There are a number of different types of site migrations and reasons for needing to migrate a site:
Hosting: Whether you’re switching site hosts to increase data storage or improve browser security, site migrations can give you a great opportunity to boost performance. You’ll need to come up with a cohesive redirect strategy as you’ll be changing your Domain Name Server (DNS) to point to a new IP address – even if the domain name itself stays the same.
Architecture: You may be looking to improve UX and the paths to conversion through a website rebuild and site migration.
Platform: If you need to fix tech issues and bugs, enhance functionality or improve security, then you may be in the market for a new CMS, which often requires a site migration.
Domain to domain: We’re talking a name change, brand refresh, or shifting focus on to a sister site. These are all valid reasons for conducting a site migration.
Subdomain to subfolder: Subdomains don’t tend to benefit from backlinks to the root domain as much as subfolders do. They also tend to be harder to manage and are often separated across different CMS’s, so are more likely to suffer from subdomain conflict. As a result, some companies decide to re-organise their content into a subfolder formation.
Nonsecure to secure: (Otherwise known as HTTP to HTTPS) It’s been reported that running over HTTPS is better for performance, SEO rankings (HTTPS is defined as a lightweight ranking signal), referral data and so on – so it’s no surprise that more and more companies are looking to migrate from HTTP to HTTPS. A common error dev teams tend to make in this context is to forget about updating internal links.
Despite popular belief, site migrations can, in fact, present you with greater search presence and ROI opportunity.
So why is it that we’re often quite happy to undertake this activity without that TRAINING, ADVICE or PREPARATION?
Why do we leave it to the IT team to sort it out?
And why do we scrimp on the migration budget?
Fair enough, SEO is not exactly the sexiest part of the whole migration process, but pumping time, effort and capital into design alone is a sure-fire way to end up buried at the bottom of the SERPs.
Jon began by showing us an example of a multi-billion pound high street bank who failed to implement the appropriate SEO prior to a TLD migration, and lost significant visibility for a range of competitive terms as a result:
(NB: Examples have been anonymised. No naming and shaming here! Although, if you are desperate to know who we’re referring to in the following examples, do drop us a comment below and we’ll get back to you)
Jon pointed out that this type of activity can be expected for a TLD migration in the initial stages, so we took a look at the aftermath expecting an uplift in new TLD performance – and how wrong we were!
Even outside of the “migratory window” or period of adjustment (14 days), during which positions should have returned to equilibrium, this site incurred a 28% loss in visibility, and after 75 days, a 35% loss – as well as a huge dip in revenue, no doubt.
This was all down to the company in question retrospectively implementing SEO post-migration. A big no-no!
Pi Datametrics Visibility Index further demonstrates this; showing overall performance across a range of key search terms, and a visibility score (Visibility Score: Position 1, Page 1 = 100).
Before the migration, this high street bank had an average visibility score of about 60, which showed strong performance (To give you some context, that would mean they returned at around position 10 – on average – across a whole host of key search terms).
After the migration, however, we can see this site never recovered from its loss across these competitive financial terms, and still hasn’t. In fact, they have since invested a small fortune on paid search to increase their visibility.
Cases like this have spawned an “Old migration myth”, which is now thought to be common knowledge: I.e. that you should expect to lose at least 20% of your traffic following a migration – a statement which we know to be simply untrue.
Even Google seems to agree with the “Old migration myth”.
We typed in “How much traffic could we lose when we migrate our website?”, and Google retorted “Did you mean: how much traffic would we lose when we migrate your website?”
On top of being overtly sassy, Google served up a bunch of results warning us to expect as much as 70% loss in traffic following our migration. No pressure then!
At Pi, we see migrations as a great opportunity to evaluate architecture, improve performance and overtake competitors.
There are a thousand and one things that you need to get right with a migration – but both Jon and the Pi team believe these to be key:
SEO should be the cornerstone of any site migration strategy. Without it, you may as well be throwing your hard earned positions and revenue down the drain.
Not crawling your site efficiently can result in a rise in soft 404s, as we can see below.
See below for more evidence…
This could have been avoided if this company had only audited their site to determine appropriate redirects. Jon pointed out that regaining authority following these 404s will take a long time.
According to Jon, your data holds everything together and acts as damage limitation during your site migration. You should, therefore, use your data to:
The migration window refers to the seven to 14 days after a migration has been executed. During this time, you may see a variety of different outcomes in the SERPs:
1. Changeover takes time: The official changeover may take some time to kick-in
2. Arbitrary drops: Random negative performance can occur
3. Brief coexistence: There may be a brief coexistence between the two domains for a short time
4. Straightforward transition: Positions may remain the same both before and after the migration
5. Positive uplift: Performance may improve post-migration (i.e. if internal conflict is resolved through an appropriate landing page/canonical linking strategy)
Jon particularly stresses this point. Without a clear line of authority to reaffirm the importance of the other 4 steps, a migration will go off-track, resulting in that dreaded drop.
As Jon says, your stage site is an exact duplicate of your live site, post-migration, and is therefore very likely to conflict with it if let loose in the search engines, so it’s vital that the pre-live site is protected.
Even if a stage URL is shared privately in Gmail – Google may well still index the site via this medium. This is based on what we’ve seen at Pi previously with our own clients.
Making a business case for SEO, and proving it’s worth to those who don’t quite speak SEO is always going to be tough. If you’re ever in need of more Pi charts to justify the role and importance of SEO in site migrations, don’t hesitate to get in touch. Just drop us a comment below, give us a call, or book a demo.
Equally, if you’ve had visibility issues following a migration, or are about to undertake one and want to find out more best practice site migration tips, feel free to book a demo to see how Pi Datametrics can help you.