Google Page Rank

Let's stop for a moment and before I charge headlong into another experiment and see exactly what Google Page Rank is.

In short, it is a measure of how popular your page / site is according to Google's latest ranking. That little Google Page Rank bar at the top of your screen is seen as a measure of success (or failure) as it turns from grey to green. Maybe too much weight is put on it's exact value by well meaning people. But Google is a major search engine, it makes this value known to the world and like it or not, it's a measure that's here to stay.

A grey barred google page rank used to mean a banned site or very new page. But now it is all too frequent and basically means an unpopular page. A Google Page Rank of 0, 1 or 2 means a fairly unimportant (or very new) site / page and it's unlikely that the page will be read by Google very often – so changes might only be picked up once per month.

A Google Page Rank 3 or 4 is about the average, but recent updates have seen a lot of falls, so 3 might be more frequent that it used to be and 4 harder to come by. At this level Google will be visiting weekly.

Page Rank 5 is good, rare and seems hard to maintain for an average site. You need to have some good links in – probably a one way link or two from a quality authority site. For example accommodation sites can get this through being listed in good accommodation directories.

Page rank 6 and above is reserved for quality sites with trusted links in and here we are getting into the realm of daily visits from Google to collect changes.

Did The Last Page Rank Update Show Such Signs?

What have people commented about the last Google Page Rank update? Basically, people say that directories and sites selling paid links have been hit, dropping a Page Rank value of 1 or more (1 in terms of tool bar, not actual value).

Given the amount of this that has happened, how? Webmasters are saying that Google has had people manually reporting sites and finding sites linked to from popular sources. But a manual process would imply processing of exceptions in the ranking algorithm – i.e. a list of sites that sell links and have to be dealt with differently.

But with my experience in the industry I know what a pain exceptions are in code. Just 1 exception makes the code harder to read, harder to maintain, far more difficult to test and every time you change the code every exception has to be retested. So surely they couldn't have done that?

So a better use of time would to have been examining the sites that are reported as selling links and seeing what they all have in common. Remember by list from a couple of days back? What do the paid for links, the link exchanges and the gateway pages all have in common? They use keywords a lot in the anchor text.

So it would be fair to say that it's more than possible. Let's find out next week.

Would Google Just Ignore Links With Keywords?

In yesterday's post I explained why I think Google would ignore links with keywords in them. I'm not saying that the algorithm is as simple as that nor that it would stop there.

Say a page is full of links that Google decides are full of keywords and it flags as potential spam to ignore. What's the next obvious step? Well if too many of the links on the page are keyword related, then why not flag the entire page as potentially spam? After all, it could indicate it is likely to be a list of gateway pages or a links directory.

What when a lot of pages in a site are then flagged as potentially dangerous? If too many are flagged as dangerous then why not punish the site – drop the page rank of the site. It's using bad practices, so the page rank is dropped to prevent its bad practices going any further.

Am I convincing you yet? I'm not saying that it definitely happens, but it could be one of many indicators that add weight in on equation. Maybe there are certain levels of tolerance and only once you receive too many flags are you punished. But it needs looking at further.

Why Would Google Ignore Keyword Links?

Quite simple – it could be seen as a sign of link building for SEO – which Google does not approve of. Consider the following scenarios.

1 – You are writing a blog about your exploits and one day someone take pictures and posts them on local website. You thank them and link to the pictures, probably with the text 'read more here' or 'see pictures here' or even 'click here'. Not exactly high SEO links.

2 – You exchange links with another site. You are trying to improve your rankings for a particular keyword, what anchor text do you use?

3 – Pay for a blogger to link to your site. Again you are trying to improve your search rankings. What anchor text do you use?

4 – You are adding some gateway pages to you site, for particular pages. What anchor text do you use to link to the new pages?

Only 1 of the 4 linking methods above is approved by Google. It's also the same one in which you are unlikely to use high value keywords in the anchor text. Still think it is unlikely that Google would ignore those links?

Does Anchor Text Affect Page Rank?

