17 replies to this topic

[aethernet]

It may have been said before, but I can't help but wonder if the SERP concept will become less relevant or a thing of the past.

Google Desktop, Google Notebook, Google Gadgets & Google Accounts will all gather end user specific information, habits and preferences. Google will most definitely use this information to target ads, but they might also use it to target more relevant search results. If so, your SERP's will not be my SERPS and my SERPS will probably be way off the norm for any websites that I do SEO for. Google will know my habits and they will be heavily skewed in towards websites that I manage.

Combine that information with all the data Google Analytics gathers about specific websites and you've got two variable end points introduced into the algorithm. Will this just add another layer of statistical complexity to SEO or will SERP's be pointless in the future?

[kensplace]

Serps are already pointless..

Unless your thinking short term that is.

[aethernet]

Serps are already pointless..

Unless your thinking short term that is.

Could you expand on this? I use serps regularly to judge how my SEO campaigns are doing and find them a useful snapshot to compare traffic analysis with. I can't reach my long term goals if I don't know what direction I'm heading.

Edited by aethernet, 14 May 2006 - 03:39 AM.

[projectphp]

[quote] I use serps regularly to judge how my SEO compaigns are doing and find them a useful snapshot to compare traffic analysis with[/'quote]
Depending upon your business model, that is, IMHO, a bit of a waste of time. SERPs, especaily if they start to vary not just from datacentre to datacentre but rather from user to user, will be a waste of time to meassure. Far better to measure actual performance (like revenue) than to focus upon rankings.

Of course, if you have a traffic based business model (AdSense based if you will), then you will probably be looking at differnt things...

[bwelford]

I think this is analogous to using Alexa Rankings as a quick and dirty method for understanding what may be happening to traffic. It's readily available and it's a nice simple tool to give you a rough idea.

Equally looking at the same keyword SERP over time gives you an indicator of whether some changes you have made are heading in the right direction. However I've already run into this problem of different SERPs for different folk. I even get it using Google either with Firefox where I have the Google Toolbar and Personalized Search History in place versus using the Google Toolbar set up in Internet Explorer. I sometimes get different results using the two browsers. Where this became most apparent my client uses IE so I always use that browser to check results.

[cre8pc]

This concern came up for me in 2004, when I began to sense that the web was getting "too big".

Search Engines No Longer Represent All Web Sites

I said:

The Web, as I said, is too big. Voting by link popularity and link connections is not an accurate measurement of web site worth. It doesn't represent the people. It's like having two major political parties. Sometimes it's a vote for who scares you the least, or whose concerns match yours. This isn't enough to elect true representation.

Mike Grehan had been writing about hubs and social networks and how search engines were messing around with networking and connecting people and search results in different ways. One of them is where they live. Another is what their interests are. And that was the tip of the iceberg. (Reference Filthy Linking Rich

With local search creeping in as the NEXT BIG thing in SERPS, what do you optimize for? Where we used to have separate versions of pages for different search engines in the old days, now there may be separate versions of pages for handheld devices, accessibility, countries, towns, genders, age groups, incomes, industries....

You get where I'm going with this? It's no wonder Google says its' computers are "full".

We will be served what search engines determine we should get. It may not be what we want.

[JohnMu]

You should take a deeper look at Google Co-op. It might not be "there" quite yet, but it is a giant step in that direction. I have not dug into the whole "social bookmarking" area very much, but from the looks of it, Google is heading into this direction.

Here's what I've spotted so far:
- you set up a profile for yourself
- you can label websites with all sorts of (your own) labels, sub-labels, with scores, etc.
- you can set up "subscription feeds" which are no more and no less than custom search results based on specific queries
- when you search for something that you've labeled, you will see similar labels on top of the results, click on them to refine your search according to your own rules
- when you search for something that is covered in your subscription feed it will show you the custom search results (on top), with text, links, etc.
- now add to that the possibility to subscribe other peoples labels / feeds
- and let them subscribe to yours
- and let the world see which labels/feeds you subscribe to
- and let them see who subscribes to your labels/feeds

It's simple at the moment (and <rant>would you believe -- they don't have any language tags?! has the web reverted back to the 90's where everything was in english? </rant>), and it is a bit complex to set up ... but imagine it in a larger scale: You have people you trust, groups that you trust, a whole network - when you search for something, it shows you the things that they have found as well, the things they have labeled.

The search results are no longer the same for everyone, not even specific for a single user or geolocation, but also contain things that you have decided that you are interested in, that you trust. Which entry in the serps would you click on: the one that "a machine" gave you based on your query, your query history? or the one that a "friend" recommends?

It will be interesting to see how this is adopted.... and misused (and how they handle that)...

John

[bwelford]

John, as you describe Google Co-op, I'm still not sure that if I do one of your searches on my computer I will get the same SERP.

[joedolson]

John, as you describe Google Co-op, I'm still not sure that if I do one of your searches on my computer I will get the same SERP.

Unless I'm off the mark, I think that's exactly what John was saying. With Google Co-op, it's even LESS likely that any given SERPs will be reproducible between two different people, on two different computers, etc.

[JohnMu]

With Google Co-op, it's even LESS likely that any given SERPs will be reproducible between two different people, on two different computers, etc.

Correct.

There are still many barriers in the way before this will be used on a larger basis, but it does seem like the general direction.

