| 5:57 pm on Apr 2, 2011 (gmt 0)|
Not only Google - all search engines struggle with the challenges of big drop-down menus. For example, that many links will disperse link equity too evenly around the website. In addition, the same keyword anchor text appears on every page, blurring the semantic relevance signals.
There are very real usability problems too. In my experience, it's only the regular visitors to a site that appreciate drop-down menus with many entries because they've "leaarned" the site. Unless the site is an online application used all the time, that usually means the business's STAFF like it, but not their customers. Certainly it boggles the new visitor.
With extensive drop down menus, there is no way for the visitor to compare all their navigational options at one time - they are not visible at once. Too much short-term memory is involved in trying to decide where to go.
I had the opportunity to test removing a mega drop-down menu on one site. Page-views per visitor more than doubled without the drop-down.
| 6:41 pm on Apr 2, 2011 (gmt 0)|
I find mega drop-downs actually increase page views.
But I'm definitely starting to think that, at least since Panda, the number of internal links per page is a big SERP factor. So it may hurt you, even if it improves user experience.
Google doesn't really care about user experience. They say build a site for your users, but there's no way they can really know what's good for your users. What they mean is build a site which fits what they believe is good for your users. Very different notion.
| 8:21 pm on Apr 2, 2011 (gmt 0)|
I noticed in the "winners and losers" discussed in this thread...
Winners and Losers in New Google Algo - Analysis
...that a higher proportion of the losers appeared to be using mega-menus than were the winners. This is an oversimplification, but overall, for sites of comparable size, that's what I'd noted.
| 8:51 pm on Apr 2, 2011 (gmt 0)|
|I find mega drop-downs actually increase page views. |
Horses for courses, then. How many total links are involved in that drop-down menu? The site I was talking about had two levels for something like 500 total navigation links on every page.
Another problematic result was poorly target search traffic - because Google was seeing the anchor text in the drop-down items as being potentially relevant to the page itself. But the visitor would land on the page and at least some of the words in their search were not visible unless they hovered on the menu.
| 8:59 pm on Apr 2, 2011 (gmt 0)|
"...that a higher proportion of the losers appeared to be using mega-menus than were the winners. This is an oversimplification, but overall, for sites of comparable size, that's what I'd noted. "
Had quick look at two winners
ehow. No drop down menus but well over a hundred links from the Home Page.
Linked In uses drop down menus.
| 9:46 pm on Apr 2, 2011 (gmt 0)|
Perhaps we should get a count of the number of links in drop-down menus for sites that have been hit versus sites that have not.
I have a navigation bar at the top with a total of ten links. One of the links will take the user to a products page that has links to pages for different manufacturers products. Clicking on that link also opens a drop-down menu with 36 links for the same manufacturers. So, users have both.
Within those 36 link are a three that have multiple categories. So, clicking on Acme Widgets opens another menu with Red Widgets, Green Widgets and Blue Widgets.
It would be interesting to see if other sites with this many (or this few) links in drop-down menus were affected.
It seems like a cleaner way to navigate. My average number of page views declined when I implemented this new design, but that's because the user didn't have to go to the same page with the list of manufacturers each time he wanted to look at something else.
| 10:15 pm on Apr 2, 2011 (gmt 0)|
Also makes it easier to maintain.
By taking all the links out of the equation, it also leaves you free to add the important links contextually or in a sidebar nav thus increasing their priority.
| 10:34 pm on Apr 2, 2011 (gmt 0)|
I've seen that approach, but I'm not comfortable with it. With that method your standard navigation links are not in the source code of every URL, only the iframed URL. I think that situation seriously impacts the circulation of link equity throughout the entire website.
| 10:49 pm on Apr 2, 2011 (gmt 0)|
>>Horses for courses, then. How many total links are involved in that drop-down menu? The site I was talking about had two levels for something like 500 total navigation links on every page.
Totally agree. There is a way to do it and a way not to do it. They can be really bad, but done right they can also be good.
The problem is Google doesn't differentiate, it has just decided they're all bad (assuming that's a factor... as always it's all just theories).
The more we talk about this the more it seems like the real problem with Panda overall is that it's not nuanced at ALL. Past algorithms were. It's all black and white. And that's not really the way the web works.
| 10:51 pm on Apr 2, 2011 (gmt 0)|
@dickbaker What you're describing sounds like exactly what I had, and I was penalized.