I've been waffling on enough about my sites that I've looked at in which links don't seem to be followed because they use strong keywords in anchor text. But, is there any truth in it?

It's long been recognised that this technique can be used to increase search engine ranking. Search on "click here" (with or without quotes!) and you will find the first result is an adobe page, without either word on it. So it is apparent that anchor text does play a huge part in Google's algorithm.

So if a site had thousands of links for an important keyword then it would seem logical that the site would be top for that search. But, is that natural results or spamming?

Click here does appear in an astonishing number of searches – but mostly within phrases. It's not a high power keyword. And in all likelihood, most 'natural' links don't use a high proportion of keywords. So it would see possible that it could be used as a flag. So I will investigate the issue further.

So could there be a difference between linking to a site with Click Here and Mortgage Rates - other than what SEO experts currently tell us?

Week 6 Results

Forgot to follow up from the Week 5 Results yesterday!

Site 1 no change – same pages cached and not cached since 14/10 and 20/10 (home page).

Site 2, again the same pages remain cached, but they have been cached this week - 4 on Saturday 17th, 1 on the 18th and 2 on the 19th with the home page also cached on the 20th.

Site 3, also had it's home page cached on the 17th, the archive has not been cached since the 10th but we now have 6 article pages (1 more than last week) cached. But only one of the posts previously cached on the 5th has been visited again – on the 19th – the same day as the new post.

All sites have the most recent posts missing, but it is interesting to note that site 2 is starting to look more popular. All cached pages have been revisited this week – usually indicative of PR3 or 4, whereas site 2 has pages not cached for over 3 weeks, normally indicative of a low PR (pr2 or less).

Keep watching – next week back to normal.

Make Your Last Post of the Month A Good One

Or at least the first paragraph. Why?

It's something that has only just struck me. Look at what pages search engines have cached for your blog. Assuming the most important pages (to search engines) list first, then which pages list first?

Usually the home page then all pages that link off that – which are usually the archive pages.

When Google shows these archive pages, where does it take the text from? If you haven't got a long introduction then it's usually from the first paragraph of the top post – or put it another way, the first paragraph of the last post of the month.

So if there are keywords you are trying to improve your blog on, make sure that your last post of the month includes these well and that they appear in the first paragraph.

There's another flash of inspiration for you!

Sometimes Sites Just Flash By…

Amid the chaos of building and rebuilding other sites I was working on Bee Bees. It was one of those sites that quite quickly comes together from the instructions provided and is then shown to the customer.

At the same time an admin system was put in to let the customer update and maintain stock information, but this passed without note as it was all done very quickly and easily. There was little of note about it.

Then suddenly I was asked to publish the site – it's all finished and approved, get it in. I had to take a step back and then review the site. Everything had gone through so quickly and quietly that I'd hardly noticed the site being built and there it was ready to go live.

A good review and there were a few details missing, so maybe it's not quite finished. But once the text is over from the customer it's done and dusted.

Another happy customer and no complaints from me!

Why Does Google Display Page Rank?

Why does Google periodically make available its page rank? And why does it seem to coincide with major changes?

Page rank is being constantly updated. If you assume that pages are always displayed in an order based on page rank (and other factors of relevance) then by displaying all pages in a site (using site:) then all pages should be shown in page rank order.

To test this, I need to view a site immediately after a page rank export – but this assumes that page rank hasn't been exported from an old version… But it would be a good indicator.

So try it now on your site – do all pages appear in PR order – a few weeks after the recent changes. Probably not. This would imply that Page Rank has been updated since – but not exported. And this is what is supposed to happen – the export just gives us an idea of what is going on, not the actual hard values that are used.

So why? It generates far more traffic to Google in looking up the values for every page you open. It means they need a copy of the PR database for the public view (given they are not making actual values public). They need to maintain and update this system.

First and most obvious is that our browsing behaviour is reported to Google. It can track which pages are in use and popular in sites. But when popular sites drop from results, this isn't a major factor.

