Google’s new social network debuted this week with several new concepts and communications tools, including a potential game-changer for online news called “sparks.”
Sparks are topics that a Google+ user designates an interest in. Google uses sharing activity, +1 recommendations and search algorithms to offer personalized content for each spark.
It’s quite different than anything Facebook and Twitter have offered. Sparks don’t just tell you what your friends have read, they tell you what you ought to read. It’s a serendipity engine, and if executed well it could make Google+ an addictive source of news discovery.
But right now, it’s a great idea with imperfect execution. That’s to be expected from a project only a few days into a trial phase and nowhere near finished. Google’s goal with the early rollout is to learn what needs improvement.
After a couple days using the new service, I believe there are three missing pieces preventing Google+ from becoming an influential driver of news consumption. All three are within its reach.
Following and interacting with institutions
Media organizations and other institutions or brands need to be represented. They have Twitter accounts and Facebook pages, but no place to interact on Google+.
I can’t fault Google for not having this yet; it’s most important to get the people-to-people networking down first. But at some point Google+ must think of a way for news providers and companies to interact.
They could create Google+ profiles as if they were people, which was the common workaround on Facebook in the early years before fan pages were developed. Some already are doing that. But Google should make clear how non-persons, such as news organizations or other companies, can join.
Forty percent of Facebook users and 25 percent of Twitter users have liked or followed a “brand” account. It’s a significant part of social networking activity that is not yet in Google+.
Filtering ‘sparks’ news by sources
Right now, a spark is treated essentially as a keyword. It brings in content from the entire Web that appears to be related to the keyword. For example, I set up a spark for “Poynter Institute” and I get top results from USA Today, NPR and SF Weekly — all articles about Poynter, not from Poynter.org.
If I want to use Google+ to keep up with articles from Poynter, The Huffington Post or any other particular site, I’m out of luck.
One possible solution, first suggested by Jennifer 8 Lee, is to enable some sparks to represent content sources instead of topics. I could have an NFL spark for all news about professional football, or an NFL.com spark intended to bring me news from the league’s site.
Even that, however, may not be enough when you consider a site as big as The New York Times. I don’t want to (nor could I) read everything the Times publishes. There should be some way to specify topics of interest within the sources of interest.
I see two ways Google+ could attempt this. One is just to create even more source-specific sparks — nytimes.com for business, another for national, technology, and so on. This model begins to replicate an RSS reader.
That works, but it gets messy and fragmented. What if I want to follow a broad topical spark for business news as well as source-specific ones for several business news sites? Now my business news is in a handful of different places.
The other approach is to keep sparks as general topics and create another layer of settings that enables me to enter preferred news sources. So my “business” spark could still bring me all the top business news, but it would place extra emphasis on the stories from my favorite sites. Maybe once news organizations have an official presence, Google can enhance sparks by favoring the news sources a user has added as a friend.
That’s the approach that seems the most reasonable and flexible. It would serve power users without forcing any complexity on new users.
Improved semantic recognition of sparks phrases
Google+ sparks work great for following news about things with unique names and labels — such as many brands, sports teams and products. It uses Google’s comprehensive picture of the Web and its assessment of sites’ authority to find relevant mentions of those keywords.
But sparks are not so good at words that mean different things in different contexts. For example, I followed a spark for “running” (the fitness activity) and mostly got results about running backs, candidates running for president and gadgets running operating systems. A spark for “watches” (the thing you use to tell time) brings me some relevant results, but also a story headlined “Turkey watches as Syria crisis continues” and another about thunderstorm watches.
If sparks are going to become a major hub of news discovery, Google will have to develop better technology that not only matches keywords but understands the correct context.
Personalization powered by the social interactions in Google+ will help sparks serve relevant items. As users build more complete social networks and accumulate a history of shared content and +1s, Google will be able to use these signals to find the best results for each user.
If Google manages to accomplish these three things — and Google+ as a whole gains momentum — sparks could become a dominant, everyday source of personalized, topical news.