New Search Engine Attrakt Focuses On Curated Content
Ever amassed a collection of Web bookmarks on a particular subject and then wanted to search through their contents? Delicious users will be familiar with link curating and sharing, but there’s no ability to actually search the contents of the bookmarked pages and sites. A few Florentines, mostly former colleagues from the Italian internet company […]
Ever amassed a collection of Web bookmarks on a particular subject and then wanted to search through their contents? Delicious users will be familiar with link curating and sharing, but there’s no ability to actually search the contents of the bookmarked pages and sites.
A few Florentines, mostly former colleagues from the Italian internet company Dada, set out in April 2011 to develop a new search engine, Attrakt, which would focus specifically on hosting specialized search engines curated by the Web community.
Attrakt’s Index Relies On Curator Contributions
As such, Attrakt isn’t a generalist search engine, with a broad index ready for everyday searches. Attrakt’s index is primarily seeded through links specified by curators of custom search engines, what Attrakt calls boxes.
As in the early days of Wikipedia, if you’re lucky enough to search on a topic someone has already curated, then you might be impressed. But for the most part, Attrakt is fairly empty, a problem Attrakt partners Andrea Dotta, Luca Ciavarella and Gabriele Miceli acknowledged won’t be resolved until Attrakt reaches a critical mass of contributing editors.
Putting Attrakt To The Test With Schema.Org Rich Snippets Use Case
Users can explore Attrakt’s potential by searching or creating a customized search engine on a topic of particular interest; the Attrakt team mentioned travelling in Italy and 80s synthpop among many examples. These two illustrate the use multimedia in addition to textual resources. I decided to try a search marketing use case: the Schema.org semantic web markup standard defined and supported by Google, Bing, Yahoo and Yandex.
By implementing Schema.org standards, websites can enable the display of many types of enriched search results, what Google calls rich snippets. Enriched results can be very beneficial for Web marketers and users alike but there’s a lot of material to wade through.
There’s the official standard, supplemented with separate documentation from each search engine. Extensions are available for many CMS‘ including WordPress, Joomla, Drupal and IBM WebSphere. Google, Bing and Yandex each offer testing tools.
Armed with around 60 selected links to pages and sites regarding schema.org and rich snippets, I created a new Attrakt box, or specialized search engine. Searches using the resulting Schema.org Attrakt box provide answers from these curated resources.
Creating an Attrakt custom search engine is a fairly straight forward process. An editor needs to provide a name for their engine, assign a category and country and add a few tags.
Each link is then specified together with a few attributes such as a link category (e.g. Blog, News, Tool, Official) and whether Attrakt should index just the URL or the entire site. The link categories allow an end user to browse through the links assigned to each of these categories.
Overall, the creation interface could be simplified: entering 60 URLs, one by one, is a tedious process. It would be quicker to paste all the URLs into a single form, assigning categories in a successive step. The box country should be optional. The topic of many curated search engines, including this one, will transcend geography.
Link categorization is also problematic: should an official Google blog post on Schema.org be tagged as Blog, News or Official? While at first glance blog might seem to be the best answer, a Google Blog Post is really an official company statement which in of itself is often a news item, which happens to appear on a blog publishing platform. Clear?
Once a custom search engine box is created, it can be modified or deleted, although the use of grayed out text for the edit link both makes it hard to find and implies it isn’t available as a function.
To be fair, the Attrakt team is continuing to refine both the user interface and the underlying search engine technology. By the time you read this article, much of what you see will undoubtedly be improved. In the future, it will apparently be possible to assign a link to multiple categories. Attrakt is also working with researchers at the University of Florence on algorithmic solutions.
Once a box is created, it can be shared with others via a link. Attrakt will periodically crawl a box’s URLs to check for changes. Box curators will see a message display when Attrakt detects changes. It would be nice if they could also receive notifications via mail or RSS, similar to the way Google Alerts works. Search box curators will be able to embed their Attrakt search box on other sites, a feature planned for May 2012.
Are Curated Search Engines Really Useful?
A curated search engine potentially has two specific advantages over a generic search using Google, Bing or Yandex. The first is it will exclude content which strictly speaking answers the query but is of low value, like the plethora of parrot posts which appear just after a Google blog post is published. It often seems that 9 out of 10 repeat the official news without adding any additional insight.