So what else can they be achieving? And is it me, or do they always export after major changes to the listings? Are they making the data available so that the real world then sees what's going on, have a guess at what took place and start to complain at what we see as unfair results within this – sites missed and sites unfairly hit?

That would then give them chance to refine their algorithms and make corrections. And it would be far easier than have a team of people sitting their trying to monitor millions of sites and the resultant changes!

Google Following Image Links

I mentioned a few days ago that a new site had been fully spidered quite quickly even though all links were anchored with images.

I also mentioned this in yesterday's post and said 'with no apparent search engine benefit'.

Well from what I have just been reading, it seems that maybe I'm wrong. There's a theory about that I've only just heard of (maybe I'm way behind the times here) about image alt texts.

It says that on the whole, Google now ignores alt text for images. The once powerful and almost guaranteed method for good search engine placement is now totally ignored (demonstrating evolution of search engines – why could they now now ignore anchor texts???). It's not quite that simple, but that's basically it.

With apparently one exception. If the image is being used as an anchor for a link then the alt text is read and used. It could even be that the link is only followed if there is some alt text in the link – it's being used the same way as anchor text used to be there.

Just wanted to clarify a couple of points and record what I've been reading!

Could Google Punish Some Links?

Would it be fair of Google to penalise links that make heavy use of keywords in the anchor text, as I suggested that it might be yesterday?

Outright no. I have a page about "Secured Loans" on my CompareMortgageRates.co.uk web site (yes, notice site name for safety!). Given secured loans is a keywords would I then be expected to change the anchor text to something strange (links as shown on the right)?

Overall, probably yes. It's a technique employed in links directories (which Google frowns upon), links sellers, spammers etc. It probably would catch more people that Google target than it would hit who should escape.

So it is something they could have tried. Just a pity that my mortgage site talks so much about secured loans, unsecured loans, mortgage calculators and other keyword specific anchor texts.

As I mentioned when I discussed Starz graphic links are being followed – and in this case followed well. Yes with no apparent SEO benefit because there isn't the chance to stuff the link full of keywords. In this site's case, every graphic link has been followed and all pages on the site are of equal page rank.

Google does evolve. Maybe this is part of the next evolution. Maybe they've tried to knock out the target sites and are now watching the drama unfold and seeing which other sites have been unfairly hit so that they can further refine the criteria and get closer to "perfection". After all, why else would Google make publicly available it's page rank? That must cost Google a lot of time, effort and cash. What do they get from it?

Google Punished Paid Advertising?

It's been talked about that the October Google update hit sites selling links. But how?

Well there is a page on Google to report such sites. But could Google have really checked every one of these reported sites and included it in a bad list? Highly unlikely. This would be a lot of processing of exceptions, which I'm sure Google would prefer to avoid.

Instead they have probably done what they have done in the past when they found spam sites etc. They've studied the pages and decided what these guys are always doing that other pages don't always do. More to the point – what patterns do the links show that natural links don't show?

Well paid advertising links, especially in blogs, will be written into an advertorial. There will be an agreed number of links from the post to the advertiser. All of the links will be live – no redirects, no 'rel="nofollow"' tags etc. And if these links are for best SEO, then the link text will frequently include keywords.

Notice that last one, we're back to keywords in anchor text. Now, if what I talked about last week about my theory that Google ignores keywords in anchor text, then maybe I'm starting to see the light at the end of the tunnel.

Also badly hit were massive link directories. Google is known to punish link farms. And don't most link directories also use keywords in the anchor text?

If this is what is happening is this fair? I'll take a look at that later in the week.

October Google Page Rank Update

A lot is being talked about the latest (October 2007) Google Update. Basically it appears that directories and sites selling links for Page Rank purposes have been penalised.

A lot of people are complaining that it's a major hit on their sites and I've heard of PR6 sites becoming PR3 – and I thought a drop or 1 was bad! It would appear that Google has targeted these sites selling links because links = "popularity" = increase page rank.

But some observers have pointed to sites that have been included that don't sell links and some that do sell links that haven't been included. On close investigation, some of those not selling have been found to use other popularity practices that are frowned upon.

