Today, we welcome another guest author to Just Behave. This week, we’re pleased to welcome Sep Kamvar, the lead software engineer of personalization at Google. In the first of two parts, Sep looks at how the introduction of iGoogle and gadgets have created a new ecosystem of push content and how Google tackles the problem of organizing this new community generated content for their users. – Gord Hotchkiss
When the personalization effort at Google started in the fall of 2003, Google’s search engine was very much a one-size-fits-all product. Over the course of two articles, I will talk about two key advances in personalization that we’ve achieved over these past four years: the personalization of our homepage through iGoogle and the personalization of search. This first article will focus on iGoogle.
When we first started working on iGoogle, our goal was to build a product that was the answer to the query “What do I want to see?” We formalized this mission as follows: Organize the world’s push content in a fast, simple, and useful manner. The key specifying adjective here is “push” – while Google has historically been in the business of providing users with information when they provide a query, iGoogle is one of our efforts to give users the information they need without a query.
We were aware from the start that we would not be able to create all the world’s push content. Our strengths are in organizing content. So we launched the Google Gadgets API with the personalized homepage to allow third-party developers to create compelling content. As that content is created, we do what we do best, which is organize it and make it accessible to our users.
This has been a win both for our developers and for our users. Gadgets from our top developers get tens of millions of page views per week, and our users can choose from more than 10,000 gadgets in the directory. Indeed, we have entered into a virtuous cycle, where increased gadget development drives increased usage of iGoogle (since users have more compelling content to choose from), and increased usage of iGoogle drives increased gadget development (since, for our developers, iGoogle becomes a stronger distribution platform that reaches more users).
The quantity of gadgets that have been developed raises the question: how do we organize this information in a way that’s useful for our users? Our primary strength is in algorithms, so this is where we have focused our early efforts. Right now, much of what we do is algorithmic: choosing default content, recommending gadgets on setup, and ranking gadgets in the directory are just some examples of problems that we solve using algorithms. My favorite example is Magic Tabs – if, when you add a new tab to your iGoogle homepage, you name the tab, e.g. travel, chess, astronomy, games, and keep the “I’m feeling lucky” button checked, the tab will populate with gadgets that other people have added to their tab with the same name. Features like this help our users set up their pages and discover new content, and we will continue to invest in algorithms to help organize gadgets for our users.
Social mechanisms are a second important means of push content organization and discovery. There are currently three primary means of social discovery of content for iGoogle. Users can share a gadget with the ‘share-a-gadget’ feature. They can share entire pages with the ‘share-a-tab’ feature. Web publishers can also add a gadget to any page on the web, an advantage to them as well as to users who can take an interesting gadget from anywhere on the web and add it directly to their iGoogle homepage. And iGoogle users can easily build shared gadgets through our Gadget Maker templates.
A third means of organizing push content is editorially. YouTube’s decision to hand-select daily featured videos is a good example of effective editorial content. I see an opportunity for third parties to play an interesting and important role here. Third-party editorialized gadget and tab directories along the lines of LifeHacker’s Show Us Your iGoogle and Google Modules, play an important role in the ecosystem, as users often discover useful content through the staff selections of these third-party sites. This is an area that I’d like to see grow. Furthermore, we have started to see Search Engine Marketers see their job not only in terms of driving visits to their client’s web page, but also driving visits to their client’s gadgets by marketing gadgets to the appropriate communities with features like the Add-to-iGoogle button or with AdWords campaigns for their gadgets. This is another editorial format that I expect to see continue.
The evolution of this gadget ecosystem reminds me of the early evolution of the web itself. Many of our gadget developers, like many early web developers, built gadgets primarily for fun, and the hobbyist developer community remains vital to the success of the gadgets platform. As iGoogle grew, a second motivation for gadget development arose: impact. Building a gadget has become a way to reach millions of users on a daily basis. We now see companies building gadgets for the same reasons that companies started building websites in the early days of the web – to reach new and existing customers.
And recently, we have seen a gadget economy arise from this user-developer-organizer ecosystem. Several companies have been built around gadgets, from gadget design and development firms like LabPixies to enterprise software companies like Appirio. We have seen the first few gadget acquisitions, where developers have sold their gadgets to existing companies. And, earlier this year, we launched Google Gadget Ventures, a pilot program that gives no-strings-attached grants to select gadget developers who have shown success to further develop their gadget, and gives select gadget entrepreneurs $100,000 seed investments to develop financially self-sustaining gadget businesses.
As with the early web, I expect to see continued growth in the users, developers, and organizers that make up the iGoogle ecosystem. I also expect to see the development of an industry to service these developers, from design firms to developer tools to venture funds. And like with the web, I expect that the people who will benefit most from these developments will ultimately be the users.
Opinions expressed in the article are those of the guest author and not necessarily Search Engine Land.