Webtrends Optimize recently co-hosted a workshop at the request of our valued client The Goodwood Estate, and together with our partners at Fresh Egg and Fullstory. The purpose was to analyse the process of website migration, a process which the Goodwood Estate had recently been through, and to highlight how you can use behaviour analytics and testing/ experimentation to de-risk and enhance this process.
Goodwood have a very strong brand with significant peaks around key events, and modernising their website to provide a better customer experience for its visitors was a key objective. They then went through a full migration process onto a brand-new website, something which is not easy with a multifaceted offering like Goodwood. We heard on the day that only 20% of migration projects are deemed to be successful when measured against their KPI’s and Goodwood are in the minority of businesses who have actively sought to de-risk this process through experimentation.
Goodwood is England’s greatest sporting estate and has been for more than 300 years. Whilst typically, Goodwood is known for its motor racing history, amongst other things, it is also home to one of the largest lowland organic farms in England, a famous Battle of Britain airfield, a celebrated racecourse, a historic motor circuit, two golf courses, a hotel, multiple restaurants, a health club and one of the oldest cricket grounds in the country, all centred around Goodwood House. They have a lot of visitors who use their website for tickets and information for a multitude of events, and balancing these many often-conflicting needs, whilst delivering the right information to the right audience, in the right way, at the right time was key.
The day was broken into three main sections:
- Things to consider when sunsetting a website
- The migration itself
- What to prioritise when the new site is live
Which gave a full overview of the website migration journey and what to consider at each milestone stage. We had an expert panel discussing each stage of the migration process, made up of James Harber, Head of Customer Success at Webtrends Optimize, Tashan Dunbar, Commercial Account Executive at FullStory, Jake Lambert, Head of Conversion Services at Fresh Egg and Simon Gardiner, Head of Digital Product at The Goodwood Group.
Fullstory Insights – What’s the data shown us?
One of their findings was that Horse Racing is the lowest converting product. Within their analysis, Fullstory identified 24 signals as contributing to abandonment. Signals represent a mixture of technical and performance issues. These issues represent an estimated 6000 lost conversions annually.
Moreover, they also noted that using the navigation bar vs the sub menu increased conversion by 1.5%, and visitors convert quicker without the sub menu. Sections of the website that have carousels on them also reduced conversion by almost 2%, and they drastically increase the time to convert.
When reflecting on the old site and how well it converted, compared to the new one, conversions were up by almost 1% (at the time of the event in June 2024). Conversion time was also down by half a minute indicating improved performance/usability. However, during their analysis, Fullstory noted that slow loading mean people are more likely to drop out the funnel which was something for Goodwood to address as a priority, alongside investigating why AOV had marginally declined.
The new site had a simplified navigation, specifically changing to a Navigation Burger menu, that drastically reduced the time to convert. Using different filters meant that conversion had increased by 1.3%. On the previous version of the site, filters were used 3 times more when accessing the ticketing section of the website compared to now.
Another change was the Book Now feature. Previously, on the old site, the Book Now button was dotted around everywhere. But now, it appears more frequently on event pages, which has resulted in increased conversions.
The overall improvements meant technical frustrations that were affecting conversion rates were eradicated, behavioural metrics were largely trending in a positive direction, and as expected key areas were identified to focus their experimentation efforts on.
Sunsetting your existing website
The first panel discussion was focused on pre-migration to highlight guidance about what you should be doing before migrating your website, or at least be thinking about. With up to 80% of migrations facing significant challenges, it is important to bear in mind there will be hurdles as you transition to a new site, and that these risks need to be identified and managed.
Whilst there are lot’s of things to consider at this stage, a good starting point is to determine what you are wanting to get out of the migration and why you are doing it. This will help to closely manage the expectations of stakeholders and communicate the upcoming change to customers, ensuring collective buy in to the project and aligning everyone to the same goals/objectives.
Aside from the above, panellists agreed that the first thing you should be doing is testing. This will help to validate your assumptions and plan future testing for when your site is live. It is important to remember that just because you are migrating your website, business doesn’t stop. So, whilst your strategy has changed, you should still continue tactical testing with the goal of optimising and hitting targets.
These might be small campaign-based tests to maintain buy in for the project with internal stakeholders or testing elements of the new site within the existing structure to de-risk the migration process. This was discussed to be two very important tactics in improving chances of success in a migration project.