So what could be happening? Well yes, fair enough if Google increases sites' Page Ranks by counting links in and people try to cheat this system them they are in their right to change their own internal ranking system. I agree with that – a lot of people don't. But how has Google identified sites selling links and hit so many so well? I'll take a look later this week!

Week 5 Results

After the end of Week 4 I have stopped monitoring the results daily. It was becoming just too obsessive!

Now rather than looking at pages cached, I'm trying to watch frequency of caching. It is known that popular sites are cached daily, unpopular sites less often and in my experience this can go to monthly or worse...

So site 1 - with no links - no displays 4 pages cached. Internal pages were last cached 14th October (2 articles & 1 archive) and the home page almost a week later on the 20th, sometime after that day's post. The most recent 2 posts haven't yet been picked up.

Site 2 - normal links - has 8 pages, all cached from the 2nd to 4th November. The home page is there, but the most recent post is missing. 7 of 9 posts cached.

Site 3 - target="_blank" links - home page cached on the 9th with 5 out of 9 posts cached on 31st October, 2nd November, 4th November (2 posts) and 9th November. Plus the October archive on 10th November.

It will be interesting to see in a couple of days whether site 2 has been fully cached again this weekend. It is getting a lot more consistency in its caching.

And Now I Really Am Confused!

I recently published MishkaOnline - a fashion website. This site uses CSS / javascript drop down menus throughout the site, so a week or two after publishing the site I wanted to check whether Google had successfully cached all of the pages that were listed through the drop down menus.

For those that haven't experienced the joys of coding CSS drop down menus (and my first 2 websites using these arrived on the same day!), basically you code the menu structure and then hide the rows using CSS. Not exactly what CSS was meant for, but it is a neat way and I'm assured fairly accessible.

In theory, Google should ignore the CSS that hide the menu structure and then just see all of the links. So I looked at Google to see what pages were cached. Apart from noticing that my descriptions were not as well written as they should be (now changed!!!) I was surprised at what pages were cached. For a start, the site map page, which I'd not linked into from any other pages, appeared. This site has already undergone "a bit" of a revamp, so possibly I had it linked in on the first version and removed later.

Apart from there not being many of the pages cached yet (I suppose it is only 2 weeks old) Google has started caching pages for the second time. This is apparent because of the restyling! But I suppose maybe Google has just started it's 2nd trawl and is only just taking in the pages a couple of clicks off the home page.

Improving Page Rank With Links

So is Google Ignoring Links With Keywords? I think it could be. It's certainly a way for it to get out a monster it (and other search engines) have created.

A while back there was keyword bombing where people teamed up and used inappropriate anchor texts on links to popular websites for a bit of fun. Google acted against these links individually. But if they realise there is abuse of the system, why not ban it all together?

It's an obvious move. People outside of Google work out how to falsely optimise sites, so Google changes the algorithm. It has happened before – alt text used to be a guaranteed way to the top, as did the title text.

It is known that many directory sites have suffered a drop in page rank – namely those selling one way links. One way links are known to really improve page rank and Google could have stopped this industry by reducing the page rank of the directory sites abusing their page rank. Including keywords in the anchor text was another way of optimising pages / sites and is often used in directories. If you are paying to improve your page rank then you want the best – so you get your keyword / phrases in the anchor text.

So why should Google not be able to work out that stuffing anchor texts full of keywords is (a) typical of directories and (b) a sign of cheating? Reduce the power of links using keyword phrases for anchor texts and drop the page rank of sites that do that a lot and you penalise the directories from 2 sides at once.

You also penalise innocent sites. Those that maybe are legitimately using these keyword phrases – possibly accidentally.

Maybe the percentage of such links should be taken into account. Maybe one day it will. But if you are link building to improve your page rank, don't put all of your eggs in one basket. Have at least some of the links using your site name or description – not just your favourite keywords.

