Follow AT Internet - Digital Analytics Blog on Feedspot

Continue with Google
Continue with Facebook


The GDPR’s has just celebrated its one-year anniversary. To mark the occasion, AT Internet has created a comprehensive guide on how you can ensure that your digital analytics is 100% compliant with the regulation. To accompany the guide, we are running blog series and GDPR quiz to get you up to speed!

In the first blog, we wade through the flood of derisive articles on how the regulation has been ineffective so far in addressing data privacy issues. Are the attacks justified, or is it too early to criticise such a complex and ambitious project 12-months in?

Here are the top 12 criticismsof the GDPR – and the case in defence of the regulation…

1. There was total chaos when the GDPR came in, with 206,326 reported breaches in the first few months alone.

Although initially there were a large number of reports cases following the GDPR deadline, many were quickly resolved. Of the 200,000 or so breaches in the first nine months of the regulation, around 65,000 were initiated on the basis of a data breach report by a data controller. Of the 95,000 complaints, some 52% of the overall cases had already been closed, with only 1% facing a challenge in national courts.

2. For the consumer, the regulation has simply created an endless barrage of rapid-fire, pop-up privacy notices that nobody actually reads. This has caused consent fatigue and has done little if anything to improve individuals’ comprehension of their data privacy.

The implementation of the GDPR has in fact significantly raised the awareness of privacy issues for consumers. Following the Cambridge Analytica scandal, a poll in the UK showed that 72% of individuals had already changed their data permissions ahead of May 25th, 2018 and were planning on sharing less data in the future.

3. The GDPR offloads too much responsibility on individuals who have a limited understanding of the complexities of the regulation.

According to Eurobarometer results from March 2019 in an article by the European Data Protection Board (EPDB), the increase in queries and complaints confirms a rise in individuals’ comprehension of data protection rights – 67% of EU citizens polled indicated that they have heard of the GDPR, 36% of them indicated that they are well aware of what the GDPR entails.

57% of EU citizens polled indicated that they are aware of the existence of a public authority in their country responsible for protecting their data protection rights. This result shows an increase of 20 percentage points compared to 2015 Eurobarometer results.

This can only increase in the future…

4. Businesses have no idea how they can be compliant and what constitutes a data breach – as reflected in the large number of complaints. According to the IAPP more than 56% of respondents subject to the GDPR said they were far from compliance or would never comply and one-fifth said that “full compliance may be impossible.” Hoping to avoid the costs of incorrect or delayed compliance, many companies are moving forward with a risky “wait-and-see” approach as they struggle to understand the best course of action.

As a brand-new regulation, the GDPR will take time to become truly effective – confusion during the implementation phase is to be expected. It is also a monumental challenge for regulators as well as for businesses being regulated. Many organisations are still running through a backlog of activity required for compliance even so many months after the GDPR entered into law. This can be very slow work and requires an experienced Data Protection Officer (DPO) with both legal and technical expertise.

Regulatory bodies spent most of 2018 getting up to speed with their staff levels and internal procedures and finishing the last pre-GDPR investigations.

5. There is inadequate support and information available on how companies and individuals can be GDPR compliant. Local Data Protection Authorities (DPAs) are not taking the necessary measures to inform businesses and individuals on the regulation requirements.

Following a record number of complaints in France, the CNIL has affirmative taken actions to support operators and ensure the successful implementation of the GDPR. These include raising awareness for local authorities with a practical guide, thematic sheets (teleservices, security, Data Protection Officer (DPO) and local authorities, etc.) and a dedicated module in its online training open to everyone. They provide support via the CNIL website and an asset and a developer kit have already been published. CNIL is also in close dialogue with professionals and “network heads” to promote the development of skills in all sectors and support for data protection officers. In practical terms, they have tools available to everyone on their website including:

  • A way to check if an organisation has appointed a DPO for people wishing to exercise their rights or ask a question to an organisation that manages their data, including contact details publicly available via a dedicated search engine
  • an online GDPR training course open to all since March 2019, already attended by more than 35,100 people, including 6,900 who have obtained a certificate of success
  • a model for a register of processing activities, which makes it possible to identify all data processing operations and to have an overview of actions relating to personal data
  • open source software to conduct a data protection impact assessment (DIA), which is obligatory for some processing operations

With 2,044 data breach notifications in France and 89,271 at European level, more than 19,000 data protection officers (natural or legal persons) have been appointed by more than 53,000 organisations. There has been also an influx of requests for information from professionals wanting to apply the GDPR framework and a growing mobilization of professionals and individuals on data protection – the CNIL website has had more than 8.1 million visits over the last year.

6. A vast new Data Protection Officer bureaucracy has been created in organisations and the associated costs are huge.The GDPR makes it a legal requirement for all companies doing business in the EU to appoint a DPO if their core business consists of data processing activities. If they process special categories of data (sensitive data, like biometrics, ethnic, health data etc.) they must appoint a one regardless of the scale of processing – public institutions also have to appoint one. Failure to appoint a DPO can carry fines of up to €10 million or 2 percent of a company’s global turnover, whichever is higher. This is expected to generate the need for up to 28,000 DPOs in the EU and the US and 75,000 worldwide.

