In April 2023, BrightonSEO – the ‘world’s largest search marketing conference’ – opened its doors to around 6,000 digital marketing specialists from across the world who descend on the UK’s south coast for three days of talks, networking and taking photos of the (surprisingly sunny) beachfront through the windows of the Brighton Centre!
Whilst the main conference took place from the 20th to 21st, the day before included various fringe events which focused on the more niche areas of digital marketing, including web analytics and conversion rate optimisation, paid social, paid search, digital PR and affiliate marketing. Our Insights team at Gravity Global Performance Marketing attended ‘MeasureFest’, the data-centric fringe event for web analysts and CRO experts.
After watching various talks on how to record, use and get the most out of data, the team has compiled the key insights from the event below:
GDPR and ePrivacy considerations for A/B testing
What better way to start a day than with cookies! Unfortunately, there was not a chocolate chip in sight, only a discussion on the “dos and don’ts” of collecting user data for the purposes of A/B testing. Rowenna Fielding did a very good job of turning data protection and privacy regulations into an interesting topic, which is no mean feat.
The talk began with the difference between GDPR (General Data Protection Regulations) and ePrivacy, that being GDPR governs how data representing individuals should be used and ePrivacy governs how devices connected over a public communications network (the internet) are allowed to communicate with each other. A common misconception is that cookie consent banners are a result of GDPR being published in 2016. But this actually began with the 2002 ePrivacy Directive (2003 PECR in the UK), which states that if it is not required for the communication to function and if the user hasn’t specifically requested it, then consent must be provided before data can be read off or written onto a device. GDPR only updated laws regarding cookie banners and the definition of ‘consent’.
Rowena then went on to discuss how the laws set out in the GDPR and the ePrivacy Directive can impact A/B testing. For instance, the basis of an A/B test is to find the most engaging format of a website/webpage/website element, which when put to its extreme could include what can be classified as ‘deceptive design’. ‘Deceptive design’ refers to website or app designs that intend to manipulate an end user into acting against their own wishes, such as hiding a ‘Reject all’ cookie button by using grey text on a grey background. While this type of ‘deceptive design’ is technically legal under the ePrivacy Directive because the option to reject all cookies is available, it is neither fair nor transparent and therefore could be considered to be a possible breach of GDPR.
Some of the possible solutions to this issue include:
No longer performing A/B tests (can you imagine how that went down in a room full of CROs?).
Get consent first (though asking for a user’s permission to be tested on is likely to result in a ‘no’).
Bury your head in the sand (which has been the industry’s preferred option since 2003 – and whilst no one has yet been litigated for this, it can cause you to fail a compliance audit).
The fallacies of A/B testing
Sticking with CRO, Ellie Hughes took an informative and in-depth look into the many common misconceptions relating to A/B testing. Myths relating to both the planning and statistical analysis sides of testing were explored, with clear and insightful explanations of where the issues are and how best to avoid them within conversion rate optimisation projects.
With planning, for instance, you may think that an A/B testing programme is better with a long backlog, and hence the best approach is to create a list of all the possible experiments you could conduct before prioritising them and then executing. However, with such a long list, it’s never going to be possible to deliver all of them – as you learn from the results of the tests that you do conduct, priorities will change, and the list will rapidly become out-of-date. It’s better to continuously look at the data you collect to look for support for new ideas, while disregarding those which have no data currently available to support them.
On the statistical side, some believe it’s possible to interpret experiment results and the associated statistical confidence repeatedly throughout the duration of an experiment. However, the results can massively vary throughout, particularly in the early stages when unique visitor differences between the variants are more influential. The truth is that ideally, a sample size should be determined in advance, and be stuck to throughout – and that all results should only be analysed with the assistance of a specialist data scientist or analyst, to avoid incorrect conclusions.
10 must-have GA4 reports for all SEOs
With the UA deadline looming, it was very unlikely that an event dedicated to website analytics would not mention Google Analytics 4 at some point. Farhad Divecha ran through a quick-fire session of the top 10 GA4 reports for SEO, how to create them and how to gain insights from them. Those reports are:
- Google Search Console (via connecting the two platforms)
- Conversion Paths (via path exploration)
- Segmentation of Organic Traffic
- Site Speed
- Referral Traffic (via free form exploration)
- SEO Landing Pages
- SEO Paths (via path exploration)
- Data Stream Breakdown
- Content Stickiness
- Scroll Tracking (using additional scroll events in GTM)
One significant difference between Universal Analytics and GA4 is the custom reports functionality. Whilst the ability to create custom reports was available in UA, the number of varying reports the platform provided as standard meant that there was often little need for custom reporting. By contrast, GA4 comes with fewer standard reports and puts more emphasis on custom options, giving the end user an opportunity to make their data their own.
The tools for an inspiring GA4 onboarding experience
In addition to Farhad’s talk on the best in-platform reports, Agata Adamiak discussed additional tools outside of the GA4 platform in their talk “The tools for an inspiring GA4 onboarding experience”. The talk focused mainly on the use of dashboards and slide decks – two data visualisation formats used every day in the world of digital marketing. Agata emphasised the importance of knowing your audience and including the right people for the build.
By knowing the audience that will see and use the data visualisation, you can determine what format, KPIs (key performance indicators) and level of detail is appropriate. A dashboard showing high-level metrics of a paid search campaign would be ideal for daily or weekly monitoring of performance by the campaign manager. A slide deck that follows a user journey, including works completed and their results at each stage would be more suited to an annual review meeting with a client or board of directors.
As well as knowing your audience, Agata discussed the importance of including the right people within the process of building dashboards or slide decks. This includes the use of data engineers or scientists to help automate the pulling, transforming and loading of the required data, data analysts to select the appropriate charts and visualisations and pull key insights and designers to bring the data to life in an eye-catching way. This should help to avoid too much manual data entry, the inclusion of unhelpful visualisations or generic templates. This is certainly an ideal scenario, as having experts in each of those areas is a luxury not many companies or agencies possess.
Overall MeasureFest April 2023 was incredibly insightful for those that work in CRO and with A/B tests or analysts that wish to get more from their data and those going through the transition from Universal Analytics to GA4 – which is pretty much all of us! Our Insights team left with plenty of new ideas to put into practice, and the reassurance that our current approach as an agency to GA4 transition and reporting is BrightonSEO-recommended.
To find out more about Insights and how to get the most out of your data, get in touch.