Starz Ceilings on DIY SOS

I was reviewing a site that went live a few weeks before the last Page Rank update - Starz Ceilings. He's already 'Starred' on these pages (sorry – couldn't resist!).

The guy's products are about to be featured on the TV this weekend – something like Starz Ceilings on DIY SOS. So we're just giving the site a once over. I noticed that the home page had gone in at Page Rank 1, not bad given the site was only 4 weeks old at the time of the latest Page Rank export. What was interesting to note was that the rest of the site was also Page Rank 1. Usually for a new site I expect the internal pages to be 1 ranking lower than the home page, increasing only as the home page rank increases.

If that was not puzzling enough, then the fact that every link uses graphics makes it even more puzzling. Traditional theorists have always said that links using images don't allow search engines to spider the site. The theory is that search engines derive no information about the pages through the link and don't follow them.

But if these pages are all cached and all equal page ranked, then Google is visiting the pages and liking the links. The fact that Google can derive nothing from the anchor text isn't hindering the work – if anything the result is better.

I only noticed this after yesterday's post. But it does make me wonder more is Google Ignoring Links With Keywords?

Consider for yourself the sites I've shown. Those with graphics instead of text do well, so do those not using keywords in the anchor text. Those using keywords are being dropped…

Google's Page Rank

More about my confusion with Google's Page Rank.

Looking at the links in the paragraphs at the bottom of the mortgage site one paragraph has a lot of popular keyword phrases. These pages are grey barred. The other paragraph doesn't intentionally have a lot of popular keyword phrases, but some probably are well searched on. Those that could be well searched on are grey barred, the rest are Google Page Rank 1.

Seems to be a big coincidence? I looked at a couple of other sites or my own that are also struggling and were once quite good. Both have been Google Page Rank 4 until recently. There's my Mobile Phone Site and the Holiday Cottages Site. Until recently both had good home page Page Ranks, but the 2nd level pages weren't page ranked. Both use (legitimately) keywords / keyword phrases in the anchor links.

So I surmise that maybe rather than it being the length of the paragraph that the link is in maybe Google is starting to ignore links that use strong keywords in the anchor text. Not necessarily always, but probably in coalition with other factors. For example, maybe if one or two links on the page use high density keywords but then there are a few more that don't then you are OK. Looks like my sites are triggering the factors and I'm suffering!

£1 to Children In Need

UK Domains are offering to donate £1 to Children in Need (takes place this Friday) when certain domains are registered or renewed.

For every .com, org and .net domain that is renewed or registered for the first time with them befire the end fi Friday will qualify for a donation.

Google Page Rank

I mentioned yesterday about my confusion with Google's Page Rank. I left the post a little open at the end as it was dragging on a bit.

What I noticed through my new Promotional Items website was that all but 1 of the second level of pages was cached with a Google Page Rank 2. The missing one was cached, but grey barred (there are now more pages like this – they have been added recently).

It's been confounding me as to why that one link page didn't have a page rank. The page has existed as long as the rest and there's never been a problem with its links. So what is different about it?

I tried something different when I set up the navigation on this site. I actually reduced the amount of optimisation for 2 reasons. First, it looks better, but also just in case Google thought it too much. So I purposely did not include whole keyword phrases in every anchor text. This worked fine, except for one link in which it still came out as a popular keyword phrase.

You can no doubt guess which link this was!

And guess how I have been optimising the mortgage site – that's right, by using keyword phrases in the anchor text. Of all of the links that appear on every page only 1 isn't a popular keyword phrase (and most are actually the most valid descriptions of the page – I'm not spamming). And that page is the only other Google Page Rank 2 page on the site.

A bit more investigation is needed here!!!

And Still Page Rank Confounds Me!!!

I was looking through my Compare Mortgage Rates site yesterday to see which pages were cached on Google and which had a page rank.

Most of the pages are cached, but with a grey bar. Which is strange - I always thought grey bar meant not cached on Google???

The home page and one other hold a very disappointing PR2 (was 5 a few updates ago). 14 pages are page rank 1, even though these are only linked to from the 2 PR2 pages. There's also a PR0 page that has not been touched or linked to for about 2 years!

