We love digital

Call 0845 485 1219

We love digital - Call and say hello - Mon - Fri, 9am - 5pm

Why We Deleted 900 Blog Posts And What Happened Next

Mike Essex

by Mike Essex on 28th January 2014

Koozai 404 PageSix months ago we made one of the biggest changes we’ve ever made to Koozai.com, and also potentially the most riskiest – we removed 30% of our content.

Here’s how we did it in a safe and controlled way and why we’re recommending you do the same.

Website content has to serve so many purposes now; to convert customers, grow a brand, prove your expertise, explain your services, report on your industry and more. That’s not to mention content created to rank for specific keywords or to solve one off needs that existed at the time, as well as content that suited SEO best practice in the past but which wouldn’t hold up today.

All of this means any website that’s been live for more than a year will find itself with a lot of content that is no longer needed.

With the likes of Panda ready to penalise websites for thin-content and other algorithmic penalties for content deemed to be tricking Google, are you 100% sure everything on your domain presents you in the right light? In this post I’m hoping to convince you to make less content before you make more.

Why We Did It

Over the last seven years we’d gone from writing three short and sweet 100 word blog posts a day, to one detailed and informative guide every day that could be anywhere from 1,000 to 5,000 words long. Our old style of writing worked perfectly back in 2007– we wrote short news pieces on the latest SEO developments, However, this wasn’t so good in 2011 when we wanted to move away from a news approach and instead focus on sharing our own knowledge.

We were certainly not ashamed of our old content, but that old content didn’t really reflect what we had become and ultimately there was always the risk that Google may flick a switch and determine that the old content was too thin. Regardless of how good the content was at the time, it wasn’t good enough any more.

We also felt that content from 2007 probably wasn’t in line with the things the SEO industry would be talking about in 2014, so keeping it up-to-date was another priority. No doubt you face the same issues with your archive, and if that’s the case you should consider whether things you said a long time ago are still true today.

Research Done

The first thing we needed to do was determine which content we no longer needed. During the early years of our blog we did have some longer articles as well as pieces that were equally true today so we didn’t want to just bulk redirect everything. It was also important we tried to preserve any value those pages had.

Assuming you’re doing the same thing, here are the steps to follow:

  • Export a CSV database of your entire blog including titles, content and URLs.
  • Filter it to only show you blog posts below a certain number of characters long.
  • Import MajesticSEO link data (or cross reference) to determine which pages have a high number of unique domains. Highlight these separately.
  • Import Google Analytics data to see which pages generated good organic and referral traffic in the last 12 months. Mark these differently.
  • If you have time, manually view each page to do a spot check for any others that are no longer relevant.

You should now have three types of pages:

  1. Pages with a very low word count / that have irrelevant content
  2. Pages with a lot of links / high traffic levels
  3. Pages with no issues.

Content Analysis

Once we’d done this we ignored the pages with no issues and then drew up a plan for the rest.

Removals / Redirects

One great thing about this process was it allowed us to identify old posts that could do with a little extra TLC. We found pages that regularly generated high levels of organic traffic that we’d all but abandoned and we invested time into improving its content and bringing it up to modern standards.

In some cases pages were generating a lot of traffic, but were still out-of-date so we created new articles on the same topic and redirected the old page. This helped us keep the existing traffic, but meant we had something to show that was more accurate for the reader. We also did this for pages with lots of links if we felt they needed updating, rather than risk losing the link value.

We’d advise you do a similar thing if you go down this route and always ensure that you keep the value by improving on rewriting existing content that is already generating results rather than removing it.

For pages that were no longer generating results we took a backup of our blog and then deleted them from the live version. We wanted to keep our blog database as efficient as we could so it made no sense to keep them up.

We then created redirect rules for each of those posts. Unfortunately as the posts didn’t all sit in the same category or have the same format – and as we didn’t want to redirect other good posts in the same place – we couldn’t do a catch all rule (e.g. redirect all posts in Category X).

These redirects were either to other relevant posts on the same topic (the ideal scenario) or to the category page for that topic. We never redirected anything to the Home page as we wanted to help people find something similar to read.

Lastly we tested many of the links to check that the redirects were working and critically we tested lots of our live pages to make sure the redirects didn’t have any weird knock-on effects that may stop them working as well.

What Happened

We were pleasantly surprised by the lack of negative effects that occurred as a result of deleting the pages. Whilst we knew we’d done all we could to mitigate the effects for both readers and search engines we assumed there’d be a slight decline in search engine traffic merely due to the nature of the change.

