Last night, AltSearchEngines hosted the first Great Debate. The topic was Continuous Search Engines and Personalization, and the debaters wwere Mark Zeman of Searchbots , Morgan Snyder of Allth.at, and Kalem Fletcher of Swamii. You can see the entire debate in the comments section of this post – and leave a comment of your own if you like!
test
Caronte Web
Federated Search
Highlight Health
MSearchGroove
Ramenos Blog
Reel SEO
Software Abstractions
The Genius Files
The Next Web
June 3rd, 2007 at 11:29 am
Can we post questions in advance? I have a *ton* of questions one of your participants needs to answer.
June 5th, 2007 at 6:48 pm
My bad! The Debate will start at 5:00pm PST or 8:00 EST (in 15 minutes) Thanks!
June 5th, 2007 at 7:02 pm
Good evening! This is our first ever debate, so we have our fingers crossed!
The three alts tonight are Searchbots, Allth.at, and Swamii. Let me go ahead and see if this works by asking Kalem to post a “Hello” comment with a paragraph about Swamii, and then Morgan post a comment about Allth.at, and then Mark of Searchbots. I presume that we may all have to refresh our screens to see the comments appear. Let’s give it a try!
June 5th, 2007 at 7:07 pm
Allth.at is the search that never stops. We will keep looking for your item on the sites you select and report new search results back to you. You can also choose to have alerts emailed to you or, you can subscribe to the RSS feed and be alerted of new results right in your favorite reader.
June 5th, 2007 at 7:10 pm
Searchbots.net is an experimental search engine that investigates the use of mythology, personification and game theory as motivational strategies in creating a sustainable search community. Searchbots has a rich history and is unique in that it allows you to search using more “human” and entertaining types of information like colour and mood. If you picked the colour red you might get a website about tomatoes, communism or angry people.
Searchbots is a research project conducted as part of a Masters in Design at Massey University, New Zealand.
June 5th, 2007 at 7:16 pm
Personalization: To personalize something means that I have to be
able to give feedback that let my tastes be known. My first experience
with personalization on the internet was my original myYahoo page…
where I could determine not only the content sources of my page but
also the color, layout, etc. I loved this.
Many years later and out of my own frustration with search (namely not
finding what I was looking for) I realized that personalization could
be applied. I then realized that personalization and continuous
search were inextricably linked. The major engines do everyday search
VERY well… certainly this is why they have the market share that
they do. However, when it came to the “hard to find” items or
persistent topics of interest, this is where I felt the majors were
falling down.
-Hard to find: This was an exercise in manually searching
sites/sources… a laborious and time consuming process. Further,
there was an opportunity cost to not finding what I’m looking for… both for me and the provider to be.
-Topics of interest: How do you continually follow a topic of interest ? At the time, the only easy answer
was google alerts. But this certainly didn’t allow alerts for other sites/sources.
By combining personalization with continuous search, a user can now search, get results, personalize or “fine tune” and then continue the search. And that’s the ultimate goal… helping people find what they are looking for.
June 5th, 2007 at 7:19 pm
Let’s get the ball rolling. I guess the obvious point is that you can’t have continuous search without some form of personalization and an on going relationship that you setup with your users. What’s your philosophy to managing those relationships over time?
June 5th, 2007 at 7:28 pm
One aspect that interests me is how active or passive the feedback mechanisms are within our various applications. If it’s personalized then it needs to collect information overtime. Do you have a passive “rate this result” button or force the user into providing feedback. Something I’ve experimented with in Searchbots is using the metaphor of the search agent actually talking to you to encourage feedback. You can skip the option to rate a website you’ve just visited but I’ve been surprised just how keen people are to engage with their Searchbot and provide valuable feedback about the personalized results.
June 5th, 2007 at 7:32 pm
Agreed Mark. I think the majors have long argued that users are lazy and just want them to “figure it out”. It seems with the recent UI changes among them, they are starting to change the tune a bit and let people give feedback to the app.
