Tommi's reports from Wonderland

January 07, 2008 Herding phenomenon in the web (or: pondering how to implement an effective feedback system for Beta Labs) Posted by Tommi at 11:48 AM | Categories: Internet stuff

As I watched this video clip of how Techmeme homepage changes over time, revolving around a limited number of hot topics, I couldn't help but wonder: isn't this a textbook example of herding phenomenon? (via Herd)

To give you some context, let me introduce you to the artificial life simulation called Boids, created by Craig Reynolds in 1986. In the simulation, boids are individual agents programmed to move around according to a set of simple rules. The fascinating part, as nowadays seen in movies and computer games, is the marvellously complex movement that emerges from the simple rules, almost as if the whole was a creature of its own.

Craig explains in his website:

The basic flocking model consists of three simple steering behaviors which describe how an individual boid maneuvers based on the positions and velocities its nearby flockmates:

1. Separation: steer to avoid crowding local flockmates
2. Alignment: steer towards the average heading of local flockmates
3. Cohesion: steer to move toward the average position of local flockmates

Hmm... Don't these points sound frightenly similar to how the blogosphere and social media works? Bloggers, including myself, while trying to be different from each other (= separation), are heavily influenced by their blog commenters, RSS subscription list, and aggregation services such as Techmeme (= alignment and cohesion with local flockmates).

Let's look what happens in practice. Here's an example Boids simulation that has added a simple goal-seeking rule, symbolized by the green dots:

Doesn't this look exactly like what keeps happening in social media, including the blogosphere and aggregation services like Techmeme or Digg? To me, it sure looks so. A "hot topic" gets born, like Scoble's episode with Facebook, and everyone and his uncle contribute to the debate. But also, a truckload of other equally important/interesting hot topics get ignored, due to the herd phenomenon.

But what if you don't like the herd thing? What if you want to avoid it?

Looking at the problem from another perspective, James Surowiecki convincingly argued in his book that for "wise crowds" to emerge, you need (1) diversity of opinion; (2) independence of members from one another; (3) decentralization; and (4) a good method for aggregating opinions.

Guess which one of these criteria is missing from the public debate in social media?

Independence. Bingo.

At this point, you might wonder why the heck I'm writing about this mumbo-jumbo here, in this blog. The thing is, we would want to take the Nokia Beta Labs feedback system to the next level, and I'm struggling with the problem how much of it we should make public, and how much private. Making the feedback system public inspires people to send insightful feedback, and to expand others people's ideas. But then again, it throws fuel on the herd phenomenon.

At the end of the day, I think we need to have both: a public feedback system (with all the latest web buzzwords) for debating the hot topics, and a private feedback system for collecting systematically feedback about the rest.

Any thoughts?

Permalink |

Comments

Wow,

Tommi that is a VERY insightful post. Lately in my own personal life I've been fighting the same sort of issue. I am in a system where compliance is the easiest way to get ahead, however it isn't always the best. This thinking outside of the box is why I have a lot of faith in Nokia.

As far as dealing with the issue of public versus private feedback... You might consider doing something like making all feedback private initially (esp. early on with a new product) and then maybe a week later suddenly it all goes public. This might allow some time for more independant thoughts to form. With the bloggers and forum-addicts, opinoins and/or reviews are often read before the product is even installed on users devices, so this is still far from perfect.

I'm very interested to see what sort of system you are able to come up with as this is hardly an easy task.

Cheers!
Malaeum

P.S. I was surprised to see not a single comment yet. Maybe the bloggers were a bit put off by your revelation... They are too busy covering CES to pay attention to this green dot. =)

Superb post! I love stuff like this, it really helps to make sense of the world, thanks for sharing!

I like Maleum's idea (ideas are private for 1 week then visible for all).

I also think that there is more diversity out there than we in the tech/telecom/internet world might initially think.

You say everyone in the blogosphere covered the scoble/facebook thing. But if we look outside of "our" blogosphere, we find plenty of bloggers who neither know nor care who Scoble is. Check out:
http://jessicamusic.blogspot.com/ a quick search of her blog shows *no* results for Scoble, none even for the term "Web 2.0" :-)

I think that in addition to asking "how do we gather feedback", we need to ask "whom should we ask for feedback". Marketing consultancies have understood this for years, and make a living off of creating diverse focus groups. We in S60 and at Nokia can adopt similar practices. We can start by asking people who are online but aren't on technorati. And then we can ask people for their opinions offline. S60 Ambassadors for example.