Initially we were concerned when Google Webmaster Tools reported a rise in the number of pages not found. On closer inspection we found many of these did have redirects in place and were false negatives. One set of unexpected errors was that the category pages had 404 errors as we now had less items in them and therefore less pages. This was quickly fixed.

The false negatives also disappeared on their own and didn’t come back, as shown below:

Crawl Errors

One of the factors that we believe helped in the process was that we deliberately didn’t make this change at the same time as changing our URL to Koozai.com or relaunching the Koozai website. We didn’t want to change too much all at once and also this meant we could isolate what had caused any issues if problems did occur.

Organic search traffic actually increased by 5% in the month we made the change (July) and then declined by 6% in August, with a further 2.5% decline in September, which then rose by 10% in October. This put organic traffic higher than it had been before the change and at a level where it maintained. The below graph puts that in perspective:

Changes Made

Ultimately taken across tens of thousands of visitors the decline was marginal and in line with fluctuations we’d seen month-on-month before the change, making it possible that the change didn’t even cause a negative dip and it was merely a change we’d have seen regardless.

Likewise we haven’t experienced any drop in rankings, even now six months after the change, and some of the pages that were redirected actually improved their positions. Site speed also saw a marginal increase as a result of the change.

Finally, we didn’t receive a single message from any of our readers to say they couldn’t find a page they wanted. We checked Social Media and our contact forms and users seemed happy with the change. For most of them they’ll probably never know what we did (until today). They found what they wanted and it would have been content we were incredibly proud of. Ultimately that’s the best possible outcome.

Lessons Learnt

What really surprised us about this change was that despite taking 30% of a website and removing it overnight, there were very few negative effects. We suspect this was because of the effective redirects and keeping the good content in place; had it been our strongest 30% rather than our weakest the results would have been massively different.

Ultimately it wasn’t our intention to experiment, we simply wanted to help our strongest content rise to the top and ensure we were happy with the overall site quality. Once we saw how well everything went we wanted to share the results to encourage others to have their own spring clean.

Every website owner should be proud of their site and not afraid of waking up to a penalty or disappointed readers. If you have any suspicions that your old content could be holding you back now or in the future then we’d recommend the process above, and please leave your comments below if you’d tried this with similar or different results.

Mike Essex

Mike Essex

Mike Essex specialises in digital marketing and everything search. A recent project of Mike’s was featured on BBC News, Radio 5Live and the Times here in the UK, whilst also featuring on USA Today and ABC News in the US. He will be writing throughout the month about digital marketing and much more...

down arrow

Your Free Whitepaper

How To Perform A Technical SEO Audit (2nd Edition)

How To Perform A Technical SEO Audit (2nd Edition)

Download this whitepaper now and get a new one every month!

35 Comments

  • Todd 28th January 2014

    Nice post guys and really interesting to hear that you didn’t notice too much change after a drastic move around.

    What are your thoughts on the affect 404 errors can have have a big move? After sorting out 132 redirects at the weekend I’m hoping there’s more positives than just ensuring historical links still pass page rank.

    Reply to this comment

    • Bart 28th January 2014

      Todd it’s a good question. Very interesting case. I will try to do on my blog.

      Reply to this comment

    • Mike Essex

      Mike Essex 29th January 2014

      Hi Todd, thanks for the feedback. In the case of Google Webmaster Tools crawl errors Google seem to be of the opinion that 404 errors can’t hurt you, although they certainly don’t help if either:

      a) the content that was there was ranking
      b) users find them and have a negative experience

      So for those reasons I’d always try to clean them up.

      Reply to this comment

      • Todd 22nd March 2014

        Thanks Mike.

        That’s the main reason I did (user experience)

        I’m also looking to redirect my old site as my new site is a rebrand and new URL. So redirecting the old posts and placing them on the new site is next.

        Thoughts?

  • 3Leaps 28th January 2014

    This is the height of SEO Stupidity.. what wrong with you guys. Just because Google is making changes in its algo does not mean that you should be deleting the old posts…. this is the reason why aliens do not speak to us.

    Reply to this comment

    • Mike Essex

      Mike Essex 29th January 2014

      Kang the Emperor actually told me to do it….

      But in all seriousness we were motivated more through a desire to better guide readers to our stronger content than we were to appease Google.