It may be a factor. Of course it's probably not the ONLY factor. So a lot of sites with them might be fine, others won't because Panda found that + 2 other things and then just said "off" to that site.
Which makes it impossible to determine if it really is factor.
It was one of the first things I changed post-Panda though and I no longer have dropdowns at all, and I have seen no improvement.
| 10:53 pm on Apr 2, 2011 (gmt 0)|
|because Google was seeing the anchor text in the drop-down items as being potentially relevant to the page itself. But the visitor would land on the page and at least some of the words in their search were not visible unless they hovered on the menu |
My biggest problem with drop downs as a searcher. Why is this page supposedly relevant?
| 11:14 pm on Apr 2, 2011 (gmt 0)|
Looking at the BBC that has a PR of 9 and for sure is white listed anyway.
The top navigation menu consists of 7 non-drop down menus - each leading you to a directory from where the browser branches off. The last menu item at the top is a drop down with 15 menu links, followed by a search box.
The bottom is littered with about 100 non drop down menus leading to sub directories.
Thus overall they have 100 plus non drop down menus, and ONE drop down menu on the home page.
| 11:16 pm on Apr 2, 2011 (gmt 0)|
I want to clarify something - I'm NOT just talking about the Panda Update. The problems I mentioned have existed for a long time, and with every search engine. They're all coping as best they can and things are marginally better than they were 2-3 years ago.
I don't think Google is intentionally targeting drop-down menus - that would be completely nuts. Google may seem somewhat nuts at times, but they're not completely nuts ;)
But those menus can often make any search engine's job harder. And that means some content may not rank as well as it could, and other content may rank well for poorly targeted queries.
| 11:46 pm on Apr 2, 2011 (gmt 0)|
|But those menus can often make any search engine's job harder. And that means some content may not rank as well as it could, and other content may rank well for poorly targeted queries. |
How would those make the SE's jobs harder? If you have a table or other structure with a bunch of <a href=""> tags, can't the SE's follow that?
I'm not trying to dispute what you're saying, just trying to understand where the difficulty would lie.
| 12:01 am on Apr 3, 2011 (gmt 0)|
@tedster - there may be something to this.
One site where we use top notch drop down menus that degrade gracefully without JS, we noticed in WMT that the single menu item WITHOUT a dropdown got bumped up substantially.
I was scratching my head because this menu goes to an old directory with some links that then go to old and outdated information we have not touched in at least six months.
I think we will dance a little with G and turn a couple of the drop downs to no drop down and gain a boost for free...easy does it, nothing drastic so as not to alarm the beast <GRIN>
| 1:27 am on Apr 3, 2011 (gmt 0)|
|just trying to understand where the difficulty would lie |
At its base, the idea of PageRank (or any kind of link equity metric) measures how "important" each page is relative to other pages. By linking everything to everything, the site is blurring an important signal about the relative hierarchy of its pages.
And secondarily, when it comes to query relevance, hover menus are placing keywords for every page on every other page - and they are often in anchor text which does figure as an on-page factor as well as a target-page factor.
As I said, Google and other engines are coping with this challenge to a degree these days.
One moment of truth came for me when I tracked which menu links were actually being used by site visitors. That analysis showed me that all those hover menu options were NOT being used by visitors. Only a small subset mattered, and there were other ways to help people find those pages.
I appreciate that my point-of-view kind of upsets the apple cart for many people. A Pubcon session where I presented my findings was called "controversial" by one blogger.
But it is not merely an opinion, it is based on data analysis of real sites. The OP asks "can a mega drop down menu create problems." I'm answering with very real problems I've seen that stemmed from using a large drop down menu system.
| 1:46 am on Apr 3, 2011 (gmt 0)|
"A Pubcon session where I presented my findings was called "controversial" by one blogger. "
Where is the link to your paper?
It would be easier to digest and analyze your thesis if it were presented somewhat like what Mr. Sullivan does - a full article with citations and so on....
| 2:54 am on Apr 3, 2011 (gmt 0)|
Sorry, it's not online ;( My complete Pubcon session was about Information Architecture, and my hover menu test and findings were just a small part of it. What I learned certainly may not apply to every site - especially when the hover menus are designed economically and with the visitor's real needs in mind. But still, given what I've seen, the more links in your hover menu system, the more problems you are likely to be generating.
