to tpay for your work. By analogy the “pay” you are looking for from “some people” is their time invested in *usefully* interacting with your content. People just aren't that accommodating.

    I still think signal to noise adequately describes the problem. As you said, in something like MMO there is a lot of feedback, just not much content: you boosted the gain by choosing MMO, but you got a lot of noise. That is, the more people you engage with, the smaller the percentage who are going to be interested in (i.e. follow) your agenda. E.g. I'm wasting your time and fouling your blogspace obsessing on SNR.

    By definition only a few people make it to the top of any social pyramid.

    A few years ago, a celebrity had a rare cancer. CNN trotted out Dr. Sanjay Gupta for a few minutes. He started with something like “I never had a patient with this cancer, it wasn't covered in medical school, and there is very little in the literature about it.” He then proceed to talk about it. I don't know how that fits here but I like the irony.

    Think about columnists, I mean “real” ones i.e. syndicated. They have a platform to distribute and that platform lends some (sense of) authority to their writing (more if it's the NYTimes, less if it's the IJ). But that imputed authority draws people and induces some to respond with more signal, and other responders are just crackpots with an axe to grind. Compare that to blogs. Who blogs with the authority of a syndicated columnist? Very few. And the largest class of responses is “you rock, man” because birds of a feather … After that there is a spectrum of bloggers who have decreasing numbers of followers, down to groups of friends who use it as an alternative to fb, and after that it's fb and other social media for people that don't want the hassle of maintaining a blog with all the accouterments (background, font, theme, regular schedule, &c).

    In the end I think your plot draws a nice line with a negative slope that characterizes the population of available (and probably possible) interactions for feedback and content density. To expect something off that line is just wishful thinking.

    Peace.

  11. I don't agree that this is a fundamentally narcissistic pursuit. Yes, it's true that I've encountered a problem and now seek a solution to address that problem – but I'm not seeking a personal solution. If I just wanted to do that, I would do what every artist and entrepreneur has to do to acquire attention, feedback, and ultimately success: endless self-promotion. To my ears, that's just another word for spam, and that's precisely what we would see less of if creations could be appropriately matched with constructive feedback. If we really want to see a higher ratio of quality content on the internet, self-promotion shouldn't be the primary vector of transmission. Ideally, content should be promoted by those who do not have incentive to promote that content.

    The system I imagine would avoid this through robust categorization, kind of like what peer review aims for. Content generators are exposed only to those that have at least minimal interest in that content. The system can incentivize input from sources of expertise by rewarding constructive feedback, and quality contributions would be rewarded with increased exposure. These are the basics of existing forums and social networks – it is no stretch of the imagination to suggest that a system could be devised that further motivates critics and creators to collaborate.

  12. I don't think my previous description sufficiently addressed the “numbers” problem, so here's a bit more detail.

    It's true that not everyone can have high exposure – the flat numbers on what gets viewed how many times will pretty much always end up being structured as a pyramid (AFIAK). I don't believe the same rule holds for feedback. Quantity of feedback is a function of the response the content generates within the viewer. Easiest example: controversial content always generates a higher proportion feedback to viewership than others. People go way out of their way to express their opinion when it's something they feel strongly about, and this is completely unincentivized.

    This suggests to me that if you can show the right people the right content and add even just a minor incentive, more feedback with a stronger SNR can be captured.

Comments are closed.