      Reply to this comment

      • 3Leaps 29th January 2014

        Delete one more post and I will find you and will help you delete the entire blog :P

  • Illiya Vjestica 28th January 2014

    Interesting idea Mike. I might just try that with some of my sites.

    Can you get the Majestic SEO data from the FREE account or do you need to sign up to be able to do what you mentioned above?

    Thanks for the informative post. Great stuff from you as always.

    Reply to this comment

    • Mike Essex

      Mike Essex 29th January 2014

      Thanks Illiya. The Majestic data you need is the “Top pages/links” reports which I believe is only available with a paid account.

      Reply to this comment

  • Graham Charlton 28th January 2014

    Hi Mike,

    Interesting post, and great that you’ve shared the data.

    Why did you feel this content posed a ‘risk’ to you? Would Google not just see that this was old content and disregard it?

    We have 12,000+ posts on the Econsultancy blog, and many of the earlier ones are shorter posts. Is this something we should be concerned about?

    Reply to this comment

    • Mike Essex

      Mike Essex 29th January 2014

      Hi Graham,

      In terms of the Google risk we were thinking about sites like Squidoo where several hundred bad pages reduced the visibility of many great ones. In the early days of Panda there was a strong belief that sites could be penalised for a few bad pages, although this has proven to be less of a factor in the long run.

      It’s certainly possible Google would have just disregarded it, and the fact we didn’t receive a reduction in traffic or an uplift would seem to indicate they didn’t actually care about it.

      With Econsultancy I’d imagine the strength of your domain would negate a lot of this risk. It’s the same reason you’d be unlikely to be penalised for having many authors (which they could see as guest blogging).

      Your content is also unique and frequently referenced which is another reason you’d be safe.

      If you do have concerns I’d recommend using Search Metrics to track your search engine visibility. If it starts to decline you should investigate which pages fell and if the pattern is that they are all thin then it may be worth taking action.

      Ultimately we made the decision mostly to point people to better resources and reflect our move from news site to an advice site. The Google consideration was more of an afterthought.

      Reply to this comment

      • Graham Charlton 29th January 2014

        Cheers Mike. We’ve just switched to all .com from individual country URLs, so need to asses impact of that first. Something to check further down the line though.

  • Bob 28th January 2014

    Pity you.

    You could avoid it if you would have published only quality content.

    There wasn’t any old or new style of writing. your so called old style of writing was not even perfect back in 2007. People were still looking for quality content, but unfortunately you were fooling your readers and writing for search engines. But ultimately you fooled, youself and this is the reason that you had to remove all those thin quality posts from your site/ blog :D he he

    Anyway, all the best, but keep this secret with you only. I don’t anybody needs it, if maintaining quality since the beginning as SEW, SEJ, SEL, etc have been maintaining.

    Reply to this comment

    • Mike Essex

      Mike Essex 29th January 2014

      Hi Bob,

      Thanks for the feedback. All of the content we published back in 2007 was included in Google news so we certainly felt like we were producing quality at the time. It was only through hindsight we thought it was a bit thin.

      But I do agree not every site needs to do this if they are happy with their lifetime quality.

      Reply to this comment

      • Bob 29th January 2014

        Appearing in Google news only proves that your content was original, but it doesn’t say that your content was quality. There are certain guidelines to get your content appeared in Google News, you just need to fulfill that criteria, doesn’t matter whether your readers like that content or not. There is a huge difference between Original Content & Quality Content.

  • Adam Beaumont 28th January 2014

    Risky strategy considering we really have no idea what Google is thinking but glad it paid off for you.

    Reply to this comment

  • Pingback: SearchCap: The Day In Search, January 28, 2014

  • Benjamin Beck 29th January 2014

    Mike,

    This is Brave! I applaud you guys for taking initiative in testing this out BEFORE anything potentially bad could happen.

    You aren’t seeing a big lift now, but with your Google Crawler bandwidth being more concentrated now I’m sure it will pay off in the future!

    Thanks again for the write up.

    Reply to this comment

  • Mike Volpe 29th January 2014

    I agree your data says there was no negative effect. But was it worth doing all this work? The stats seem to also say there was no positive effect… so why bother with all this work? You could have been writing more articles! Just curious.

    Reply to this comment

    • Mike Essex

      Mike Essex 29th January 2014

      Thanks Mike.