The strange thing is that the main pages - those that are linked to from every other page on the website, are all grey barred (but cached). Whether my hosts have had some regular problem around the time of the page rank update that I was not aware of seems an unlikely explanation. More likely is the positioning of the links on the page.

I've realised that the links on every page to the main pages form a nice neat column. But those that Google seems to be following are a paragraph. The one PR2 internal page is not only in the column of links but also included in the text.

I've noticed on other sites using a similar layout that the home page is PR4, but the internal pages are not page ranked.

I'm starting to wonder if Google is trying to detect link exchanges by using some algorithm along the lines of if the link is on a line of it's own delimited by <br> then ignore it.

A lot of link exchanges would be detected this way and it would throw weight into proper links within text. And looking at various sites that link to a well performing site, not one is using the line break. But it means that internal links structures in a vertical format are penalised.

It's also interesting to note that the one links scheme that I notice always does very well allows the link to be placed mid paragraph.

It wouldn't fully explain all of the strange results I'm seeing, and another new site of mine (Promotional Gifts) seems to go against the theory. It has a very similar row of links, it's home page is PR3 and all but one of the 2nd level pages are PR2. Only 1 2nd level page is grey barred. Something else stands out about that link. But that's enough for today...

Your comments and observations are welcome. Please post a comment!

Week 4 results for target="_blank" experiment

I mentioned in the week 3 results of by SEO Experiment that something strange happened on day 21. I didn't want to say what had happened to see if a pattern formed, and it certainly did.

First, it's interesting to note that although both blogs that are linked to have links on the same places, Google is reporting that blog 2 has 7 links, blog 3 only has 6 links. It seems that 1 of the links was missed by Google.

So, we might expect site 1, with no links, to be least popular, site 2 with most links to be most popular and site 3, with almost the same number of links to have almost the same popularity. But do search engines ever do what we expect?

Well on Yahoo, site 1 is ignored - it isn't getting the pings. But sites 2 & 3 are similar. But then I never questioned whether it took note of the target="_blank" statement.

On Google, here's the latest days' results for number of pages cached...

Day 20 - Site 1 still has 3 pages cached, site 2 8 pages but site 3 suddenly takes the lead with 9.

Days 21 & 22 were the same as day 20.

Day 23 and site 2 lost a page.

Day 24 and site 1 increases to 4 pages, 2 drops to 7 and 3 drops to 8. Interesting to note that site 2 was mostly cached end of October, site 3 beginning of November.

Day 25 and sites 2 & 3 both lose another page.

Day 26 no change.

Day 27 (today) and site 2 has dropped to 2 pages.

I can't explain why the site with no links has 4 pages cached, the site with most links has just 2 pages and the one in the middle has 7 pages and that's the only one to have been cached recently.

So as far as Google is concerned, the site with no links is doing better than the one with most links. It could be that because of the better links site 2 was cached and dealt with earlier and is showing a pattern the other 2 sites will show but 8 - 10 days or more further behind. I'll keep watching...

One Way Link Building

Yesterday I talked about one way link building. Can I just point out that for the forums I was not suggesting spamming! All that I meant was that if you already actively take part in forums then change your signature to include your main (relevant) website address if you haven't already done so. Please don't go posting just to link to your sites - these places are far too clever for that and detect this almost instantly.

Right, now I've got that out of the way, how can you get loads of links from blogs? The obvious answer would be to post comments on blogs, but most blogs (sensibly) use the rel="nofollow" tag to prevent this. And I agree with this.

What I have noticed recently is schemes that allow bloggers to be paid for mentioning and linking to other websites. And this is what I think the guy I was looking at was doing with his website. It's obviously a great way to get one way links, but never having tried it I cannot say whether or not it brings much success.

To be able to be paid to blog you must have a blog of 90 days old. Unfortunately, this blog is just under that, so I can't sign up and find out what it's all about for a couple of weeks.

