Tuesday, October 21, 2008

Twine went public

Finally, Twine went out of beta.

To introduce this service, in Thinking Space I had written two pieces (first impression and second impression) of post that were synchronized with Twine's steps of going public. I am interested in this service because it is a representative Semantic Web application until now. After talking and debating Semantic Web for so many years, Twine is among the very few successful (or at least workable) real-world Semantic Web applications. Just by this mean, it deserves our special attention.

For readers who are still not familiar to Twine, Paul Miller had a fairly comprehensive discussion of the service at ZDNet that is worth of reading. Also, you may directly go for Twine.com to try the service by yourself. In this post, however, I continue my style of service analysis following the first impression and second impression.

The up side of the public version

1) from producing knowledge networks to producing interest networks

I vote this update of mind to be positive. Knowledge network is a great concept. Producing knowledge networks is probably the ultimate goal of Semantic Web. However, this intent is too broad to manage. How to project this broad concept of knowledge network onto a narrower and easier-to-manage alternative is a challenge. I am glad that Nova seems figuring out one---interest network.

Personal interest is a subset of personal knowledge. Moreover, personal interest is probably the most heavily cared portion of personal knowledge. By reducing knowledge network to interest network, it decreases the amount of potential information processing and narrows the size of the application domain. Hence it could significant improve the performance of the service if the service designers truly understand the meaning of this change.

2) emphasized personalization and privacy

Nova has emphasized that "What Twine will be doing will be to make excellent PERSONALIZED AND TOTALLY PERSONAL AND PRIVATE recommendations to users." Moreover, Nova mentioned that Twine would monetize its service through the coupling of recommendation and privacy protection. I agree to Nova. This is the right direction.

3) enthusiasm of the users

Nova reported that at present Twine users spend 12 minutes per session on average. This is a very encouraging number. It means that the service is sticky in some sense, or at least among the ones who really use the service it is sticky.

4) good quality of content

I use Twine regularly. Based on my experience, the quality of bookmarking service Twine currently performs is fairly good. When I search information inside a twine, the quality of searched results is satisfactory in general.

The down side of the public version

1) move towards 2.0 rather than to 3.0

I was said that Twine was likely a Web 2.5 application in its beta. Then I expected that Twine would gradually move towards 3.0 when it went public. However, my prediction is wrong. By inspecting this public version Twine, it is more similar to the other Web 2.0 applications than becoming a revolutionary new service. In the other words, Twine is moving towards 2.0 rather than to 3.0.

At the beginning we might instantly tell the difference between Twine and another standard Web 2.0 service. But now it has been very hard to tell why Twine is not another Web 2.0 application. By just adding a few semantic computation is not enough for Twine to claim Web 3.0.

Twine needs to hire some visionary service manager in contrast to (or in addition to) the experienced service manager it has now. By experienced Twine could only go for what is popular now. Only by visionary Twine might lead the trend into the future. Twine is losing innovation.

2) too low traffic flow in monthly visits

Based on the information provided by the CrunchBase, the number of Twine.com's monthly unique visitors at the past September is only about 50K. Be note that Thinking Space has monthly unique visitors in the same September more than 5K. In this month, the unique visitors at Thinking Space is close to 10K. But Thinking Space has no invested money at all, with respect to Twine.com has $13 million only in its series B round of fund raising. Isn't it a serious problem of Twine?



Referenced resources:

7 comments:

Jennifer said...

Take a look at http://www.youbundle.com which is in a similar space as Twine, but doing things in a different way (and maybe different audience)

Lethe said...

I'm also a Twine user, and over all of the social bookmarking sites I am registered under (approx 12), Twine is the one I keep coming back to. Why is this?

I like the Twine Digest that comes to my email every morning with my favorite twines, but other services such as Mr. Wong and Faves has the same.

Perhaps it's the uniqueness and peculiarity of the community which lends itself to surpisingly similar interests, or interests that are close enough to mine I would be willing to try them out.

Oh, and I found this site through Twine. So it must be a good thing.

Yihong Ding said...

@Jennifer

thank you, I have looked at the site. It is quite interesting and apparently it looks like Twine too. But what is the underneath technology? Is YouBundle also produced based on the Semantic Web techniques?

Yihong Ding said...

@Lethe,

Sure, I am glad having connected with you through Twine. As you said, Twine is a great service thought it also has a few problems. But we still look for it being better and better in the future.

Reputationist said...

I appreciate your insights. I have questions on several levels from which I will evaluate Twine and it comparable sites. The sites I am using in this ongoing work are Swurl.com Popego.com Lifestream.com YouNoodle and to a lesser degree is Diigo.com.

On the deepest level - the question is about "trust" and "knowledge". As a anti-structuralist I tend to see the creation of discourses that valorize what some positions over others seem more statements of power not science.

Then there is the issue of "trust" which is in it own right a tangled issue. My simple point is more related to reputation - connecting with those less dangerous or risky in some form of invitation. The invitation simply means that what can be lost is of less value than what is being offered under the "to be trust" with.

Lastly, I suggest what will make Twine work or not is the value of the suggestions of people more that subjects and the daring people demonstrate with adopting those suggestions.

Yihong Ding said...

@reputationist,

I agree to you. "Trust" is a critical issue. Actually, it is not just critical for Twine; it is critical to all the present Web 2.0 applications.

The questions you have raised is not unique to Twine. It is a general issue in all the Web 2.0 applications. The present social networking services have provide too shallow the technology to support the distinction of varied levels of trustworthy. I believe that this is a central problem that must be solved so that we may approach to the next level of the Web.

However, how to solve the problem is non-trivial at all. We need innovation. And probably it will happen during this economic downturn. Let's be optimistic on it.

Cheers,

Yihong

jennifer said...

@Yihong

I have had some time to explore both YouBundle and Twine over the last week or so and come up with the conclusion that they started with a similar vein, but are reaching for different aspects of the market.

In a crude sense, Twine seems to be an evolution of the discussion group with the median being links. With Twines being created on subjects and people throwing everything they have into the pot

YouBundle seems to be more of a socialized personal publishing platform (blog) however the median being again Links. Allowing users to curate and present a portrait of a subject through their hand selected links. there is also a cool widget which can port the content to blogs.

As to whether it uses semantic technology, I am still trying to understand the real meaning of that word as it applies to the current trends.

But if we could define Semantic as analyzing the content of a page and a users actions and recommending content, then yes it is semantic.