Prior to your migration, it is recommended that you set benchmarks of what you are going to measure and why. However, it is important not to get caught measuring websites on a like for like basis. Your site is going to be different, and you are changing it for a reason, so whilst you want to make sure you are improving on the pre-agreed metrics whether it be Performance, User Experience, etc. You should not be comparing on a like for like basis. Turbulence during migration is to be expected, and sites need 1-2 months to bed in before the top-level metrics can be confidently reported to be trending up or down.
Throughout this stage, teamwork and unity is arguably more important than all the technical factors and decisions. Simon Gardiner (Head of Digital Product at Goodwood) highlighted that having trust within your colleagues to help you down the line is a really important element of success. Some days will look different to others, but having the same shared vision and building the right team around you is going to remove barriers and inefficiencies in the long run.
The Migration
Depending on who you ask and what their experience is, there are different approaches to a website migration. Though the big decision boils down to whether you’ll phase the migration or do a hard swap between sites.
A phased migration would mean that you changed elements of a website slowly rather than doing everything all at once. A hard swap is where you change everything from the old website to the new website instantly. Whilst many people prefer a hard swap, if anything was to go wrong, then it could result in damaging the brand’s reputation. A phased migration can be more beneficial for SEO purposes as you aren’t changing all the links at the same time, but without careful planning can lead to confusing Google which should be avoided!
A good place to start is to have a clear set of metrics you want to achieve and assessing each decision on these bases. Have you actually fixed things that weren’t working, or have you just moved where a problem is occurring somewhere else in the funnel? Is this opinion or an actual problem? Which links back to the importance of the insights generated by Fullstory revealing which aspects Goodwood should be focusing on to most improve the visitor experience and reduce friction in key user journeys.
There is no right answer on whether to do a hard swap or phased migration as there are risks with both approaches. The advice from the panel was to break things down and look for opportunities to de-risk where you can. Phase the aspects which can be phased, and test items/concepts that can be tested in the existing site where you can, which will minimise the impact of a hard swap. Though the main take away from the day was the importance of continuous communication throughout, to both customers and the internal team, is key to ensure this process is as a smooth as possible.
Research to help inform decision making
Fresh Egg performed extensive research to identify to Goodwood which areas they should be focusing on and improving as part of the migration.
Some of things they identified that could be improved upon were:
- Improve the conversion rates from articles rather than just writing them for SEO purposes (these are some of their highest entry pages to the site)
- Utilise product recommendations in different areas of the site; this again could be through articles with more of a story attached to recommend products
- Track preferences and engagement metrics to deliver a more personalised experience
- Increase the use of filters so visitors can tailor their experience based on their preferences
This again highlighted what to take into the new site, which areas to prioritise in the experimentation strategy, and what data they should be looking at.
Post migration
When your new site is live, there can be a lot of expectation for things to be ‘perfect’ straight away. However, it is important to set new benchmarks for what you want to achieve, and if things aren’t working instantly, you need to allow sufficient time for the site to bed in before making large scale changes and if there are any red flag metrics that if they fall above/below a certain threshold you intervene. Having flexibility in your post-launch roadmap will mean you can adapt and make tactical changes as you learn the behaviour of your visitors and gather data.
And remember to keep testing! You should keep optimising as you normally would and continue to test at all points of your journey. As pointed out by Fresh Egg, only 33% of your test ideas actually win. So, if you don’t test, you don’t learn, grow or change. Crucially, A/B testing can be used during migrations to protect against losses or decisions/changes that can lead to increased negative behaviours.
What can we take away from this experience?
The workshop ended with an audience discussion about the top takeaways from the day, an interesting exercise when there are so many different approaches to website migration. Using Goodwood as a live case study allowed us to explore how you would tackle a multi-faceted website with a range of different offerings.
The five main takeaways were:
- Involve stakeholders from the start and all the way through
- Don’t stop testing
- Identify your success metrics and stick to them
- Have a single source of truth for your customers (and colleagues)
- Data always trumps opinions
But mostly importantly, there is no ‘right’ way to do it. We hope this helps as you consider/ begin your website migration. To learn more, feel free to get in touch and we will be happy to elaborate on any of the advice above.