A second benefit is to surface specific content which might not have otherwise showed up in a generic search. At the time of this writing, Bing’s Schema.org testing tool is poorly indexed in Google, but can be found browsing the Tools links of the Attrakt Schema.org search box.
Attrakt Wants Crowd Sourced Curated Search To Go Social
The quality of a custom search engine depends on the ability of the curator and on the technology providing the search results. Currently, Attrakt will display the user name of a box curator, but there’s no way for a casual user to determine if the curator is authoritative or not.
Attrakt currently collects profile information during the user registration process; it would be nice if they would also display profile information together with other box attributes. Such display would also give search box curators a greater incentive to contribute quality search boxes to Attrakt.
Attrakt highlights some hot search boxes on their homepage, but it isn’t possible to search for a search box on a topic nor is there a directory of search boxes.
In some cases, search boxes are listed next to results in a general Attrakt search, although there are some geolocalization issues still to work out. Attrakt primarily relies on a search box’s author to promote a specific curated search engine.
Attrakt search boxes are curated by a single user. It would be interesting to allow collaborative search box editing, similar to Wikipedia’s support for multiple authors. In a world filled with spammers, this isn’t so simple to successfully implement.
Attrakt plans on adding a robust social layer – a layer which will undoubtedly make Attrakt much more attractive to curators and search users alike.
On this point, Attrakt says:
Users will soon be able to follow each other and each others’ boxes, create elaborate user profiles to declare their areas of expertise. A Twitter stream will soon help animate search boxes with up to date content. Boxes already have static URLs for sharing, and sharing will be encouraged by share buttons.
Outside of the Attrakt platform, we’re planning on using twitter to drive traffic to peoples’ search boxes using relevant hashtags, and if a twitter account is present in the user’s profile we’ll tag that user.
Attrakt Is Still In Early Stage Of Development
A reader would be forgiven if they have the idea that Attrakt is very much a work in progress. The Attrakt team began armed with a dream and a modest €50,000 in self-financing. They’ve come a long way in a year and attracted an additional €500,000 in funding but are still months away from realizing the product they have in mind.
Mindful of this, they’ve eschewed seeking out press coverage, preferring to release and refine features out of the limelight. However, a site focusing on Italian news tipped me off to this new source of traffic, so now you know about Attrakt too.
Services Similar To Attrakt
Google’s customized search is probably Attrakt’s closest competitor. Many websites have added a basic version of it to supply a quick and easy site search, although it also includes features like Attrakt’s link categories, what Google calls refinements.
Perhaps the primary difference, beyond Google’s algorithmic prowess, is that Attrakt says its embedded search boxes will be advertising free (Attrakt does carry advertising on their site).
Attrakt also facilitates the sharing of curated custom search engines and plans on further differentiating themselves with additional social features. Readers can compare the search results for the same Schema.org resources indexed by Attrakt and Google.
Mahalo has positioned itself as a human edited search engine, but they don’t crowd source their contributions.
Delicious is by far the most used collaborative bookmarking service, however they don’t actually index the content of the saved bookmarks.
Volunia is a social search engine also from Italy, however their focus is more general.
Attrakt Opportunities For Marketers
Attrakt offers marketers the opportunity to demonstrate their personal or company expertise in a particular subject area through the creation and promotion of a curated search box.
As with all things social, the rules of the game have changed. The most successful search boxes will include comprehensive coverage of a subject domain, even if that means including resources from competitors.
What Webmasters Should Know About Attrakt
Attrakt robots.txt Support
Attrakt supports the robots exclusion protocol, known as robots.txt. Attrakt’s crawler is called attrakt. If Attrakt is crawling the site, web server logs will contain the user agent:
Mozilla/5.0 (compatible; attrakt/1.0 +http://www.attrakt.com)
Meta tags are not yet supported, nor is there support for sitemaps.org and schema.org protocols.
Keyword Tracking In Google Analytics And Similar Systems
Attrakt provides search query information in URL referrers using the defacto standard q=<query> URL parameter name/value pair.
Users of most digital media measurement systems will need to add recognition logic for Attrakt searches, otherwise Attrakt will show up as a simple site referer. Google Analytics asynchronous tracking code users can add
['_addOrganic','attrakt.com','q'] to their tracking code.
The story behind the Attrakt name? When asked, the Attrakt team demurred. Its origins seem to be lost in the sands of time.
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.