How would you handle 12,000 "tag" pages on Wordpress site?

[M_D_Golden_Peak]

We have a Wordpress site where /tag/ pages were not set to "noindex" and they are driving 25% of site's traffic (roughly 100,000 visits year to date). We can't simply "noindex" them all now, or we'll lose a massive amount of traffic. We can't possibly write unique descriptions for all of them. We can't just do nothing or a Panda update will come by and ding us for duplicate content one day (surprised it hasn't already). What would you do?

[Klarke]

I would "noindex,follow" them.  Don't block them with robots.txt.
With that many pages, you're certainly running the risk of being hit by Panda.

Those tag pages shouldn't be ranking, instead the individual posts should be in those positions.  If I were you, I would take the chance and do the noindex, with the expectation that Google will appropriately rank the posts in their place.

I'd say those are better odds as against losing 50 - 80% of traffic in a panda update.

[evolvingSEO]

Hi

I'm actually going to be addressing this exact question on a post for Moz in the coming weeks - so keep an eye out for that.

But in short, here's what I do;

Analytics

1. run a report for landing tag pages (with a filter) - over the last three months
2. apply an advanced segment to see google only traffic
3. dump the report into a CSV

Webmaster Tools

1. view a impressions / clicks report by top pages (not keyword) - also zoom out as far as you can
2. filter for web only (not images)
3. dump the report into a csv

VLookup in Excel

using a VLookup in excel - combine the two reports matching data to the URLs (you'll end up discarding some non-tag pages from wmt) - the end result will be a master spreadsheet, with the following columns;

• URL
• Impressions
• clicks
• avg position
• visits
• pages/visit
• avg visit duration
• % new visits
• bounce rate

(These are all the default report metrics. I actually prefer a custom landing page report in analytics, but this works fine.)

Analyze

Then, you do your sorting, filtering etc - to decide how valuable the tag traffic has been. In general, you're trying to look for an overwhelming reason for the value add of having those pages in there. they might get visits, but what's onsite behavior? maybe they get visits, but perhaps only from a small handle of tag pages?

In the post I do, I'll cover more about how to analyze this report etc.

As Klarke put so well, the actual posts should rank in their place. Those tend to have better results when people land on those.

Remove

If you decide to remove, do so carefully. Do it on a weekend or just before a downtime. If you use Yoast simply select to noindex tag archives.

Also, rememeber to exclude tags from your XML sitemap.

Then watch webmaster tools etc and watch for their removal.

--- I did this process on a site with 9,000 tag pages in the index and results were very good.

-Dan

[Klarke]

Looking forward to that post, Dan.

[M_D_Golden_Peak]

Thank you both...and, we're thinking alike. I recently went through our 60+ Wordpress sites addressing the issue of non-indexed /tag/ pages and also ensuring they weren't in the sitemap via our Sitemap plugin.

For the sites that had hundreds or thousands of /tag/ pages, but very little traffic in Google Analytics (Search > Organic w/ Landing Page as "primary dimension")...I just went ahead and set them to "noindex").

For sites where the /tag/ pages were driving a fair amount of traffic (10% of site total or more), I had our editors write unique descriptions for the top 50-100 (like we do with category pages) and then we set the rest to "noindex,follow" via the meta robots tag.

For this one site...I just haven't found an easy solution that didn't leave an uneasy feeling in my stomach. It's tough to give up 25% of your traffic in hopes that Google will get it right and rank your real content higher in place of these /tag/ pages.

Uh oh...I just checked Analytics and or organic traffic started creeping down @ July 13th. When I look at just the /tag/ pages in the organic landing pages section, I see that they dropped in traffic @ 50-60%. Something bad is happening. I am setting them to "noindex" immediately.

Definitely can't wait to read your post. I'll be writing my own on www.kernmedia.com in the near future as well.

[Klarke]

Are you looking at your analytics as far back as early 2011?

I'm come across people who were hit on known Panda update day that weren't aware they were ....as strange as it may sound.

[David_ODonnell]

Additionally, make sure your posts have rel=canonical.

[M_D_Golden_Peak]

Definitely. I start with the 30 day view, then go to YTD, then push the start date back to 1/1/2011. That's my 3 step process every time I'm investigating a situation.

I've seen at least 20 of our sites decline in traffic in the past few months due to the April & June Panda updates. The dates of decline in Webmaster Tools (Traffic > Search Queries) line up perfectly with the various recent Panda updates.

Fixing /tag/ issues is one thing...but we have a monumental task of rewriting massive amounts of product descriptions next. We also have a fair amount of "no-indexing" or canonicalizing to do with our syndicated content. We'll be better for it in the end. I only wish I knew about these situations much sooner.

As I tell everyone, protect your unique content with all you've got...and keep duplicate content nowhere near your site. It's just too risky.

[M_D_Golden_Peak]

Yep, already implemented. Good point though.