A few months ago, anyone launching a new search engine would have been reasonably advised to have their head checked. Of course, that all changed with the PRISM scandal and the uncertain level of involvement by web giants such as Google and Microsoft. So far, private browser DuckDuckGo has been the most visible beneficiary of this shaken-up search landscape, but these are early days yet.
Hence the decision of the team behind Archify to launch a new crowdsourced public search engine called Blippex. It’s not a pivot – Archify, a personal search engine that archives the users’ browsing, is still going – but rather an experimental repurposing of some of Archify’s core technology. For example, Blippex’s ranking is largely based on the same DwellRank technology, used to order results by time spent on pages and the level of user interaction, that underpins Archify.
“It was always our plan from the beginning that we wanted to make a public search engine, and we thought now was maybe a good time,” co-founder and CEO Max Kossatz told me. “Because we have the infrastructure already for Archify, it took us two weeks to make it.”
More private, more transparent
Google’s PageRank algorithm is rather opaque (hence the company’s EU antitrust woes) but it generally leans more heavily on incoming links as an indicator of a webpage’s importance than it does on user engagement. Blippex, which relies on browser plugins in order to track how much time people spend on pages, only has around 2 million pages indexed at the moment, but according to Kossatz its results are “already quite different from Google”.
“We thought we’d try it out and see what happens,” he said. “Until now, every search engine which is new tries to get the same results as Google does – we’re trying something different to see if it gives you different websites.”
Of course, getting people to install the browser extension is the key, and to that end Kossatz and his co-founder and CTO Gerald Bäck are shouting loudly about Blippex’s privacy aspect. All the plugin sends them is the page’s URL, the current time and the amount of time spent on the page – as with DuckDuckGo, no IP addresses or other identifying information are recorded, and neither are search terms.
The team is also pushing transparency as a big selling point. They say they intend to open up their search algorithm data, starting with the inclusion of their scoring details in a public API (the same one Blippex’s official front-end uses). The extensions are open-sourced on Github, and Blippex will also publish a dump of its database on a monthly basis.
What’s more, users get a degree of control over the weighting of the search algorithm, with a sliding scale that lets them choose how much influence DwellRank should have in the rankings, versus search term frequency.
So, what’s the business model? “We don’t have one yet,” said Kossatz.
To a certain extent, this worries me – any time there’s a new service that touches on privacy and there’s no stated business model, you have to wonder what the adoption of a money-making aspect down the line will mean. After all, even Google would probably be more privacy-focused if it hadn’t built all its services around targeted advertising.
That said, there’s not really any private data involved here, so I’m not overly concerned about this factor. Ditto the fact that (like DuckDuckGo) Blippex uses infrastructure-as-a-service from U.S. provider Amazon.
Blippex is available now as an extension for all major browsers barring Internet Explorer, and also as an Android app. Bearing in mind it’s an experiment of sorts, I’m keen to see how it pans out in terms of both the privacy aspect (see also DuckDuckGo and the Dutch outfit Ixquick, which has been around for over a decade) and the engagement-centric ranking method. Search is so dominated by Google in most countries that the scene can really use credible new players.