Although initial reports warned that the GDPR would create the need for vast numbers of DPOs, many companies will outsource and the likely figure is a considerably lower. Smaller companies are also allowed to share a DPO and they can be in the form of an existing employee as well as a new hire.

The GDPR may seem draconian by demanding that each organisation has its own DPO, but data regulation was in drastic need of modernisation to bring it in line with a rapidly changing online ecosystem and make the law fit for purpose.

By fostering a culture of data protection within an organization, the DPO will save money by avoiding potential fines.

Businesses are realising that aside from focusing on the fines they might receive, GDPR compliance can be viewed as an opportunity to engage with customers and build loyalty, as a driver of increased trust and overall business growth. This is one of the reasons mobile banking, streaming services, and tech companies, have been extremely agile when it came to GDPR compliance. DPOs must be allowed to act independently and cannot receive orders from the management which also creates an atmosphere of confidence in a company’s data privacy practices.

7. EU data authorities are not adequately staffed or coordinated to handle the demands. EU regulators claimed they had insufficient resources or funding to cope with the new GDPR workload and that they need needed a “substantial increase in resources and staff”.

2018 was a transition year and the various agencies needed time to staff up. EU data authorities have stated that “while the cooperation procedures are robust and efficient, they are time and resource intensive – SAs need to carry out investigations, observe procedural rules, coordinate and share information with other supervisory authorities.”

The Irish DPA ‘beefed up’ it’s 27 staff to more than 130 to cope with the steep rise in complaints and this is expected to rise to more than 200 in the next year.

The framework surrounding GDPR fines is still being created. One of the likely reasons for the uneven GDPR fines across countries, as well as the slow associated process, is that those in charge of making legal decisions don’t have legal precedents to guide their actions.

In 2019, people should expect regulators to become more concise in their interpretations of the law.

8. There is little international coordination – the DPAs of several EU countries also need to harmonise the wide range of fines being handed out.

To provide more guidance as to how a data protection agency should calculate the amount of a fine, the Dutch data protection agency, the Autoriteit Persoonsgegevens has released a framework to determine how severe a penalty will be. The Dutch framework (in Dutch) has four categories of violations, and each category has a defined “default” fine, along with a range of possible fines depending on the severity of the violation.

The UK’s Information Commissioner’s Office (ICO) are coordinating with both the Dutch and Norwegian DPAs to create a harmonized framework or matrix ‘toolkit’ for various watchdogs to give them a foundation for calculating fines in the future.

Data Protection Authorities across the EU will soon be publishing annual reports, which should give us a wider and better picture of the level of compliance one year  into the GDPR.

9. The threat of huge fines never materialised, with the exception of Google’s token slap on the wrist by the French. Only 91 fines were imposed in the first eight months of GDPR and (removing the €50M Google fine) the average GDPR penalty was approximately €66,000. Even Google’s fine was merely a snip of the company’s $136.2 billion 2018 revenue, approximately .04% – far below the 4% potential.

The fines that were handed out were extremely measured and reflect that DPAs are not out to trip companies up and were moving forward diligently.

The ‘token’ GDPR fine from French CNIL against Google for “lack of transparency, inadequate information and lack of valid consent regarding the ads personalization” can be viewed as a shot across the bows for the tech giant. It also demonstrated that DPAs have the will and muscle to implement fines on larger companies and that the major players are not untouchable.

The first GDPR fine was for German social media platform Knuddels. Although the data breach was a major one, compromising the email addresses and passwords of 330,000 users, the company’s proactive approach to addressing the breach meant that that the fine was only €20,000. They not only responded quickly by notifying the German data protection authorities and customers, they also worked rapidly to bring in the recommended security procedures. 

Stefan Brink, state data protection commissioner for Baden-Wurttemberg, commented on its €20,000 fine against Knuddels: “The LfDI is not interested in entering into a competition for the highest possible fines. In the end, it’s about improving privacy and data security for the users.” Although it remains to be seen whether other DPAs take a similar approach, this case sets a valuable precedent.

In March 2019 when the Polish DPA showed a highly measured response by only issuing a €219,000 fine to a company who failed to inform six million individuals that their personal data were being processed. While the Danish DPA only fined a company €161,000 for holding on to personal data longer than allowed under GDPR.

In June 2019, the CNIL dished out a second fine of €400,000 fine to the French real estate company Sergic. Although the breach was a serious one, i.e. a lack of due diligence in addressing vulnerability and the fact that the accessible documents revealed very intimate aspects of people’s lives, CNIL took the size of the company and its financial strength into account and imposed a measured fine. 

10. It is a weak start from the regulation considering that DPAs have already handled around 100,000 self-reported breaches and user complaints.

The complaints are still being processed and the impact has yet to be felt, in that we haven’t yet seen significant enforcement activity, both in terms of volume and amount.