June 5th, 2007 at 7:36 pm
Our first glitch. For some reason I had to approve Kalem’s comments. Please let Kalem catch up and go next.
June 5th, 2007 at 7:41 pm
We’re always going to be niche engines but I see that as a strength. Users that are more interested in search and interacting with the search results. I’ve just run a user survey and it was interesting how many people loved Google’s breadth and speed, but really that just wanted to interact with what was spit out. “Never show me this result again.”, “Show me a thumbnail preview”, “Turn off ads”, “Turn off commercial websites’. It seems to me people are very keen to contribute. Does anyone know what happen to Wink and why they veered away from rating Google results?
June 5th, 2007 at 7:48 pm
This has been my experience as well… users really do want interaction. I suppose this gets to how the alternatives compare to Google and Yahoo
Alerts
Google and Yahoo alerts both work well. The problem is that they are both married to their index… which can only be updated so often. The best example of this would be when you are looking for something and they serve an ebay result, you click on it and get a message like “bidding for this item has ended.”
By using a combination of API’s, RSS and index results, we are able to get more timely results because we are most time directly linked to the source database through the API.
June 5th, 2007 at 7:58 pm
Hmmm… how up to date id the API you’re accessing though? Alerts are great but the problem is that they are just literally “alerts”. I think the benefit we offer is personalization over time. For example at a very simple level Searchbots keeps a full history of what you’ve visited or been sent in your reports. It then uses that history as a filter so only new websites with a high quality rating you haven’t seen before are sent to you. I’m not sure if Google Alerts is linked to Web History yet? It should be?
The idea of machine vs human or index vs social network start to come up. How do we merge the two in an ongoing fashion. Eurekster takes a crack at it but not in a continuous way.
June 5th, 2007 at 8:11 pm
Yes, it seeems that all of these things (machine, human, social network) are all somehow on a collision course… I don’t think there is one “right” answer. I’m also not sure that the next great algorighm is the answer. It certainly is interesting to watch and be a part of.
Personally, I’m in favor of a neutral, search platform… a search application (not an engine) that can leverage all of these great things as they evolve. This way, when the next thing does come along, it can be added as a source to the platform. And ultimately, users can find what they are looking for.
I’m done. Thank you Charles!!
June 5th, 2007 at 8:11 pm
Well it’s been an hour, and I’ve got to let you go. Feel free to stay and see if we have comments from readers. But please check back occasionally as many people may read this conversation later. I will keep it posted. Thank you very much for being our first debaters!
June 5th, 2007 at 8:17 pm
Readers: Even though this live debate is now over, feel free to comment on the coversation above. The three debaters, Kalem of Swamii, Mark of Searchbots, and Morgan of Allth.at may pop in later and respond to your comment. We’ll consider the comment period closed after 24 hours.
June 7th, 2007 at 5:01 am
Hi guys,
Welcome to the blogosphere, Alt Search Engines!
Thanks for the interesting debate, guys. Mark, I’m particularly interested in your comments about keeping user history as a filter. As you’re no doubt well aware, privacy is a major concern among users. What has reaction been to having Searchbots store user activity, and what is your privacy policy?
I’m also intrigued by your comment that you’re always going to be niche engines. In one of my recent posts I proposed that the search engine network currently has a ‘winner take all’ topology — essentially, that Google has become such a monolith that the industry no longer follows a power curve (I pinched the terminology from Barabasi’s book ‘Linked’). Do you think that’s true? Or do you think there are any search engines out there that stand a chance of occupying more than ‘just a niche’?
Best regards,
Kaila Colbin
August 29th, 2007 at 7:22 am
[...] first debate! Continuous Search Engines with Allth.at, Swamii, and [...]
November 13th, 2007 at 7:13 pm
[...] of our past debates in case you missed them. Just click on the first word for the whole debate. Continuous Search: and [...]