SEO testing sw

[davidd]

Has anyone got any recommendations for SEO stuff?

I have seen this

www.instantposition.com/members.html

which looks pretty good but wondered what else was out there.

Thanks

David

[J_S_G]

It's a fairly black art... the Search Engines tweak their parameters regularly, for a start, so whilst you could get a measure of whether a site SHOULD be successful, it may not tell the whole story. It's a lot like accessibility testing tools - whilst they can cover the basics such as "are all the necessary tags, descriptions, etc." in place, when you get to more subjective things that ACTUALLY determine its fitness for purpose, you're kind of on your own.

Using something to automate the grunt work - ensuring unique titles, meta tags, number of links, etc. is a good idea. Beyond that, an expert pair of eyes are about the best option.

Shout if you want me to have a quick look at a site gratis to spot anything immediately apparent. Beyond that, we should definitely be able to help if need be...

[davidd]

Interesting you use the phrase black art as that is exactly what I tell people when they ask me about it

All I'm looking fo is a tool that we can use to give a benchmark report for a site, sort of a before and after. Hopefully the results will do the talking.

Cheers

D

[J_S_G]

davidd said:

Interesting you use the phrase black art as that is exactly what I tell people when they ask me about it

All I'm looking fo is a tool that we can use to give a benchmark report for a site, sort of a before and after. Hopefully the results will do the talking.

Cheers

D

Hmmm, depends what you're trying to achieve... the impression that you've improved things for them or actually having achieved that. You could easily achieve the former, no tools I know of that can quantitatively achieve the latter.

I tend to use some of the standard Accessibility tools to give an overall "health" of a site in terms of semantic markup, etc - things like Watchfire. But never use a tool for something as specialised as SEO. The only results that tend to do the talking are the Google PageRank after the fact. A lot of the problem around this is the reliance on external factors to weight the impact of your SEO attempts.

[davidd]

I think what we are tying to acheive is something which will show them where the site is obviously wrong now.

So all the tag stuff, making sure the content is in line with the search criteria, that images are tagged, that the search engine can read the site etc, etc.....

We can then talk about accessabilty etc.

We then fix the site, or educate them. The work takes place and then rerun the same set of reports show how the site has been improved....

We then look at pagerank , stats etc to make sure that it is working.

D

[Plotloss]

Quick related question if anyone is about.

Keywords inside H1/H2 tags carry more 'weight' with the algorithm.

If you encapsulate your keywords in said tags and then override the behaviour of those tags in CSS so it looks like normal text, in keeping with your design parameters, does the weighting still apply?

[J_S_G]

davidd said:

I think what we are tying to acheive is something which will show them where the site is obviously wrong now.

So all the tag stuff, making sure the content is in line with the search criteria, that images are tagged, that the search engine can read the site etc, etc.....

We can then talk about accessabilty etc.

Not quite what I meant... the accessibility tools check for a lot of the same things that make for good SEO - it's all about semantic content. I.e. they'll check that there are page titles, that all images have alt tags, that links have descriptions, etc.

davidd said:

We then fix the site, or educate them. The work takes place and then rerun the same set of reports show how the site has been improved....

There aren't any tools that'll do this really effectively. Sure, you could write something that spiders a site, reads in each page as XML (as it should be XHTML), parses the meta tags, etc. and checks that they're all unique, etc. But how can it know that one tag of "twas brillig and the slithy toves did gyre and gimble in the wabe" and another of "lorem ipsum dolor sit amet", whilst unique, are totally pointless. Similarly, even if you made it smart enough to know that the site's about, say, cars, how will it know that that means that whilst one page should reference "cars", another should be talking about "engines", another "brakes", etc.

Any tools that you can get won't give a realistic view of whether a site is/isn't optimised... if you do do stuff like that, and make it your goal to go from a "fail" to an "excellent" on what reports say, then you're short-changing the client; you're doing it for the sake of a report, not real optimisation to drive their business.

davidd said:

We then look at pagerank , stats etc to make sure that it is working.

Problem with that is it can take a LONG time to get those stats back, meaning you could be at it for months.

[J_S_G]

Plotloss said:

Quick related question if anyone is about.

Keywords inside H1/H2 tags carry more 'weight' with the algorithm.

If you encapsulate your keywords in said tags and then override the behaviour of those tags in CSS so it looks like normal text, in keeping with your design parameters, does the weighting still apply?

Yes, it does. But if you have an excessive amount of the site marked up as H1s/H2s, firstly they'll most likely spot this attempt at skewing the page rank and mark you down/blacklist you for it, and secondly, that's TERRIBLE from an accessibility point of view... if it's a popular corporate/retail site, that could get you into trouble very quickly.

[Plotloss]

Modern browsers shouldnt see any difference though shirley?

The way my page is laid out would look odd if I started encapsulating keywords in the text in bolding tags, I'm just trying to highlight keywords to the spider without them being actually highlighted to the user.

Yes a tiny proportion of browsers would ignore the CSS so it would look very odd and Jakob Neilsen probably wouldnt be pleased but it would in effect be transparent to the user who is my market segment is generally tech aware. Analytics tells me that most visitors have some fairly high end kit, latest browsers, big res, 32bit colour etc.

[J_S_G]

Plotloss said:

Modern browsers shouldnt see any difference though shirley?

The way my page is laid out would look odd if I started encapsulating keywords in the text in bolding tags, I'm just trying to highlight keywords to the spider without them being actually highlighted to the user.