Be assured when this blog hits 90 days old I will be doing so and telling more about it!

The New Link Building

So just what was this guy doing in yesterday's post that the millionaires and big sites had missed out on? How can a shoe string budget site so well trample sites with thousands, even tens of thousands in marketing budgets? What can we learn for our own sites, like my Compare Mortgage Rates site?

The answer was remarkably simple. I looked through the first couple of pages of links to his site as reported by Google to see what he was doing. I figured that if Google lists these links first, they are the most important links. What I noticed was that there were loads of links to his site from 2 main sources. Neither of them link exchanges:

1) From internet forums. He was posting in discussions, sometimes related sometimes not and in his signature linked directly to his home page.

2) There were lots of blogs (just like this one) pointing to him.

The forums seems easy to do - sign up, make sufficient posts and survive the trial period then continue to post and include your URL in the signature. Start with one or two, as you get the hang of it join more.

The blogs aren't so easy. There were too many for them all to be his doing so what I suspect he's doing there is paying people to mention and link to his site in their posts. As look would have it, I'd already started to investigate this before I saw his links, so I recognised what he was doing.

I've started investigating one such scheme, I'll let you know more tomorrow.

Google Is Getting Cleverer...

Yesterday I was talking with a customer about link building and general search engine optimisation.

We were looking at various sites, including my Compare Mortgage Rates site and working out why that had dropped. Looking through Google, it's 550+ links that I've checked are still in place are reported as only 8 links. This massive drop is now been blamed for the site's massive drop.

We then went on to look at my customer's other site and his competitors. He named the top site traffic wise (out of his research) and it was a big name I recognised, using affiliate links, ppc and a huge warehouse setup. But the surprise was the second placed site. This was a guy he knows who runs the business from a spare bedroom with stock in a friend's barn. He wasn't a rich man when he set up in business, just very clever at what he does. He's beating the next best sites we looked at by miles - and these are millionaires happy to throw their pounds into marketing their sites.

When we started to look at how he was getting his site to the top of the listings - the tricks he was using. It didn't take us long to see that he was using techniques that I'd heard about before, just maybe not invested enough time in trying them out.

But, on the whole, he wasn't wholesale link building. We looked at another site I maintain (for a customer) and of his 600+ links only 12 show on Google. Less than 4%. Seems that Google is getting good at detecting link building and just ignoring it.

So in a day or so I'll at what this guy was doing and what we can all learn from him. And I don't think the answer is spamming people for link building. Gone are the days of Ms Yeager & co sending me masses of emails telling me they've just visited my site and how good it looks....

Flash Web Design

I talked yesterday about including flash in web sites and I'm all for it - in moderation. But is there ever a time when a complete flash web design is called for, or at least can be justified?

I suppose that if you aren't in need of search engine traffic finding your site, your users are happy to wait whilst it loads and you need a lot of graphical displays then yes, flash web design is useful.

But what about people without flash, or accessibility. Again, I suppose there's times when these may be overlooked - I can't think of an example, but there's bound to be someone to argue the case. But aside from that you should consider a non flash site alongside the flash site. This means that you are producing twice the work.

In this case, would it not be better to have some of the site as standard HTML and just link into flash elements as required?

That's my reasoning behind one of the sites I'm currently working on - when more material comes through. It needs to look spectacular, but be search engine friendly. So most of the pages are HTML, but the products go to flash web design instead. It's not an e-commerce site, just showing off the products. So this combination works great.

But I'm not convinced there is a reason for building a totally flash web site.

To Flash or Not To Flash???

Should you include flash on a website or not? That's a big decision and opinion varies a lot. Some people are horrified at the thought of having flash on their website, others consider it essential.

I think it's like chocolate and wine. Both of these we are told are good for you - in moderation. Have too much and you'll make yourself ill. Abstain and you are missing out on the feel good factor and a few beneficial properties.

The same goes for flash. If you build an entire site in flash the search engines won't be interested. But if you include a little snippet or two, in small files, then the effect can be worth the effort.