Any true feedback mechanism for the broader market has to include feedback from non-bloggers. That means getting outside our offices and talking to people. At trade shows, at Evenings with S60 (not everyone there is a blogger), at the local barbecue, in queue at the bank, wherever!

Great stuff, Tommi.

Tommi, this is extremely thought provoking. I'll have to think about it for quite a while...

I like Daniel's point about diversity though. Mommy bloggers or knitting bloggers don't give a rip about Scoble. And Scoble doesn't do a lot of knitting on his days off.

But I'm reasonably sure that we have a lot of herds, each created by the subculture(s) we belong to. For political bloggers, my blog with zero mention of the Iowa Caucus results would be unthinkable.

Yet, they are all spinning on their same rule sets driven by their internal communities.

That is really deep. Each meme takes on a different set of rules - a different set of communities to infect. It makes different groups turn and follow and bunch in different ways.

Now ... the ruleset is simple. But is the outcome predictable?

Great stuff Tommi! Posts like this are the jewels of your blog in my eyes.

I think Tommi You got really lucky this time. Malaeum provided a great answer to your question in the very first comment. I will also attempt to use that proposal in my work in some form. Thank You Malaeum!

However leaving my answer just like that, from my side would be the example of the hearding phenomenon in working so here is another proposal:

People who just want to agree with something mentioned earlier should propose something new also even if it sounds crazy to keep new issues rising and keeping the horizon wider. This carries the risk of popular ideas geting less popular due to some keeping silent not to be forced to propose something, but it might also lead to the "cutting up" of the heard...

So will I propose something extra also - more than just agreeing with Malaeum - to comply with my proposal?

Hups, I just did! :-)

Great stuff Tommi! Posts like this are the jewels of your blog in my eyes.

However leaving my answer just like that, from my side would be the example of the hearding phenomenon in working so here is another proposal:

So will I propose something extra also - more than just agreeing with Malaeum - to comply with my proposal?

Hups, I just did! :-)

Open it up, let people read what other people say, let people rate what other people have said. Remember the conversation we had a few months ago? Feedback inspires feedback. I love reading other people's thoughts because they might make me think about something I would have never otherwise considered.

Nice to hear that this thinking resonates with you :)

malaeum: good suggestion! However, I think you can't keep all feedback private from day-1. If I close the discussion threads here, they will emerge elsewhere. What we *could* do is making it mandatory for the application teams to write a summary of feedback the have received privately, say, from after the first week.

Daniel: yes, to get a balanced feedback, you need to do ask from "regular people" too (and Nokia sure does this). But this is outside of the scope of beta labs: we focus on the lead users here.

Jim & Daniel: good point. There are several distinct herds wandering around the blogosphere: some in techlology pastures, others in knitting fields.

Stefan: yes, I remember our conversation. But I think you missed my point. Beta Labs feedback system is not intended to entertain journalists & bloggers like you. It's intended for getting valuable feedback to Nokia developers. And I'm becoming increasingly convinced that a "herd feedback system" is not the optimal way to do it.

---

> the ruleset is simple. But is the outcome predictable?

I'm afraid not. Isn't this the main point of Chaos theory?

Beta Labs is focused on lead users? I'm confused on two counts. 1) It's not clear from the Betalabs home page that betalabs is only concerned with lead users. Yes, beta is for lead users, but the apps themselves look they they could be for the majority of people once they get out of beta. Apps like SportsTracker, for example, look like something a customer in the early majority might use.
2) If you are only looking for what the lead users want, then why try to avoid a herd mentality in the first place? best way to figure out what the herd wants is to follow where the herd is going. trying to parse results further than that will only fracture your efforts.

Interesting discussion and great post Tommi.

With regards to the herding phenomenon, the private and public system might be the easiest to set up but I'm not sure the private system will deliver "more accurate" feedback. After all the blogosphere is open to everyone and also the niche that sends private feedback is different from the niche that discuss in public.
By the way, when voting something or someone in a poll, the vote is done before watching the results. In a blog is quite the contrary, you read the comments first and then you comment further. Even if you were the only blog on earth, you would still have a flawed system for unbiased comments.


» Subscribe » Favorite Links » What is S60? » Freeware & Trials » S60 devices » Hints and tips » About this blog	» 3rd party S60 applications (78) » Devices (44) » General (137) » Internet stuff (13) » Nokia Beta Labs (38) » Nokia S60 applications (78)

Comments

Post a comment