Site Architecture: How flat is too flat?

[GalcoIndustrial]

There's a lot of debate out there as far as too many internal links or too many levels of a website.  I've seen the videos from Rand and I've read a lot of the posts here on Moz, but I just want to know where everyone stands on this.  Anyone have experience with architecture while working on a large E-commerce site? We're talking Millions of pages and over 1,000 links off the homepage alone.

Anyways, I don't want to get too specific.  I mostly just want to hear about experiences of the community.

Thanks!

[DonnaDuncan]

I hear ya! I find one of the biggest issues ecommerce sites have with rankings has to do with their data. That and a lack of canonicalization to help Google figure out which are the money pages.

[GalcoIndustrial]

We first rolled out the Homepage link reduction project back in July of 2013.  I was hoping to make it a full year, but it seems as though the higher-ups are ready to move on (seeing as though it's been close to 10 months, I can't really blame them).

However, I should mention that we did see a hefty increase in the overall number of pages in Google's index. It's really too bad that those indexed pages never really converted in to web traffic, but I would say that we have some site clean-up to perform and some page quality issues we should be checking in to as well.

I suppose it's nice to know that when we end up making the proper changes to our site content/quality, we could try and adjust our site architecture again (and this time we'll have some figures to benchmark against).

Thanks again!

[DonnaDuncan]

Hey Gerry,

How much time elapsed between implementation of new directories and did you put redirects in place to ensure you didn't lose any link equity in the process? If I were you, I'd give it several months before I passed judgment. Also, you might not feel a benefit if traffic and links are low, but will later down the line as both of those things pick up.

D

[GalcoIndustrial]

Thanks for getting back to me, Donna!  I would have thought the same thing as what you mentioned, however, we programmed a bit of a workaround to reduce the links on the homepage, yet, putting one more step between the homepage and the product pages (i.e. if the product pages used to be 3 steps away, but with over 1,000 links on the homepage, we put the product pages 4 steps away and reduced the total number of homepage links to under 300).  The results weren't terribly affective (effective?), but the reduction in the homepage links did not boost traffic at all.  In fact, there was an almost immediate drop (although, not too significant) in traffic across the board.

There's no doubt in my mind that the workaround we implemented was not the cleanest way to roll out the changes, but I would have thought we would see just a slight change in traffic - in the right direction, that is.

Either way, thanks again for getting back to me.  We've decided to rollback the changes that were made so I'll be sure to keep you updated on the trends we see afterwards.

Regards,

Gerry Mendiola

[DonnaDuncan]

I have experience with a very large ecommerce site with hundreds of thousands of pages and a completely flat architecture. It's not good. Think of the distribution of your link equity like you would an organ donation. Would you rather give two kidneys to two people or 200,000? Which choice is going to make a difference to the recipients?