The original recommendation from Google was 100 links per page - in the beginning that was a limit to what they could spider. That limit has since been softened a bit as they've upgraded their technology - but the basic idea of not flooding any page with a lot of links still holds true. And nothing floods the source code with links (that you forget about) like a drop down hover menu. It becomes like a junk drawer and a way to avoid the hard work of creating a truly useful IA.
| 2:59 am on Apr 3, 2011 (gmt 0)|
Here's a more complete treatment of the mega menu challenge. The thread is from 2008 when I first realized the problems it was creating. The "Mega Menu" Problem and Google Rankings [webmasterworld.com]
| 8:48 am on Apr 3, 2011 (gmt 0)|
What if those dropdown menu's are limited to specific subjects? So you link every page to every other page about the same subject? Would it still be bad for SE?
Are links in a dropdownmenu with <a href="#">link</a> (a hashtag) links that will spoil pagejuice?
And if dropdownlinks aren't used by users because they first have to hover, how would you setup any navigational menu's if you have a lot of pages?
What do you consider a site with a good site architecture / internal linking?
| 3:27 pm on Apr 3, 2011 (gmt 0)|
Is there a Robot txt file one can use dissallowing the SE from indexing the Mega Menu?
| 5:18 pm on Apr 3, 2011 (gmt 0)|
The easiest way to describe good Information Architecture (and ultimately it is NOT an easy job, it is very difficult) is that IA serves the visitor above all. That is, the visitors' needs are what matter most, not the business structure, pet ideas or even what someone thinks Google wants.
Sure, the website wouldn't exist if the business didn't have something to push, and it wouldn't get far by completely ignoring search engine requirements, but those should be quite secondary. SEO comes after Information Architecture. They can blend, but if SEO dictates the whole thing, then the site tends toward failure with its visitors.
Information Architecture is a field that pre-dates the web and it originates with Library Science. It is not trendy and the foundations don't shift much as the years go by. There's a series of threads from 2004 that begins here: Information Architecture for the Small Site - part 1 [webmasterworld.com]
| 6:49 pm on Apr 4, 2011 (gmt 0)|
I guess the large Brand Names would not be affected by a Mega Menu..
One solution might be using a 'flyout' on all the inner pages. Anyone have thoughts on this?
| 7:04 pm on Apr 4, 2011 (gmt 0)|
Why not use a combination of JS and AJAX? Essentially you load the menu items as needed while people hover. The amount of information being transferred is small so the user won't even notice it. This way you get the UX benefits while not harming your SEO.
Win/Win, which are the best kind of wins.
| 8:03 pm on Apr 4, 2011 (gmt 0)|
|I guess the large Brand Names would not be affected by a Mega Menu |
I work with some very large brand names, and I can say for certain that they ARE affected. Discovering that is a big part of why I wrote the original post on this topic.
Branded search terms are never the problem for big brands, but ranking well for generic search phrases most definitely is. Mega hover menus have definitely confused Google in this area, and also for certain trademark searches that the big brand actually owns.
You may think you'd like all that backlink power and apparent "special treatment" that big brands get in search. But once you see the challenges of big brand SEO, you might be very happy to be running a fleet of affiliate squeeze pages ;)
| 11:21 am on Apr 5, 2011 (gmt 0)|
ismailman: I recently implemented menus that AJAX load on hover for one of my sites. In our case, the menu items were not where we wanted our pagerank going, we have other links elsewhere on the page for that already.
If you do implemented AJAX on hover or iframe for the menu items, you need to make sure that you have SEO links elsewhere.
| 4:03 pm on Apr 5, 2011 (gmt 0)|
Did the AJAX menu help you to get the pagerank where you wanted it to be?
| 12:12 pm on Apr 6, 2011 (gmt 0)|
It certainly means that the pagerank is no longer in those submenu item pages. In our case, they were various marketing initiatives, account related pages, and company info pages. None of which have search volume associated with their keywords. If you do site: searches on google for the site, it no longer pulls those pages up.
Presumably, the pagerank is better directed at our product pages that are linked further down the page. I don't have any evidence that we increased rankings or traffic to our product pages because of the menu change though.