Forum Moderators: Robert Charlton & goodroi
The site was already successful in their niche - and it's pretty competitive, with 400 million Google results on the big trophy keyword. The pages looked sharp, but the main navigation (top and left) was created with images of text with alt attributes, not anchor text. And javascript rollovers, naturally.
The only change so far was to switch this setup to CSS styled text menus (links in list elements) a few weeks back. And funny thing, the look is so close to the original that no end user or affiliate has yet commented on the change. Some of the business staff haven't even noticed.
At any rate, with the Google changes on Sep 15, we got our first look at the new rankings. On the trophy keyword, the home page went up about 45 positions, from out of the running to a second page contender. With lots more improvements to go, this picture is quite hopeful. The power of anchor text on the main navigation turned out to be BIG.
Lots of other major keywords went to #1 or into the top 3, from previously being 10, 20, or more positions down. In all, it looks like 200 phrases have improved from the benchmark sample we are tracking. I'm sure as the new conversions start to show up, we'll learn about lots more keywords that are worth tracking and tweaking, too.
Lesson for me -- don't let the issue slide when it comes to using text in main navigvation. It's a lot more important than I used to think, and I used to think it was pretty important anyway.
[edited by: tedster at 1:38 am (utc) on Sep. 19, 2006]
The power of anchor text on the main navigation turned out to be BIG.
Absolutely...there is nothing quite like having the spider see your most important top level navigation links (and associate anchor text) throughout your site in a consistent available "crawlable" fashion..(not to mention how this affects in a positive way the actual human usability of the site...)
The "hypertext" link (and associated anchor text) came before the search box and is still one of the "main root" connection points for information online....
This is definite proof that the egg came before the chicken... ;-)
egg came before the chicken
So now the only picture on the home page is the logo :c)
A mid-level navigation page started ranking #1 for a keyphrase that brought in more traffic than any other for several months.
There were no external links to that page. Internally, every page above it, including the home page, linked to it with that term consistently. Every page below it linked back to it with breadcrumbs.
On the page, the title and file name contained the term. That was it. So it was definitely the navigation that gave this page its ranking for that fairly competitive term. Years later, it has never left that front page.
One advantage that internal anchor text has, is that Google seems to recognize it as navigation and does not consider the repetitiveness of the anchor text to be a bad thing.
I have some sites where I use java script rollover navigation images.
HOWEVER, I also always use plain text links on the bottom of the page.
But lately, I've been switching to CCS generated navigation (but still with bottom of the page text links)
Do you think it would be worth it for me to go back and convert all of the sites to CSS (and remove the rollovers) given that they have the bottom of the page text links?
I think that a test case analyzing one site is a small sample and it's not possible to draw an accurate conclusion
Uh, a "case study", by definition is the analysis of a single case.
The problem is not that you cannot draw an accurate conclusion, but that you must acknowlege the limitations of the study, just as you must do with any study. It is standard research methodology.
Of course, there should also be citations of other supporting and dissenting research by others, if it was real research. But it was just a post, believe it or don't. Or even better, do some of your own research and report back.
We had 3-4 major links at the top of the page that used to be text and I made those into images (that look like the text) so that msn wouldn't do that, however we do have a side menu in css which is all text and does have those same links in it.
We do have a few links at the bottom where users expect to find them .. sitemap, contact etc .. and now I'm finding that msn is using some of those for a description instead of the page copy .. go figure.
on a site I worked with they wouldn't let me do anything but their graphics were so bad they wouldn't notice if I did exactly the same thing with text and a little CSS magic.
worked like a charm, larger jump than your case Ted but a much smaller set of G results.
it's just the power of anchor text, we know it works, and we also know internals get play but we sometimes don't realize quite how much.
Uh, a "case study", by definition is the analysis of a single case.
I never said that a "case study" is NOT the analysis of a single site, in fact I said the opposite, so not sure what you mean
But it was just a post, believe it or don't. Or even better, do some of your own research and report back.
sure it's just a post, just like mine was just a post, I'm just saying that "IMO" there could be several reasons for a jump
Question: what if the client wants to keep the images? I once put text next to the image and hide it with CSS (text-indent: -1000px). The text is telling the same thing as the images, so I think it's a white hat technique.
I think it's a white hat technique
thanks for the case study Tedster!
have you ever tried SIFR?
Internal anchor text can be very powerful indead. All of the corporate sites I look after are changing their look and feel this month and are all moving to a text based navigation system.
Also, so many sites are moving to text based navigation systems these days.
Internal anchor text is the most powerful seo weapon at your disposal. I simply cannot believe the amount of savvy seo's that overlook this simple method.
After your inbounds are there (a nice mix) apply some decent themeing and appropriate anchors to your navigation and your apples. Also don't just build links at your homepage, deeplink your site. Get those internal sections you deliberated over so much some decent PR and attention.
