Freelance SEO Consultant and Content Marketing Specialist

6 mistakes I’ve seen cause a website migration to go wrong

Tom Crewe imageTom Crewe
May 6, 2022

Before I start this blog, I feel like it’s my responsibility to say this loud and clear, if you are thinking of migrating your website, get an SEO involved AS SOON AS POSSIBLE! Ideally, you should involve an SEO before decisions about the sitemap, information architecture or website design have been made. Of course, I work in SEO so I would say that, right? But I have seen so many websites lose visibility, traffic and of course conversions/revenue because of mistakes that were made during the website migration, and it’s pretty difficult to reverse these damaging effects once the migration is complete. Prevention is better than cure with website migrations, so please do take this advice seriously!

I’ve worked on countless website migrations to help clients maintain and even improve their website performance when migrating, whether that be to a new domain, design, platform, content or even migrating multiple merging websites into one new domain (read my case study on this here). I’ve also performed analysis for many website owners that have recently undergone a failed website migration in which they instantly lost traffic and rankings, so I’ve seen a lot of mistakes that have contributed towards that loss.

While there are some situations where rankings and traffic loss might be unavoidable, for example when moving a large website from a well established domain to a brand new domain, there are still certain strategies that must be followed to try and minimise this loss. Below are six mistakes I’ve personally seen cause a failed migration over the years, so avoid these wherever possible!

Unnecessary URL changes

If your existing website is already indexed and ranking well, that means that Google already knows the location of those pages within the sitemap and trusts the content on the ranking URLs. When a URL is changed, Google then has to reprocess and rediscover those URLs, essentially reviewing the new URL as a brand new page. While a proper redirect setup can help with this process, URL changes still carry a lot of risk, and when URLs are changed en masse it can quite often cause a significant and noticeable drop in organic performance.

One of the worst cases I have seen of unnecessary URL changes was a client that not only added two new completely unnecessary folders to their high ranking content (i.e. the URLs went from /blog/blog-title/ to /blog/news/articles/blog-title/), they also changed the optimised URLs to a string of random letters and numbers (e.g. from /blog/blog-title/ to /blog/news/articles/fnieqwDEdna8jddewD89/). Needless to say, traffic tanked!

Advice: Maintain your URL structure where possible. This won’t always be possible of course, so at least try to maintain and replicate the URL structure of your highest performing pages, and don’t unnecessarily add new levels to your URL folder structure which may increase the crawl depth of each page or remove any clean, optimised URL structure.

Unnecessary canonical changes

Your website should specify a canonical version of each URL within its HTML. Essentially, this is a fairly simple piece of HTML which specifies to Google which version of the URL should be indexed. This is necessary when your site has more than one version of a URL, for example if you have products with multiple colours or sizes (as long as users don’t search by colour or size for your products!) Even pages that aren’t a duplicate should have a self-referencing canonical, just to make Google’s job easier when it comes to crawling and indexing.

Now, one of the issues I have seen with canonical setup is when website developers have not properly considered which canonical version of the URL was ranking before the migration. For Google, a version of a URL with a trailing slash is different to a version of a URL without a trailing slash, for example:

Is not the same page (in Google’s eyes) as

This means that if you’re not telling Google which version is the canonical, Google either has to decide for itself or will just get confused and rank both intermittently, which will likely impact the ranking position.

So, it’s important to include a canonical URL, but the main mistake I have seen here is a lack of consideration for the canonical version of the URL that ranks pre-migration. I have seen a website migration where the pre-migration ranking URLs all had a trailing slash, and the canonicals all had a trailing slash, but on the new version of the website the developers had canonicaised all URLs to the non-trailing slash version of the URL. This meant that Google had to reprocess and rediscover URLs across the whole site, and inevitably resulted in a huge loss in visibility.

Advice: If you aren’t making changes to your URLs, then make sure you crawl your existing site to discover which version of the URL is currently canonical, then replicate this across your new site. If you are changing URLs, ensure the canonical element contains the newer version of the URL and not the old version!

Site wide redirect chains

If URLs are being changed, then redirects are entirely necessary and are the right way to try and retain organic performance. However, 301 (permanent) redirects do result in a small loss of authority when moving one URL to another, so the last thing you want to do is accidentally create a redirect chain in which one URL redirects to another URL which then redirects to another URL and so on.

