This is part of a longer note on internet content in general and Demand Media in particular. The full report can be downloaded with this link: Demand Media Crapification (pdf).
Do you Quora?
A new kid on the block has captured the attention of alpha geeks and early adopters and has certainly entered the hype cycle as the “next big thing.” Quora starts with a focus on questions and answers, adds to it elements of social networking (like Twitter) plus a crowd-sourced ranking mechanisms found in Slashdot and made popular by Digg. This sounds a little esoteric at first but it ends up being effective at turning online information and the energies and knowledge of the many into productive results.
The crux of Quora is that it allows users to determine what questions to ask, answer them, and then provokes the community to rank the best ones so they are at the top. In this way they combine elements of Wikipedia with Digg; however, the ability to follow topics and have networks of followers adds another dimension.
There are a few reasons Quora is proving more useful than other online networks, so they are worth a closer look:
Topical focus: Quora uses questions and answers to provide a framework for information sharing. Users can submit new questions as well as answers to existing ones. There is an approval filter on questions to guard against multiple instances of the same question and “junk” questions.
Notion of authority: Quora has been blessed with a very high level of initial user. For example, answers are from individuals who are clear authorities. It’s not unusual to see a founder or CEO of a company answering questions like “how did ABC company get their name?” But anyone can end up with the best answer to any question by dint of knowledge and their effort to express it.
Crowd filtering and ranking: Questions tend to elicit multiple answers, and in addition to authority the users vote for the answers they think best address the question. In this way, the best information “bubbles up” to the top of the list and silly or wrong answers are pushed out of focus. Over time this leads to an inventory of “right” answers to interesting and useful questions. This is a filtering mechanism that works very well.
Intersection with networks: You can “bring your network” to Quora by following people you already know and also start following new ones. The combination of following people and following questions leads to a more productive filter that still allows for discovery.
There are some other benefits that result from the Quora approach. The first is that people stop adding their 2 cents when it’s not needed. After a question has been addressed for a while, the best answer is right on top and is typically quite good. This helps prevent subsequent visitors from adding content that doesn’t improve the answer and instead encourages them to vote, comment, answer other questions or ask new ones.
Quora is also a good place to build your network and meet new people because it’s based on shared interests, knowledge and constructive interaction. In other words, it’s about the content that people generate more than the jobs they’ve had or people they know. LinkedIn takes the latter approach and is not nearly as effective.
Is Quora a flash in the pan? That’s a question that has some people worried because they enjoy the service so much. Part of what has made Quora special is the quality of the dialog and the fact that so far the company has been willing to forgo a business model. Will a huge increase in the number of active users result in a “dumbing down” or a “smartening up and broadening out” of the content? Quora does have a few mechanisms that may protect them from the adverse effects of a very large user base but these will need to be tested.
The other main risk for Quora comes from within. Users are building something that may be of great value and the company will want to monetize it to stay in business. Advertising is the most common path but doing so without jeopardizing the character, energy and quality of the site is a challenge. Right now the costs to run Quora are low and if they stay that way it may be possible to preserve the aspects that make the site special. Craigslist is an example of a site that was largely able to preserve their free and open model and only charge for specific things like job postings.
What can Google do?
There are some deep concerns about Google being able to address the shortcomings of its search results. The focus on speed and general answers for search may fail to capture enough semantic processing to deliver improved results.
Of course the folks at Google know all about the problems we have described here, and have myriad solutions they could apply. Technologies like instant search make it clear how deep their technology reservoir is. Google now finds itself facing the innovator’s dilemma.
Google has been disruptive by giving away valuable services “for free” since they could earn billions of dollars in advertising fees. Today that’s a huge and very profitable business for Google. If the path to continue to delight users happens to involve showing far fewer ads, is it one that Google can afford to take?
Based on recent statements, it sounds as if Google will make noise about improving results but is unlikely to do anything radical. As the market leader it can afford to move slowly, at least for now. However, there are a few fairly simple but powerful changes Google could implement to improve results:
User-defined site rankings: A simple way to provide positive and negative feedback on sites would make a big difference. A heavily bookmarked site generally has good content. Offering the opposite would be sensible. There are many sites for which I would check the box “never show search results from this site.” Google could make this easy.
Leverage social and professional networks: There are social and professional graphs emerging that also go a long way to defining some notion of authority or at least superior judgment. Google already can show me a search result that has been bookmarked by someone I’m following on Twitter. Ranking sites bookmarked or tweeted by people one follows would be a helpful option.
Penalize sites with high advertising content: This is crux of Google’s problem. Users want more content and less advertising. We’ve reached the point where pages have more ads than content and viewers are beginning to find other places to look. Google isn’t in danger yet but if they don’t do something about the advertising proportions of returned results they’ll lose share. It might take years, but once a decline starts it’s hard to stop. [Look how many years it took for Microsoft (MSFT) Internet Explorer to shrink from having the vast majority of market share to now less than 50% and falling.]
More collaborative filtering: Taking bookmarking and combining it with search and networks would allow Google to show users new sites in results lists that are more likely to fit the nature of the searcher. This would generally improve the quality of new sites being discovered.
Making enough changes to produce high quality results would could put a big dent in advertising revenues.
The race is on in terms of who will do the best job integrating social graphs and networks into filtering information and search results. One likely outcome is that search traffic will spread out more evenly among more players and be less concentrated in Google. It’s already more natural to search Yelp if you are looking for restaurant reviews, Amazon (AMZN) if you want to buy a book, eBay (EBAY)if you want to buy a used motorcycle, Foodily if you want to find recipes, and Craigslist if you are looking for an apartment. This trend will continue and erode some of the power Google has in the market.
Time for a New Search Engine
Internet content management will be an unending cycle of massive undifferentiated expansion followed by the tools and attenuation need to turn it into real intelligence. Recognizing that it’s a cycle scientifically and realizing that business models end up getting embedded into the existing players means that starting a new search engine company is a good idea.
However, “search engine” is probably not the right term and should be replaced with something more active like “content finder” or, more charmingly, “librarian.”
As noted, users are already searching on individual sites that tend to offer either the general content they like (as in National Public Radio or The Economist) or better answers to specific questions (like Yelp for restaurants and Amazon for books.)
There’s a “curation” craze on right now that should die out and be replaced by intelligence based on actions (bookmarks) and relationships (social networks) that can scale more effectively.
This is a huge global market that offers extremely high returns on invested capital. Why aren’t we seeing more investment in better information finders? We’ve seen a few like Blekko and Wolfram but there should be more. It turns out it only costs about $25M to create a fully functioning, scalable search engine. Adding some intelligence to the results and offering improved results would create a large multiple of value on that investment.
There are also some existing companies that deserve a second look. For example, a search property like InfoSpace [(INSP) – $8.54] has a small market share but enough of a presence to do something interesting in this market and improve their fortunes. Other companies like Oracle (ORCL), IBM Inc. (IBM)and HP (HPQ)are in a strong position to create technology and online services that rival Google. We’re due for a little disruption in search.