Information Architecture
[webmasterworld.com...]
[webmasterworld.com...]
[webmasterworld.com...]
Theming
[webmasterworld.com...]
[webmasterworld.com...]
Some are kicking around the title seo siloing, but to me it's theming 202, although siloing does have a nice mental connotation to it.
- Ben
[edited by: Bennie at 7:55 am (utc) on Sep. 19, 2006]
Thanks Tedster.This reminds me to get around to changing a sites homne page that I have where the first letter of each internal anchor text has a 'strong' a 'font'and 'color' attribute for purposes of decoration.
You could use a pseudo-class in CSS to change the appearance of your first characters without breaking the strings. Won't work with all browsers, but might be worth a look.
nerd.
But just how much more powerful is it?
Quite a bit as evidenced by your case study. And I'll add some of my experience to help solidify yours.
Over the past couple of years, I've been doing quite a few site reviews. I have a scoring system that I use based on years of experience and following many topics here at WebmasterWorld and other communities that I participate in.
I start off with a Franklin T. On the left, I list those items about the site that are positive. On the right I list those items about the site that are negative. From that list, I come up with a quality score.
For each positive, I add a point. For each negative, I deduct a point. In reference to text navigation, it's at the top of my list. If I see a graphic navigation, I subtract a point. If alt attributes are not being used, I subtract another point. If title attributes are not being used on those linked navigation images, I'll subtract another point. And, if there are alt and title attributes, I'll add a point for each. But, I'll deduct a point if they are not mirrors of the graphic elements.
Internal anchor text is one of those areas where you really need to focus your efforts. Think of it as "providing links to your own site". They have power and lots of it if used properly.
The one thing that really powers internal anchor text is the site architecture. I've found it very helpful to have a visual map of linking structure to determine the path a spider make take when reaching one of my sites. Brett's topic on pyramids comes to mind.
Plain text will trump whatever may be in an image. It always has and I think it always will. You can make a graphic link usable and accessible, but it will never carry the same weight as an actual text link. Look at it this way, if you were a spider, which of the two scenarios provides you the shortest and most succinct description of the link you are following?
<li><a href="http://www.example.com/parts/">Parts</a></li>
or...
<div><a title="View our Parts Catalog" href="http://www.example.com/parts/"><img src="/images/parts.jpg" alt="Parts Catalog"></a></div>
Using a list element for the links -- beats links separated only by <br> tags.
Personally, I follow what the W3 Guidelines recommend. Most links that are grouped together qualify for a list element as they are typically navigation style links. You have both vertical and horizontal list layouts.
Many don't use <li> because they may not be familiar with CSS and how to control the look and feel of those list elements. It's easier to do the <br> thing. ;)
Personally, I think items wrapped inside certain containing elements have intrinsic value. For example, if I'm creating a glossary, I'm surely going to use the <dl><dd><dt> elements. The same applies to a list of links. I'm surely going to wrap those in a <ul> or <ol> element. ;)
I like the <ol> element as it makes site assistance much easier. Having a user go to link 22 in the menu is much easier than having them go to "the link half way down, right under that one, yes, that's it.". But, if your menu structure is divided using <h> elements and other visual separators, the <ol> is not required and definitely not a common practice.
I have been struggleing with this a lot lately. I have a few pages that rank lower than the rest of my (usually well ranked) site. They are the most competitive terms and the top ranking sites have run-of-site links back to their page. On my site these have the most internal links. Initially I created all those links in emulation of the top ranking pages but I suspect it's hurting more than helping. It appears that it takes a long time for these changes to take affect (either adding or removing links). What do you look for to check if you've got too many links; or if a change you've made is in effect (e.g. "link:www.xyz.com/page_i_care_about.html")?
It would seem bizarre (and confusing to the user) to go through my site and change the text in my breadcrumbs.
Links in HTML documents
[w3.org...]
I don't do 22 links in one menu.
lol! I try not too but there are instances where secondary navigation menus have upwards of 25-50 links.
Excellent!
Tedster is talking about repeating KW in the anchor text, BigDave about repeating the anchor text itself.
The latter is usually safe unless it includes the former ;-)
Links in HTML documents
[w3.org...]por - that was VERY interesting..I must admit, I've not seen the link usage before..that might be very handy for some of my rich content sites..
A good CSS designer can do a lot with an unordered list, it's surprising to see what they can pull off given some pretty boring HTML underneath.
Unfortunately I don't have a good before-and-after case study, because all those sites are getting constant updates and have consistently climbed ranks, so it's pointless to presume what the winning factors were. But I do know that the sites usually get crawled top-down, are well indexed, and have a good top-heavy link structure.
I haven't made a Fireworks-esque images-and-rollovers menu since... about 2002. I do not miss them at all.