There is the danger that this is rapidly take to the extreme. One customer saw that I'd put in a small flash animation and insisted that the size was increased and the number of frames doubled - then was horrified when my prediction of a slow to open page came true.

Everything in moderation - including flash!

Week 3 Results

Time for a recap of the latest results on the SEO Experiment.

Recap: The Second Week ended with the first site - that with absolutely no links, not cached. Site 2 - with normal links, had 8 + pages cached and several pages with PR0 whilst site 3 - linked using target="_blank" had only the home page cached. In that case the home page had a PR0 and the rest were grey barred.

This week: A week of changes!
Day 15 - site 3 was also cached with one of my own blogs (the majority are syndicated articles).

Day 16 - Site 2 lost a page - down to just 7 pages cached, whilst site 3 was up to 5 pages. At this point I thought things were balancing out...

Day 17 - Site 1 appeared again, with it's home page cached; site 2 was back to 8 pages and 3 had 5 pages. I also added new posts across all 3 blogs - all posts similar themes but different writing. Monitoring when these posts appear, but (2) and (3) did show on Google an hour later as being updated.

Day 18 - Was the same as 17.

Day 19 - Site 2 also had an archive page, but this was identified as duplicate content. Site 3 is up to 6 pages cached.

Day 20 - Something strange has happened. I'll leave this until next week when I've had time to watch what else follows....

Basically, after 19 days:
Site 1 - no links - just the home page is appearing.
Site 2 - links - all posts of over a week old appearing (cached about a week ago).
Site 3 - links, but target="_blank" - home and a couple of posts appearing.

There is, at the moment, a difference between the 3 sites. Site 1 is being cached - because of the pings. 2 & 3 are showing a lot more search engine interest, with 2 slightly more interest and better page ranks.

At the moment, it does look like for a new site the target="_blank" does affect the link. Not completely like other ways of blocking the link. But from the results to day 19, I'd much rather be in site 2's position.

Customers Are Sent To Try Us

Sometimes it seems that customers are sent to try us out! Yesterday I had just finished publishing a new website and was in the process of handing it over when the question came - could we change the background colour of the links.

And change the colour of another part of the screen.

And move the links to another part of the screen.

And change the entire menu structure of the website.

And change all of the graphics on the home page.

And change the text in some buttons.

You get the idea. The back end of the site stays the same but the entire website look and feel is to be changed. How we managed to get so far through the project without him raising these issues I just don't know. We were passing latest samples of the site to the customer a couple of times per day and small changes were coming back, but how it becomes an entire restyle the day it's published, I just don't know!

It's back to the drawing board for that site. It will go live again in a few weeks!

When Sites Get Dropped

Have you ever suffered pages, even an entire site, dropping from Google?

At the beginning of October I wrote about a site that totally vanished from Google and I suspected Duplicate Content Filters.

The background - I'd published a one page website for a customer and then about 60 to 75% of the page content was duplicated onto a local business directory. A short time later, the new site vanished from Google. I checked links in, did some links building etc, but week after week I couldn't get the site listed again.

It was then that I discovered the directory listing. The text on the listing was changed, which didn't take too long, then it was left for time to take it's course.

A couple of weeks later and Google had cached the new version of the directory page and magically, the new site had reappeared in the listings.

This teaches us 2 important lessons:

1 - don't copy someone else's text - make sure your text is original, no matter what it takes.

2 - if you are creating a listing somewhere - use new text! You might be doing more harm than good if the directory is older than your website.

October Roundup

October seemed to be one of those months when I just don't stop, but have very little to show for it. Hardly any new sites went live during October, but hopefully that's because there are loads to do over the next couple of weeks.

The first goes live later today and another customer is desperate to be live within 10 days and others are just finalising sites. Looks like November could be a very busy month for publishing!

My site that's dropped down the rankings is still hovering between 5th & 7th pages, 8 days later. Customers who have experienced the same I managed to recover in a week - so why not my own site! It's infuriating. There's no obvious reason for it to have dropped mid week. Just one of those things, I suppose.