      I agree if our primary goal was to get better traffic / rankings then we failed but our goal was to better help people find our stronger content which converts better and will help us keep our readers for longer. We also wanted to remove any risk of future issues.

      In that respect there was a positive outcome, just not one you’d measure with raw metrics.

      Reply to this comment

  • David Black 29th January 2014

    I dream of getting to 900 posts never mind deleting them – I’ll give your strategy a try in around the year 2024 when I’ve got enough content for a clear out. I’ll let you know my results if you’re still around.
    In the mean time, keep up the good work.

    Reply to this comment

  • Doug Roberts 29th January 2014

    It’s not _just_ about traffic though is it?

    If your site’s visitors are now more likely to hit a substantive, relevant, up to date article (rather than the thin, old content), has that improved your engagement metrics such as bounce rate, shares etc?

    Things change. Reviewing your content to make sure it still supports the goals of your site/business just seems like common sense.

    Reply to this comment

    • Mike Essex

      Mike Essex 29th January 2014

      Exactly right Doug. The real positive of the experiment was just that, we could push people towards more recent and better researched content.

      We’ve seen a rise in whitepaper downloads and sales from the site in the time that followed. Sadly it’s hard to isolate that just to the page change which is why I didn’t include it as a result, when the likelihood was the two were connected.

      Reply to this comment

  • Anoop Srivastava 29th January 2014

    Mike, Thanks for sharing it. It encourage others those have low quality content and they are not deleting it because that they will loss the traffic and index pages.

    Reply to this comment

  • Butler 30th January 2014

    I think the problem here isn’t the fact that you’ve done what you’ve done, but that you’ve framed it in the context of SEO, Panda, and potential Google updates.

    This should be seen as quite the opposite: a long term play to increase the ratio of good to bad content as an investment n your brand and ongoing content strategy.

    Reply to this comment

  • Maria 30th January 2014

    Completely agree with some of the commenters above. Deleting old posts to please Google is an odd approach, to say the least. If you started off writing thin content simply to please Google News, surely that’s an indicator that you (we, everyone) is better of producing quality than working to Google’s latest whim. Surely it would have been better to pay someone to review and rewrite what you already had to bring your entire blog up to a better standard.

    Reply to this comment

  • Pingback: E2M Eagle’s Eye - 31 Jan 2014 | E2M Blog

  • Tim Soulo 2nd February 2014

    Oh you guys… I can totally understand you!!!

    It’s not about SEO and Google.. it’s about people coming to your site and reading stuff that no longer represents who you are.

    I don’t have too many articles on my own blog, but just recently I’ve decided to go through the early ones.. delete some.. edit others.. I just can’t live with the fact that there’s some weak content on my blog and someone might see it.

    And as for Google.. I think they actually should think about endorsing this kind of practice.. Matt? Do you hear me? :)

    Reply to this comment

  • Mike 10th February 2014

    I think we’ve all been here with one site or another, at least you got round to it :p

    The thing with this sort of thing is that it does just need doing once in a while. Like cleaning the back of your cupboard.

    Reply to this comment

  • Liz 28th February 2014

    Very glad to have found this since I’m planning on a similar weeding out of my 800+ posts. The judgmental comments boggle my mind. There are all sorts of legitimate reasons for doing this – including reasons related to SEO, etc.

    Reply to this comment

  • Ruud Hein 3rd April 2014

    Love the idea. I wonder how you went about matching up your WP data with GA though. Pulling in both goes fine but lining them up…. Whether I work with page title (reasonablish) or URL (horrible; to much crap) it seems like a lot of manual work must have gone in?
    There’s no way to upload a CSV with URL’s and get stats back from GA, I guess…

    Reply to this comment

    • Mike Essex

      Mike Essex 3rd April 2014

      Hi Ruud, thanks for the comment. In the end I believe we rewrote the URL’s slightly after they were exported so they matched up and then we used a pivot table to combine the data. I hope that helps.

      Reply to this comment

  • Abhi 26th May 2014

    i have deleted some of my posts and changed category slug… and lost 90% of my traffic since feb-2014… do you think it will take some time to recover… 301 redirects are in plance…

    Reply to this comment

    • James Perrin

      James Perrin 28th May 2014

      Thank you for your comment.

      Deleting your posts and changing the category slug creates a new URL so it’s important the correct 301 redirects are in place.

      If the correct 301 redirects are in place you should see it recover but it will take time.

      Reply to this comment

Subscribe To The Koozai Blog