The mistake I have seen here is when website developers redirect to the wrong version of a URL, which then has a ‘catch-all’ redirect set up to redirect to the right version of the URL. As previously mentioned, Google sees trailing slash and non-trailing slash URLs as two separate entities, but this is true in other scenarios too. In this instance, I have seen developers redirect each page to the HTTP version of the URL, which then redirected to the HTTPS version of the URL, causing a redirect chain across every single page on the website. This meant more authority was lost for each URL which contributed to rankings and traffic loss.

Advice: Ensure that your redirects are set up correctly, so each URL you are redirecting points straight to the new version of the URL, with no steps in between. Redirects should be tested on the staging site before going live, ensuring that each 301 goes straight to the 200 (OK) version of the page.

Redirects not set up at all

This one should be obvious, but if you have a high performing page on your existing website and you change its URL on the new website but don’t redirect old to new, then you’re going to have a bad time. Google needs clear instructions, so you shouldn’t rely on the fact that the content on the new page is the same as the old content if the URL has changed. You need to specify exactly where each old URL is moving to on the new site to try and retain any value that page had.

In this instance, I haven’t actually ever come across a site that has not implemented redirects at all, but I have seen high performing sections of a website not redirected. While most of the main content was redirected appropriately, some of the highest performing blog pages, which were driving both traffic and conversions, were not redirected. These pages accounted for around 30% of the traffic to the site, which of course, was instantly lost without the redirects.

Advice: Ensure 301 redirects are set up across all pages. Crawl the old site so that you know you have a full list of URLs that need redirecting, then with your exported crawl data in Excel, switch out the live domain for the staging site domain and recrawl that list once you have set up the redirects in the staging site. This should then show you all of the pages that result in a 301, 200, 404 etc. so you can discover any areas that have not been redirected appropriately.

Removal of H1 tag

Changes to the design of a page, even without changing content, can still have a negative impact on organic performance (Google said it themselves!). There are certain elements on the page, such as header tags, internal links, map embeds and content hierarchy on the page that can impact performance.

Header tags, such as the H1 on the page, are of particular importance. This is one of the first places Google will check to get a better understanding of what the page is about. For one failed migration I analysed, the design of the page had changed, but not the content, however they had not used a H1 as the main header and just used large, bold text instead. This caused rankings to drop for a lot of their products and was such a simple error that could have been easily avoided!

Advice: Analyse the current web design for anything that might be contributing to rankings and ensure it is carried across like-for-like. Even if the current site does not have header tags, using them on the new site will only be a good thing! The design of the new site should be carefully considered for both SEO and UX in order to improve performance.

No consideration for what was working on the old website

Finally, the most common reason I have seen for a failed website migration is a lack of detailed analysis and consideration of how the existing website performs and how that performance might be retained. By blindly making decisions about a new website, whether that be the page layout, content, site architecture or anything else, a lot of value can be lost during the migration process. Each page needs to be analysed to understand its value when building the sitemap, which is something I recently wrote about in my Search Engine Land article, ‘How to maintain organic performance when merging multiple websites’.

This mistake is a bit of a generic one, but the reality is if you haven’t implemented a proper migration strategy for SEO, then there are plenty of mistakes that can be made along the way. If the goal is the retain and improve upon the value of the existing website, then you have to consider what is already working and set up the new website with this in mind.

Advice: Analyse every single page on the existing website, looking at which pages drive traffic and conversions, have keyword rankings or backlinks or potentially support the rankings of your priority pages. If your website is a major source of income for you, then definitely consult an SEO specialist as early as possible in the migration process, before key decisions have been made.

This last piece of advice isn’t meant to just be a sales message to try and get you to use my services for successfully migrating your website (although that would be a good idea!) – it is genuine advice that could prevent a negative impact on your business come migration day. So, whether you use me or another SEO, just make sure that you don’t shoot yourself in the foot when planning a website migration by not taking the SEO setup seriously. If you do want to talk to me about your website migration, contact me today!

Share
Tom Crewe image

This article was written by

Tom Crewe

Tom Crewe is a Freelance SEO Consultant with over 8 years’ worth of experience in Search Engine Optimisation and Content Marketing. His genuine passion for SEO is clearly demonstrated in his client work, the articles he writes for industry leading publications such as Search Engine Land, the talks he delivers at events such as Brighton SEO and his eagerness to learn more each and every day.