Avid readers will recall that I’ve covered Hakia before. But that was almost a year ago. Since then Hakia has added a new feature that gives it more of a social feeling. It’s called Meet Others (MO). First, you ask a query and receive your search results. Then, in the top-right corner of the results, you will see an icon that says “Meet Others who asked the same query.” Clicking on the icon takes you to a room (if it exists) of people who have performed a similar search and decided to leave a message in the room.
Once you’re in the room, you can post a message yourself or contact someone who has already posted a message. All you need to do to post is authenticate your email address. If you want to be contacted, you can choose to receive messages via email (which Hakia masks) or IM (Skype or MSN). There’s also a voting system, which helps determine how long messages stay in the room.
Here’s a screen shot of the room I saw after I clicked on the appropriate button from a query about social search engines:
What the image doesn’t show you is that Hakia also provides links to rooms with similar queries (in this case, “search engine”) and popular rooms. At the time I made the query, these included rooms discussing cancer, New York real estate, and San Francisco computers. And as you’d guess, the system is not immune to spam; the “Search Engine” room included a post from someone apparently advertising their services as an “experienced SEM professional.” You don’t have to be registered in any way to report a particular post as abuse, though; just click the “report abuse” link that appears with every post.
But plenty of posts weren’t just advertising. Some people asked questions, such as which search engine offered the best contextual advertising solution. Others posted links to articles. When I entered the cancer room, I noticed other kinds of posts: book reviews, advice, and more. It had the feeling of a cross between browsing a general forum and one of the groups on Searchles, especially when a conversation gets going around one of the links posted to the latter. Not everybody will want a social element with their searches – but this is an intriguing way to implement it for those who do enjoy it from time to time.
When I first heard the concept behind Enetez, I thought they were nuts. Instead of using your keyboard, you search with a series of drop-down menus. How do they think that’s going to cover all of the topics about which someone might search?
The answer is that they don’t expect to cover everything. When you arrive at their portal, you see that they have a regular search box fed by Google. You can choose to use their sans-keyboard search feature. You start with country and category drop-downs, and then get more specific. Here’s how my search started to shape up after using a few drop-down menus:
So what kind of a result did I get from this? When I first clicked into the result, more than half of the screen was taken up with what I thought of as extraneous information. In a large rectangle on the top right was a box that displayed my search. Below that was another rectangle that ran the full length of the screen filled with sponsored ads – all for Fort Lauderdale banks, not movies. But once I scrolled past that, my results weren’t too bad:
What do I think of this? It’s really an interface and not a search engine in its own right, which is why it talks about being “enhanced by Google.” It reminds me very much of the “phone trees” you sometimes have to navigate through when calling insurance companies to get to the right person or department. On the other hand, if you’re just looking for something entertaining to do in your area and haven’t really focused on what yet, this can help you narrow it down. There are many other categories of information that I didn’t list here, so it’s not just for entertainment.
The interface works as advertised. It really is easier to use. Who needs to come up with keywords when they’re listed right in the drop-downs? I can see myself bookmarking this search engine to use for certain very specific sorts of searches. It’s not a full answer to every search, but it does certain things fairly well.
True Knowledge, alas, is in private beta. It does feature a fascinating seven-minute video on its home page, however, that explains the technology. They say they’ve created a technology which translates natural language questions into a form that computers can understand. This addresses what they see as one of the fundamental problems of search: the fact that computers cannot understand the content of web pages.
As TechCrunch observed, this potentially puts True Knowledge in competition with Powerset. There’s an important difference though, which TechCrunch also points out: Powerset is indexing the web as well as trying to convert natural language queries into a form that databases can understand. True Knowledge is not indexing the web. Instead, it is collecting data from structured databases. It is also accepting data (and corrections to data) from users. The video at least implies that they have put some protocols in place to prevent the kind of vandalism that Wikipedia experiences from time to time.
Web surfers are an adaptable bunch, though. We’ve had to learn how to ask questions of search engines in a language that they understand. It may not be natural language, but it’s not that far off for experienced users. For example, one of the natural language questions that Google has problems with, according to the True Knowledge demo, is “is Jennifer Lopez single?” And it is true that, if you put the question in that form into Google’s search box, Google doesn’t really seem to know what to do with it.
But look what happens when you think about the question a little differently. I put “Jennifer Lopez spouse” into the search engine (without quotes) and here’s what came up:
Is this an instant answer? No. But clicking on the third link revealed that Lopez was married at least as of November 8, and clicking on the first link revealed that she’s been married three times. Don’t get me wrong; I’m sure that True Knowledge’s technology is a major improvement on what we have now, and may even be a godsend for some of us. I’ll probably use it myself once it goes public. But I do find myself wondering about the real size of the problem, and how well True Knowledge will solve it.
With each of these search engines and their new features, it seems to me that what we’re seeing is parts of a whole solution when it comes to the problem of search. Each one shows a different way to approach the matter. I’m tempted to say that each one approaches the question of search a little differently, but the truth is that each one poses slightly different questions, so of course you will get different answers.
For True Knowledge, the question seems to be “how can we help the user get the most accurate, most relevant answer to a natural language question as quickly as possible?” For Enetez, the question is more along the lines of “how can we help the user zero in on the exact thing they’re looking for?” And Hakia’s approach is one answer to the question “how can we help the user find additional relevant information about their query, perhaps in a social context?”
In the search field, there doesn’t seem to be one right answer – because there isn’t one right question. Searchers themselves will approach search differently. Two different people can put the exact same query into a search box, and may be satisfied with slightly different results. For that matter, even the same user will approach different searches with a different attitude.
For example, if I put the query “Thanksgiving in Disney World” into a search engine, am I looking for discounts, good places to eat, another family’s experience, or what? I might be looking for a hotel discount at one point; later in the planning, I might be looking for the best place to eat in Disney World for a Thanksgiving dinner. That latter might be a good topic for a social search, incidentally; everyone who has been to Disney during Thanksgiving is bound to have an opinion (as far as I’m concerned, it’s the Liberty Tree Tavern on Main Street, but I’m just one person).
Since everyone has different goals when they search, it is a good thing to have different approaches to search. Hakia and Enetez have the right idea by offering a default search mode/interface that most people are used to, but providing the option of doing something differently. Even True Knowledge offers links to related documents as part of their search results. With all this idea fermentation going on, it won’t surprise me in the least to see the major search engines adapting themselves more to how their users approach search – and giving them more options – over the next year.