BrazilChallenges in web analytics and tips for implementation
Amid the constant noise of the online revolution, using web analytics to garner insights and drive better web performance is an essential way of enhancing a company’s online footprint. Consultancy.eu sat down with Anton Bies, a data analytics consultant at Digital Power, to understand more about the challenges in the space, and how companies can best tackle them.
Tracking online performance is important if any company is to stand out among the millions of other websites around. There are a number of tools designed to help with this, like Google Analytics or Adobe Analytics. However, despite this, consultancy Digital Power finds that in practice the web analytics statistics of businesses are often inaccurate.
Digital Power consists of a team of experts who are passionate about data, helping to enable both small organisations and multinationals to work effectively with data. The firm specialises in data analytics, customer experience, data science and data engineering, and is the strategic partner of Dutch companies in the field of data driven digital transformation.
Since joining the firm in 2017, Technical Web Analyst Anton Bies has seen first-hand the problems companies face when it comes to analysing their data. According to Bies, there are often serious issues with the implementation of tools clients use, leading to important actions on the website not being tracked, or not tracked correctly. They often therefore do not know how their customers use their website, sometimes even when they think they know.
“In our experience,” Bies explained, “it is often not the type of site that determines the quality of their web analytics, but the level of digital maturity of the company. Companies that are more mature are more often focused on analysing important behaviour on their website with the goal to improve the customers’ experience... With that comes more knowledge of digital analytics and therefore the realisation that a useful web analytics implementation requires serious attention. That is why a higher level of digital maturity leads to more accurate analytics statistics.”
The simplicity of web analytics may strangely be behind the failure of many firms to adequately use them. According to Bies, because it seems so easy to just add a “single line of code” to a website and have a web analytics tool on a website, many online teams never learn how to get real value out of the tool,” and fail to invest enough specialised knowledge as a result. Just because Google Analytics has a lot of standard information it collects, which gives a company many data points without any customisation, it does not mean everything automatically works the way it is needed to.
One example Bies pointed to was how a very popular set of reports in Google Analytics relates to traffic sources of visitors. He stated that to ensure sessions end up in the right traffic source bucket, firms needed a decent approach to campaign tagging, but this was “rarely done well, even though it has a huge impact on the data.”
Balancing act
Just how should companies go about balancing accuracy versus usability and actual usage, then? Bies suggested that, first and foremost, they should bear in mind that they will “never work with 100% accurate data,” and build systems to compensate for this. Technology limitations, for example, or cross device visitors that do not log in, or even the use of ad blockers can prevent web analytics from being completely accurate.
“Companies should, first of all, only use web statistics as a way to gain insight in their way of working, not to provide accurate and specific numbers,” Bies warned. “And then a company should control what they can control as well as possible. And accept what they cannot control as unavoidable error rate… Then, firms must address the need to be successful in translating analytical data into real action. Many website owners only use a fraction of the possibilities they are offered by web analytics data, which is a missed opportunity.”
When asked what companies can do to more effectively gain strategic insights through web analytics, Bies commented that they should begin by defining a core mission, vision or strategic goal. From that basis, define website Key Performance Indicators (KPIs) and ensure the most important visitor actions and KPIs are tracked first.
He expanded; “With that fundament available, you can get a first view on how the site is performing and start looking for things that have high improvement potential. When you start thinking about how that could be achieved, more questions will come up that will require additional analysis and therefore more advanced analytics tracking.”
Implementing web analytics
As with all strategies, however, it is ultimately worthless if it cannot be implemented properly. It is standard procedure for consultants to recommend starting a change project with quick wins – they bring benefits and can be reaped relatively easy – and the implementation of a new web analytics regime is no different.
One example Bies pointed to is the potential to perform a quick review lasting a maximum of two days, which can uncover many issues that can directly be fixed, such as broken tracking, or missing tagging on things that are clearly important to measure. In the end, the biggest wins will come from setting up a clear structure. In the context of web analytics, this could mean documenting an analytics framework which is used as a foundation for the implementation.
“It is often not the type of site that determines the quality of their web analytics, but the level of digital maturity of the company.”
– Anton Bies, Digital Power
“Many companies want to do ‘really cool data science things’ while the basics are not in place,” Bies cautioned. “In a situation like that, you might end up spending a lot of time and money on a product recommendation engine, while the checkout flow is broken on some browsers. Actually, many data science applications use web analytics data as input, which means data quality issues strongly influence what can be achieved with data science. Garbage in is always garbage out.”
On top of this, he stated it would be wise to implement web analytics step by step. This would mean, when implementing ecommerce tracking, for example, firms should start with the most important part of the implementation – in that case transactions – and work back from there. To help clients put all this into practice Digital Power has developed its Building Blocks Model.
“To coach clients through the steps required to go from business goals to actionable data, while we continuously spend time on accuracy," Bies said. "Moving step-by-step, and teaching a company to continuously consider actionability of data they collect and analyse, is what really helps them grow in digital maturity.”
