Lionel Cherpin, founder of Empirik agency, was kind enough to partake in our interview. Discover his expert opinions on digital analytics.
Hello Lionel, can you introduce yourself briefly?
Hello, my name is Lionel Cherpin. I’m the founder of Empirik, a company I created in October 2012. Empirik is a digital marketing and web analytics agency. Our main services include digital analytics, tag management, natural search (SEO), paid search (SEM), and A/B testing. We specialise in advanced usage of analytics solutions, firmly guided by the principle of making data meaningful and transforming information into concrete action. We work for large and small companies from several different sectors. Our best-known clients include AFNOR, Legrand, Yves Rocher, Trigano, SEB group, and the Paris Philharmonic Orchestra.
From a more personal perspective, I have 15 years of experience in this exciting field, and I’ve known AT Internet for a long time because I started using its solutions back in 2003, when it was called Xiti.
You help many advertisers with their digital analytics strategy. What are some of the main issues they face?
The first issue is data reliability. That is an essential requirement. No analysis can be performed without absolute confidence in the quality of the data. You also must make sure the quality of the data does not change over time, and ensure that the tags don’t break with every product launch.
Once these requirements are met (and that is already no small feat), the main issue obviously becomes how to transform this data into effective action.
Our clients often feel a bit helpless and overwhelmed by this plethora of information.
As an agency, our primary role is to interpret this data. Experience is essential to properly assess whether an indicator is good or bad. But we cannot do it alone: it is crucial to contextualise the analysis and tailor it according to the client’s current activities, industry, and competitors.
This process can also be facilitated by creating customised dashboards, either directly within an analytics platform such as AT Internet’s, or with data visualisation tools.
Overall, I believe it’s best to take a “frugal” approach when it comes to data: It’s better to follow a limited number of indicators that are quickly actionable, rather than piling on dozens of KPIs.
We’re witnessing an explosion of performance optimisation solutions on the market. How do you guide your clients in picking the right tools?
There is certainly a broad range of choices — however there is one dominant player, which is Google, and alternative solutions, such as AT Internet. We try to remain unbiased and objective, while educating our clients about the potential risks of going “100% Google.”
As I said earlier, our passion isn’t installing tools; it’s transforming data into action. The advice we give clients regarding which tools to use is based on their needs, above all else, as well as their organisational structure and level of maturity.
What are the issues associated with Google’s dominant position in the market?
The data sampling that we are seeing in the free version of Google Analytics is a rather serious issue for sites with large audiences. Sampling has a serious impact on the reliability of data collected.
The main problem, however, is that Google is playing both sides of the court. Advertisers (for the most part) can no longer do without AdWords campaigns if they want to build an audience and generate revenue. So the same entity that is selling you ad placements is also measuring the results! Let’s face it: that can be dangerous if you’re spending hundreds of thousands or even millions on AdWords! It’s almost like giving the IRS full access to your bank account. Would you do that?
What should be the greatest strengths of an analytics platform?
It is imperative for an analytics solution to offer open systems, with APIs and technologies that are compatible with third-party solutions.
Secondly, it should be above the fold when it comes to the reliability of data collected, its position as an independent third party, and its ability to offer customised services.
Regarding the use of data (from collection to interpretation), what can make a difference in terms of efficiency?
Let’s say an advertiser wants to use analytics solutions to answer two key questions:
- What is my current level of performance?
- Which factors are behind changes in my performance?
Answering the first question is pretty simple, but the second is a whole different story.
Advertisers want to identify to what extent endogenous factors (acquisition campaigns, website developments, content creation, etc.) or exogenous factors (competition, seasonality, weather, etc.) have impacted the product’s performance.
Based on these findings, we will be able to make concrete decisions.
Therefore, a data project must be structured around this goal every step of the way:
- As you learn about clients’ needs, you will gain an understanding of the characteristics of their market, and their concerns about the relevance of all digital choices (acquisition campaigns, editorial strategy, social media presence, technical or ergonomic changes, etc.);
- Tag implementation must be a technical reflection of their needs, while ensuring that statistical reports remain readable;
- Finally, it’s essential to visualise the data in dashboards in order to quickly identify any unusual changes. The use of a data visualisation solution is also vital, in my opinion, because analytics data alone is not enough to truly understand how performance is evolving over time. The aggregation of third-party data (digital marketing solutions, ERP, CRM, etc.) can also make it easier to understand a particular phenomenon.
Once this setup phase is completed, we move on to analysis and monitoring. Communication between the web analyst and the advertiser then becomes a key issue, because analysis without context is just pure speculation.
In your view, what are some key technological investments your clients should make in 2017?
The explosion of mobile is creating new challenges in analytics. But tracking a mobile application is not always as simple as it is for websites. As I previously mentioned, I think that data visualisation tools will become increasingly democratised.
Also, DMP and marketing automation technologies are gaining momentum because they fulfil the promise of completely data-driven management: automated marketing purely based on data.
And which ones call for more caution?
I think we need to be careful with the issue of attribution. I’m not saying that we should go back to the last-click model. It’s always better to have a comprehensive view of all channels that contributed to a conversion.
But we should just be aware that a single transaction now involves a growing number of devices and platforms (mobile, computer, and tablet). The higher your average basket value, the more significant these cross-device phenomena may become.
Many thanks to Lionel for sharing his experiences with us.