June 22, 2006
Everytime anything changes with Google’s algorithm, there seems to be a cacophony of calls that Google is Broken. With the latest update, it seems these cries have been going on longer than average – and there’ve been indications of problems which are rather more unusual than the average "dropped from SERPS" reaction, as well. In the past couple of weeks, three threads have popped up at Cre8asite Forums discussing different Google issues:
The common threads in most conversations seem to be that 1) Google displays data that shouldn’t be indexed, 2) Spammers have gotten through Google, and 3) Google isn’t putting my site where it used to be!
I’d like to just go on the record as saying that Google isn’t broken. They have no obligation to index what you want them to. They have no obligation to rank your site where you think it should be. From a webmaster or search engine marketer’s perspective, the recent Google fluctuations have created new challenges for optimization. There are new issues, sites may have fallen, but others will have risen. This is the ongoing process of search strategy – constant effort is required to maintain a quality site, because you never know what your competitors may be doing.
Yes, I agree that Google should not be indexing binary files. However, this is not something which will greatly effect most users. I agree as well that "t1ps2see", a now-famous spamming technique should not have gotten through Google’s algorithm – but it did. This is not because Google is broken, but because the thousands of brilliant minds at Google can not compete 100% with the billions of minds out in the rest of the world. It’s the infinite monkeys paradigm – an infinite number of users (or spammers) can accomplish anything. It will always be the case that there are as many or more people attempting to use Google for illegitimate gain as there are people trying to prevent that activity – and both parties will win their share of battles.
I think it’s important to think of Google from a user’s point of view at all times. Did your site not come to the top? Well, what did? Are these useful sites? If they are, then Google is doing it’s job. Again, it is not Google’s responsibility to support YOUR site – it’s their job to find quality, relevant sites. If you aren’t making it to the top, then you need to change your strategy.
There are particular searches which will never produce great results – these searches are too general, too basic, or too popular. I read an blog entry recently from Evan K Roberts complaining of a flawed search. His search was for "exotic hotels", and his complaint was that Best Western came up as the 11th result.
Sorry, Evan – but that’s not a good enough search. You searched for exotic hotels – but where? Considering the open-endedness of the search, I’m not the slightest bit surprised that a massive hotel chain with a huge website came up in your results. I think it would be wonderful if Google, or any other search engine could automatically understand our search intentions – but they don’t.
Any given search using Google, Yahoo!, MSN, or any other search engine is merely a snapshot in time – the results for this hour, this day, this data push, this index, this data center, this algorithm. Each company is constantly tuning their results and attempting to bring better results to their users, because that’s what their business depends on. Sometimes an update won’t work quite right – but as often as not, this is dependent on the perspective of the user.
The responsibility of a search marketing consultant is to make certain your site maintains the traffic, search engine results, and conversions you want, and to continue encouraging them towards growth. The responsibility of a search engine is to provide to a searcher the best possible results for their search. The fact that you think your site is the best does not mean that the search engine will.
I haven’t dealt with the serious problem of "made for adsense" sites. This is, to some degree, related to Google’s problems. Results are swamped with problems Google has created through the low-bar to entry of their advertising program. This is not something that should necessarily be corrected with an algorithm; instead, it needs better oversight.
Okay, that’s enough for this little rant.
June 21, 2006
-
Huckabuck
-
1. A homemade frozen treat consisting of a paper cup filled with a sweet liquid, such as Kool-Aid. Native to New Orleans.
-
2. A meta-search engine drawing results from Google, Yahoo and MSN amongst others and featuring a unique "search tuner" feature to adjust your results. Also Native to New Orleans.
It should come as little surprise that I’m not intending to write about the frozen treat. Despite the fact that it’s looking like a hot, muggy day and I certainly wouldn’t mind one – writing about frozen treats won’t bring them into my hands. The meta-search engine, on the other hand, I can easily take a look at, without needing to travel to New Orleans.
Huckabuck’s Technology Offering
Huckabuck.com is one of the rising crowd of new metasearch engines. They’re big selling point is the use of an equalizer-like search tuner, allowing the user to define their preferred blend of results. Not happy with your search? Perhaps you wanted a pinch more Technorati in the mix. I’ve mentioned before that I think user-customizable search tools are an important new development in search. This is probably one of the more user-friendly examples of this kind of technology I’ve seen.
I’ve reviewed attempts at customizable search before – I was highly critical of MSN’s Search Macros and not incredibly thrilled with the anemic offering from Mojeek. Both great ideas, but neither particularly well implemented. Huckabuck, on the other hand, seems to have a very nice, clean implementation. Unobtrusive when not activated, yet very easy to understand and use when you choose.