Just wondering what the next steps would be ....
Imagine something that works like pagerank / trustrank, but for people. You trust "cre8", many people that trust "cre8" also trust "eKstreme". Since you truse "cre8" you trust the feeds they publish and want entries from those feeds to show up higher in your serps. Since "eKstreme" is "related" (and weighed, perhaps based on the # of common trust partners), it could be assumed that I also want results based on eKstreme's feeds in my serps. That would almost get to the point that the results for searches are not based on SEO but mostly based on the direct and implied trust I place in them. Could it really come like that? That would almost mean .... good, unique content is king?

John

[joedolson]

Could it really come like that? That would almost mean .... good, unique content is king

My god...no way!...

But seriously, I think that the biggest barrier to this change is in getting people involved in projects like Google Coop. After all, when it comes down to it, the majority of searchers still aren't using the Google toolbar, or the Firefox google search, or, frankly, most of Google's or other companies "special" projects which woud allow search engines to gather customizing information.

Although I don't have any numbers on this, naturally, the opinion I'm inclined to spout is that the majority of people conducting searches are still just visiting a search engine and typing in their search. They aren't using any of the customizing features which would allow Google to provide these adjusted results - and I'm not certain that a change in that direction is going to be seen with any rapidity.

In certain areas, particularly technology related, I suspect that this will have a stronger effect sooner. However, in the most seached fields (law, health, shopping?) it'll be a very very slow percolation.

One of the biggest problems I can see with continuing to monitor basic SERPs is that people in the search marketing industry are likely to be users of these products - are we going to have to start having secondary computers, unmarred by special software that we use just for checking SERPs, because we need to know what the "default" search is as well as a wide variety of customized searches paths? I hope not...I'd rather see SERPs disappear altogether. (As a relevant element to monitor, at any rate.)

With the importance of the "long tail" and a diminishing verifiability that SERPs are consistent between different searchers, it seems that it's unlikely, whether projects like Google Co-op come to fruition or not, that monitoring SERPs can be considered really meaningful.

[FP_Guy]

One thing that I thought about is when localized geographical search ended up in the picture the SERPs were already different in different regions of the country. Depending upon what they were searching for anyway. Correct me if I'm wrong, but don't they read the computers IP address and bring you SERP results depending upon your local from that? Or was that 'one' of the things that they were going to try?

[JohnMu]

Hi Michael
being in multi-cultural Switzerland has it's advantages . We have 4 official languages + lots of english as well, it's a good way to play with Googles algorithms..

I noticed differences (without testing the geo-ip-location):
- using the different country Googles (eg .com/.co.uk/.de/.ch/.fr)
- using the different language filter options (eg "&meta=lr%3Dlang_de")
- using the country filter options (eg "&meta=cr%3DcountryCH")

Some samples (I don't know how they end up on your end ) - all searching for "cre8":
http://www.google.com/search?q=cre8
http://www.google.de/search?q=cre8
http://www.google.de.....ta=lr=lang_de
http://www.google.ch/search?q=cre8
http://www.google.ch.....ta=lr=lang_de
http://www.google.ch.....=cr=countryCH
(endless variants possible )

Keep in mind that by going to "google.com" often / usually it will automatically take you to a country / language Google (I think I have this overridden with a cookie here, not sure). So even if within the "same Google, same settings" the serps are the same, everyone is likely to have different settings anyway.

I did notice however, that Adwords geo-targets extremely well, much better than the Google search results, but that's the way they have that system set up - it would be hard to push the same or a similar system into the search results...

John

[MorgZ]

It can already be difficult explaining even the upper levels of SEO to potential clients who just want their page to be "high". I try to access success of a campaign by monitoring the site traffic and, where information is available, how the traffic is getting to the site.

If i can show that

a) site traffic has increased XX% and b ) traffic is finding your page from appropriate, relevant searches

then i feel i can show that the SEO / Marketing is working. You might be saying that i am missing out the crucial "revenue" measure but in a case where my company has not designed or developed the website in question (and the client has specifically said that they are happy with their site - even if its naff) its better showing success through a) and b ) and if they are still not getting the revenue they desire then its much easier pointing a finger at their website.

Edited by MorgZ, 17 May 2006 - 03:27 AM.

[korske]

Umm... What does SERP stand for?

[JohnMu]

SERPs = Search engine result pages
Usually you would track your ranking (how far up you are for important keywords) by checking the search engines (just doing a search like any user would do it). The question here is if that is relevant any more and if it can be checked at all anyway.

[korske]

Thanks SoftPlus. I've been here before but I'm still not very good with all these acronyms.

[AbleReach]

Keep in mind that by going to "google.com" often / usually it will automatically take you to a country / language Google

Lately I've wondered if some SE rankings change according to local region. Google's got the power: Adsense can be localized.

As an American I'm cool with getting English results, but I wonder if we'd be more rounded if we had more frequent exposure to the existance of other languages.

I hope I'm being superstitious about very local SERPS. If I wanted to see Seattle/Tacoma results I'd put Seattle in the search field.


Right now I'm making a site for a craftsperson who works in a niche that is almost exclusively WOM. Her keywords return manufacturers, not artisans. Keywords won't make as much difference as traffic and name recognition from a few respected authority sites. We're still negotiating about my desire for 400+ words per page. Among her peers' sites large images and very few words are the norm.

Edited by AbleReach, 04 June 2006 - 02:40 AM.