Yes a tiny proportion of browsers would ignore the CSS so it would look very odd and Jakob Neilsen probably wouldnt be pleased but it would in effect be transparent to the user who is my market segment is generally tech aware. Analytics tells me that most visitors have some fairly high end kit, latest browsers, big res, 32bit colour etc.

This isn't about CSS... it's about the semantics:
1. If you deliberately overuse H1s and H2s, Google et al will no doubt notice this and mark you down for it eventually, possibly blacklisting you if they spot what you've done and think it's deliberate (bear in mind they're VERY good at this stuff)
2. Bear in mind the way that the H1/H2 will "chop up" content, such that if somebody tries searching for a string that appears across both the P and the H1/H2, it might return a miss because of those tags in the way (same potentially in the preview content in Google results)
3. Accessibility isn't about how it looks, it's about the semantic meaning. Screen-readers will find an H1 in a P and tell the user it's a new section, etc. This'll SERIOUSLY p*ss off anybody with disabilities, and bear in mind there's no cap to the fines, etc. imposed on you for not making every reasonable attempt to make your site accessible.


Technically, an H1 can't appear inside a P, either, from a document structure point of view, but that's by-the-by nowadays

So, overall, I'd avoid doing that. I'd be more tempted to have additional headings that cover those key words, or a nice juicy, content-rich FAQ, or something really swish like a descriptive list of DL/DT/DD elements, with titles in the DTs, contained in H2s/H3s (save H1s for what they're REALLY for) and content in the DDs. You could then do some nice CSS jiggery pokery with DT.H2 and DT.H3 elements to make them "less intrusive" shall we say.

[davidd]

J_S_G said:

There aren't any tools that'll do this really effectively. Sure, you could write something that spiders a site, reads in each page as XML (as it should be XHTML), parses the meta tags, etc. and checks that they're all unique, etc. But how can it know that one tag of "twas brillig and the slithy toves did gyre and gimble in the wabe" and another of "lorem ipsum dolor sit amet", whilst unique, are totally pointless. Similarly, even if you made it smart enough to know that the site's about, say, cars, how will it know that that means that whilst one page should reference "cars", another should be talking about "engines", another "brakes", etc.

I think we have the same point of view on this, especially from an accessability point of view, we have been having a discussion for ages just about the value of alt tags and yes if an automated tester sees anything then it will pass a checker no matter what rubbish is written there (jpb001.jpg for instance).

As part of a wider discussion we have been thinking about means to give access to the data containted within a site to those who are blind or have very poor eyesight. The browser based web as we know it is a visual medium, there needs to be a completely different interface (I feel) for the blind.. If only I could sort it out

Anyway, thanks for your input so far. You still egging?

D

[J_S_G]

davidd said:

J_S_G said:

There aren't any tools that'll do this really effectively. Sure, you could write something that spiders a site, reads in each page as XML (as it should be XHTML), parses the meta tags, etc. and checks that they're all unique, etc. But how can it know that one tag of "twas brillig and the slithy toves did gyre and gimble in the wabe" and another of "lorem ipsum dolor sit amet", whilst unique, are totally pointless. Similarly, even if you made it smart enough to know that the site's about, say, cars, how will it know that that means that whilst one page should reference "cars", another should be talking about "engines", another "brakes", etc.

I think we have the same point of view on this, especially from an accessability point of view, we have been having a discussion for ages just about the value of alt tags and yes if an automated tester sees anything then it will pass a checker no matter what rubbish is written there (jpb001.jpg for instance).

davidd said:

As part of a wider discussion we have been thinking about means to give access to the data containted within a site to those who are blind or have very poor eyesight. The browser based web as we know it is a visual medium, there needs to be a completely different interface (I feel) for the blind.. If only I could sort it out

That's the whole point of the semantic web... people with minor visual impairments can make use of features such as CSS overriding (not great), or page zooming in the likes of Opera, etc. But for people with more serious disabilities, that's why screen-readers like Jaws exist. It's worth seeing how this reads a page out to appreciate the issues that crop up from normal websites. Things that can be fixed REALLY easily, and that mean you can have a very visually strong site that works for those with disabilities.

Speaking of Egg, I really pushed this there. Ignore a lot of the brochureware before you log in (that was next on the list!), but if you've got an account with them, log in and view the sourcecode. OK, if you use Egg Money Manager, the JavaScript that needs to bring in your other account details is painful, but look at the HTML for the .NET pages (.aspx extension)... you'll see that almost EVERYTHING is text, with background images to provide the brand. And that the tabs across the top are actually list elements with a hidden title to let you know they're related. And that proper FIELDSETs, LEGENDs, and other such tags are used properly to give real context. THAT's accessibility. And it's also what search engines love as it means the content's there, clear, and structured for them to spider.

"The browser based web as we know it is a visual medium"
Disagree on that point, to be honest... the web is HTML. And HTML is a descriptive markup language that presents documents with very clear, identifiable structure. The problem is people just think of it as the output of a DTP/graphics package. If you structure the content right, then it's perfect for rendering into shiny visual pages, or for reading aloud. Think of XHTML as just being XML that's suited to human-readable documents... CSS is the stylesheet (XSLT, if you will) that makes it visually rich, and Screenreaders are the translation for audio rather than visual medium.

Happy to wax lyrical on that all day long if you're interested!

[quote=davidd]Anyway, thanks for your input so far. You still egging?

Nah, left Egg about 2 months ago now. Had pushed things as far as they were willing to take it, amongst other reasons.