April 21, 2005
They launched on accuweather. I can see them on the right-hand side. Searching shows results within the accuweather domain and provides the map. Clicking on a location also provides the profile page on the accuweather domain with a map. The maps aren’t the interactive kind that is on google maps but, instead, the static kind like all other maps we’ve seen.
At the top are a bunch of google ads.
User-agent strings found out there seem to imply they are google-related…I have my doubts.
April 20, 2005
So it looks like google is at it again. I have already mentioned how disappointed I am in the search engines’ personalization arena.
Well, by that time, google had already launched google personalized which really didn’t impress me a whole lot. I didn’t find the resultset to be very helpful. It involved selecting a bunch of categories you liked but, since my searches tend to be very targetted (I am very specific in my searches, using lot of keywords I know must be in the document and remove them to expand as necessary), this wasn’t very helpful.
Now, google is launching Search History which allows you to take all your searches/google pageviews with you as long as you can log into your account. This seems like both a useful tool (kind of like being able to see your browser’s history remotely)…but keep in mind that this is just another tool that could be used to get people to register for better targetting/tracking.
This tool is definitely better than the others so far. It allows you to pause and resume the tracking and, when you view the search history, you see your search activity (number of searches) by date on a color-coded calendar. You also get a list of searches with items you clicked on that search below it as well as a list of searches you did not click on at all.
Other cool features include tracking when you clicked on a result (it shows up on the saved searches page *and* on the web results page next to the page you clicked on)…and yes, they do provide you with a way to clear history and pause tracking.
I can see privacy advocates up in arms now. Are they really naive enough to think that they couldn’t/wouldn’t do it before. Every site does analytics to track and optimize their goals (leads, conversions, etc). All they did was make this available to users as well.
Looks like intelliseek sent me another robot with the following user-agent string just minutes ago:
Looks like the domain was registered less than a week ago and is still serving up a godaddy parked domain page.
What are the makers of cool tool “blogpulse” up to?
Snap is a small search engine still being incubated by idealab…as a sidenote, I’ve worked for at least 2 idealab companies
They recently launched a tool similar to google’s suggest except that rather than show you the number of matches, it shows you how many people searched for the term.
Snap is also providing substring matches. I may find this useful 1% of the time but 99% of the time it should be anchored to the beginning of the text. For example, I went to the site to search for airlines but, since the results are sorted by the most searches, “southwest airlines” was well above “airlines.” In most cases, it won’t be a problem (that “southwest airlines” shows up before “AAA”) since, as you type (especialy for short strings), those results disappear. However, this isn’t the case for the “airlines” example. I suppose I may come up with other cases where I would want to use this (more than the 1%) but they aren’t coming to me right now.
Overall, a pretty cool tool nonetheless. Kudos to Snap for putting a cool product together.
After Yahoo announced earnings that said they doubled their profits due to online advertising, they also announced the completion of migrating overture to Yahoo Search Marketing.
It seems these days as if all CPC providers (or soon-to-be providers) are moving towards a one-stop-shop for online advertising. Yahoo’s Search Marketing not only contains Overture products, but paid inclusion, travel, shopping/products, etc. Expect to see more of this.
Seems like they all want users registered as well. This is partially due to a better user experience (which may not always need registration anyway) but could just as easily be used to track and serve ads specific to their needs. I noticed, for example, that google has taken their gmail accounts and tied it in to froogle so you can save things to a shopping/wish list, etc.
next up for reporting is google…
In the past, I noticed Google Ads on some of my new pages would show Charity-related Ads. However, I noticed in the past few weeks that they are now serving relevant ads.
What used to happen was that (and this is a guess based on what I observed) pages that were already in the google web search index would get relevant ads immediately. This implies they looked up context from thei cache (or somthing else built from their cache).
Now, however, they seem to use their media bot to crawl in realtime and serve up the page in real time. It happens pretty quickly but I have noticed on rare occasions that a page would hang for a second while waiting to load a google ad.
I think somebody realized that they are losing a lot of money because of this and that is why they fixed it.
April 18, 2005
then check this out.
They provide you with lists of these types of domains…just keep reloading the page until you find one you like.
Brought to you by the same people who brought you buzzword bingo…a few of you out there know what we used that for
MSN is testing clustering (examples: apple, jaguar) similar to Vivisimo’s Clusty.
Clustering is an interesting idea that allows the user to narrow in on what they’re looking for…search engines aren’t min readers and they cannot default users down a specific path they “believe” to be correct because there will never be an answer here that is 100% correct (and the suer will never be able to get to the alternate meaning).
In local, data is structured and is usually classified correctly because of it. Web pages, however, are unstructured, and must have better semantic translations. This usually entails the same technology as the “suggesters” most search engines already have but the clustering adds the additional overhead of only showing those alternates that have results behind them rather than blindly listing a bunch of alternates (and, in some cases, providing the total number of matches).
As you can imagine, the overhead of attaching various classifications to data can become expensive when done “on the fly” as data is retrieved from the search engine (which is what vivisimo’s technology claimed to do).
Tivo is in talks with google and yahoo about something that has been talked about for years since the early days of commercialization of the Internet…remember the word “convergence?”
Pretty interesting what they must be talking about, especially since both yahoo and google are making pushes into video (which go far beyond just video search) which seems to be pushing th envelope on convergence (were many Internet companies before them had failed…but they do have the funds that those previous companies didn’t).
As a side note, regarding tivo, check out these tivo-AJAX hacks.