The implementation is in AJAX, with all the fancy graphical interactions available to that technology. However, appearance is not what I’m reviewing; I’m interested in results and usability. Here’s the tuner itself:
The use should be pretty clear from the outset. A set of sliders, to disable or adjust the use of these six data sources for the metasearch. Also, above those sliders, arrows to navigate between a variety of presets emphasizing technology (Digg), social search (del.icio.us), research (Google), shopping (MSN), or blog search (Technorati). Each preset places an emphasis on a different engine. Interestingly enough, no present places any emphasis on Yahoo – perhaps Yahoo is lacking in any kind of specialty?
As an aside, this isn’t necessarily a negative – Yahoo may simply provide the best results for a generalist search. Although, with the growing importance of contextual search and vertical search, this is a dubious positive.
Continuing to the right side of the selection panel, we have the ability to change theme (not interested), an autocompletion option (intriguing), a virtual keyboard (curious), and the ability to change the number of results per page (very nice). Some of these are useful, some are window-dressing, but they are all very clearly conveyed – and that’s what matters to me.
I played with the autocomplete feature a little – it’s a lot like Google’s Suggest. A neat time saver for common searches. If you’re looking for something a bit more off the beaten path, it doesn’t have much to offer, of course. The virtual keyboard was also a little intriguing. I’m not sure this is really a worthwhile addition – perhaps, in combination with autocomplete, it could be somewhat handy. However, to me the primary purpose of a virtual keyboard should be as an aide for mobility impaired users. This particular implementation will certainly be usable by that population – as long as they browse with JavaScript activated. The AJAX implementation of the tool does mean that, from an accessibility perspective, it may not be perfect.
In fact, without JavaScript, none of these tools will function at all. Still, there are some advantages to browsing without JavaScript – the advertising won’t function as well.
What about the search results?
Well, there’s not much to say there – Huckabuck isn’t using their own algorithm, so the results are rather at the mercy of the engine they’re drawing data from. However, the customization available using the search tuner gives them an edge – if you don’t like their results, it’s very easy to change. I do wonder, however, whether it might not have been more user friendly to use their presets concepts as their slider categories. As it stands, you adjust the importance of a data source in the results – what if, instead, you adjusted the importance of, for example, commercial data in your search and the interface adjusted MSN’s importance in the background? Many users may not be aware of what engine would be best to adjust for their results – but would probably be more quickly able to identify what type of data they were more interested in. Huckabuck’s slider presets provide this data, but the sliders themselves are a more obvious tool than the selector for the presets.
Other thoughts about Huckabuck
I like the fact that the owners of Huckabuck, Chris Schultz and Blake Killian, are deeply involved with the New Orleans community. They are sponsors of BrainJams New Orleans, which is doing a lot of work to help rebuild New Orleans businesses.
I also love the fact that they donate a portion of their search profits to charity. For a startup, this is a exceptionally wonderful choice. Currently, the beneficiaries of Huckabuck’s Social Mission are Tipitina’s Foundation, working to rebuild New Orlean’s music culture, and the Backbeat Foundation, also supporting musicians in New Orleans. As a musician myself, I certainly find their goals uplifting.
More information:
June 20, 2006
Peter Da Vanzo and Barry Schwartz have each posted recently on a question asked at Google Answers – specifically, the question was:
What percentage of Google searches are contextual? That is, what percentage of searches occur in response to content a user is reading on a page (e.g., reading an article about Qualcomm and then doing a search on "Paul Jacobs") versus "from scratch" searches that are not related to content a user is already reading. These contextual searches sometimes also occur via a "right-click" on a highlighted word within Firefox to search the Web for [highlighted term].
Google did not deign to answer or allow the question to be answered, which is the reason this has created a little bit of a stir in the blogsophere today. Both Da Vanzo and Schwartz described this as a form of censorship – which is an interesting question.
Now, on the basis of the question asked, I would have to say this was never going to be answered. This kind of statistical data is fairly consistent with information commonly covered by Google’s non-disclosure agreement, and certainly not publicly available at the moment, so censorship of the answer doesn’t appear to be the situation. No data was removed; and a failure to provide it does not constitute censorship.
However, Google removed the question as well – thus removing the possibility of some person attempting to answer it who may not have had that proprietary information. It seems foolish to prevent somebody from attempting to answer the question in a more discussive context. It also seems highly questionable to make the specious answer
"We’ve removed your question because you can find the answer on our main site, free of charge. All publicly available information about Google is available at: http://www.google.com/about.html.". The second part of this response is true – but the first part, as far as can be ascertained, is not.
I can understand why Google may not have wished to provide this answer – however, I think the choice to remove the question and answer it in this manner was unwise. A simple statement that the answer to this question is not available due to Google’s information disclosure policies would have been much more sensible and acceptable.
« Previous Page | Next Page »