Radar’s Twine Emerges from Productive Beta

By SemanticWeb.com on October 21, 2008 2:48 PM

Jennifer Zaino
SemanticWeb.com Contributor

The much-discussed Twine, from Nova Spivack’s semantic web startup Radar Networks, comes out of beta today. At its formal 1.0 release, expect to see improvements mostly suggested by the user community during its seven-month beta stint. These are primarily usability-related features, including a:

Revamped interface
New, more intuitive navigation system

All-new interest feed

Public content within Twine is now indexed by search engines

Batch bookmark import from Delicious, Digg, and browsers

Improved recommendation engine

Ability to invite people to Twine from online email address books

New and improved semantic search.

On the way: a Digg-like rating system and the ability to automatically get the linked pages related to an article you bookmark. Another thing Twine learned from its beat users: “Be clear on one thing Twine is for,” says Spivack. The “interest networking” service does a lot of things, including letting users author content and create groups, but the tagline it’s going by now, Spivack says, is that, “Twine is a smarter way to keep track of all your interests,” letting users collect and share bookmarks, notes, video and other content, and using its semantic magic to learn more about the user as that user’s profile of interests expands.

According to Spivack, 500,000 unique visitors were invited to the beta. About 50,000 are now active, repeat users. Those 50,000 have created about 20,000 Twines — little sites about topics– and have contributed about 1 million different items to Twine so far.

“But I’m most proud that they spend about 6 minutes on average per session — that’s a very high level of user engagement. Six minutes gets us in the direction of some of the social networks that have high levels of engagement but the site is like a tracking and discovery site,” ala Digg, says Spivack. And sites like Dig and del.i.cious have smaller audiences than social networks and much lower levels of engagement. On the other hand, social networks don’t monetize as well because they are not focused around discovery but around communication.

So, why does this all matter to Twine?

“We are a site where people go to discover stuff, but the level of engagement is more like a social network. That combination should monetize well with our new form of direct marketing,” Spivack says.

In early 2009, Spivack, who says Twine has already signed up large national advertisers for testing, plans to deliver sponsored content — whether for products or events — right within interest feeds. Not in a sidebar, and not segregated at the top, where he contends they are easy to ignore. “That’s marketing, not advertising,” he says, noting that sponsored content will be labeled as such. “I think ads are a pull-based model where you may or may not see them. Marketing is more push-based, where it gets sent to you,” with the idea being that this is targeted content that a particular user will want to receive — and have control over whether or not to receive it.

If, for example, Twine figures out from certain interests and bookmarks that you are looking for a new car, Twine can recommend things related to the kind of new car you want, but once you buy a vehicle you can indicate that new car updates are no longer an interest for you.

Spivack notes, however, that Twine is being careful about privacy — advertisers who contract with Twine can target to particular types of communities and users in aggregate (like all users in the Bay area who are interested in new cars), but not individuals in particular.

“Everyone [in the advertising community] realizes there has to be something after display ads and email marketing,” says Spivack. “This is a new way to make money on the web and it’s more beneficial to parties doing the marketing and the people they’re marketing to. Ultimately you want to shift control back to the user, where it belongs, to safely track their interests without giving away their identity or email address.”

Also to look forward to in 2009: an API for enabling data integration with sites it’s not integrated with.

“Today Twine looks like a better bookmark or a better way of tracking, but as we open up the platform you will see you can build applications on Twine,” he says. That’s an evolutionary step for a web company, and one to be taken only after you’ve launched and found out who your users are and what they want; formally launching your product or service; and are making money off of it. “Then you can afford to be a platform; otherwise, there’s no point,” he says. Being a platform “helps you expand your franchise and open it up. But don’t open up your platform if you can’t afford to run it, because it’s expensive to do that.”

Comments