Many cases are also ongoing on an international and national level – there have been 446 cross-border cases which tend to be far more complex and take longer to resolve. Of the large number of reported to EEA Supervisory Authorities (SAs), over 144,000 queries and complaints and over 89,000 data breaches have been logged. 63% of these have been closed and 37% are ongoing.

11. The GDPR is ineffective against the tech giants as they have interpreted the regulation liberally at best. There is also a major lack of transparency. Facebook has been in the headlines for its reintroduction of facial recognition software and data sharing with its recently purchased subsidiary WhatsApp. Paul-Olivier Dehaye (a privacy expert who helped uncover Facebook’s Cambridge Analytica scandal) stated that “big companies like Facebook are 10 steps ahead of everyone else, and 100 steps ahead of regulators – there are very big questions about what they’re doing.”

The EDPB released a report stating that out of the 206,326 cases reported under the GDPR across the 31 countries in the European Economic Area (EEA), the national Data Protection Agencies (DPAs) have only resolved only 52 percent of them. As regulators work through this backlog, businesses can expect more fines of greater amounts in the coming months.

The CNIL fine against Google was a critical first step, and created an important precedent to evaluate the impact and reach of GDPR. More fines are to be expected in 2019 and beyond.

12. Google continues to turn a blind eye in terms of its failure to get consent from users “before sharing data among its fast-growing line of networks and products—from YouTube to Google Photos to Gmail and more.” Nicolas Vinocur at Poltico refers to a significant loophole being exploited by the tech giants. Chief policy officer at Brave, Johnny Ryan

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Born in the early noughties, the Hackathon has traditionally been a tech-based “competitive event in which people work in groups on software or hardware projects, with the goal of creating a functioning product by the end of the event”. However, here at AT Internet, we have taken the hackathon one step further!

For the fourth year running, this year’s Ship It was a searing, sizzling, sweltering, romper stomper of an event! It was action-packed with creativity, inspiration, perspiration(!), slamming music, moshing and enough luscious food and drink to feed an army – in short, IT ROCKED!


AT Internet started running Hackathons (a combo of ‘hack’ and ‘marathon’) back in 2016 with the aim of accelerating innovation and showcasing the dynamic and progressive nature of the company.

Ship-It gives employees the chance to let their imaginations run wild and rise to the challenge of developing and delivering new innovations in just 24 hours. It’s a great way to let teams show off their skills and generate project ideas that could be useful in the future – both internally and for our customers.

It’s also a superb way for employees from different teams to meet up and share the positive company vibe!

Feeling hot hot hot…

This year’s Ship-It coincided with Bordeaux’s first heat wave of the summer and temperatures rose to over 40°C. Luckily the re-fuel team was on hand to provide refreshments in the form of smoothies and cold drinks galore! And the survival supplies didn’t stop there… There was a regular flow of snacks and sweets throughout the day to feed the neurons of our creative teams. There was also a food truck serving burgers and chips in the car park to soak up some of the ice-cold beverages on offer at the company party.

Superkramp (AT’s home-grown band), took the Thursday evening by storm with a pumping concert on the terrace. Employees relaxed on the comfy pallet furniture had been built that very afternoon by one of the Ship It teams, and those in the mood got stuck into the mosh pit – ever seen a company CEO crowd surf?

Breakfast on day two featured rounds of sweet pastries and healthy smoothies. For those that needed a boost of rocket fuel to get through the morning, there were non-stop rounds of English bacon rolls, fried up by our resident Londoner and bubbling to the sounds of Bob Marley.

The projects – teamwork, hard graft and a touch of genius!

This was the 4th edition of our Ship-It Day and the teams were looking stronger than ever. A total of 14 groups broke sweat all hours to come up with an intriguing selection of projects. The topics covered everything from creating software that automatically generates digital charts by capturing the image of a physical one, to installing a private cloud on our on-premise infrastructure to build on-demand services for end users. There was even a project to design a new customer service chatbot!

And the winners are…

The clock stopped ticking at the 24-hour mark – 1pm on Friday. Each team then had 10 minutes max. to present their project and convince the audience why it deserved to win. The presentations were hosted by our marketing director who added his touch of je ne sais quoi…

Coming in Third…

Analytics for Green

The “eco project” looked at the impact of day-to-day actions at work on the environment and is part of AT Internet’s strong commitment to ecological issues. The team aimed to create ways to collect data and track activities such as water/electricity consumption, carbon emissions from the daily commute as well as more specific behaviour, such as which type of rubbish bins are most used, and how many bikes park up each day.

They then displayed the information in a dashboard and report that could be regularly shared with the company to track our progress. Using this data, they then aimed to use automation to act on the measured results. For example, controlling when lights/screens turn on and off using technology such as Raspberry PI, Jeedom, temperature measurements, controlled power sockets, automation scenarios, tagging, and dashboarding.

In at second was…

The Real data collection counter!

This project looked to design a physical counter that could be used throughout the company to display a wide range of information. They came up with a large illuminated display counter with 10 digits (7 segments with RGB LEDs) and displayed how it could be used to display the number of hits collected on our site, hits collected for a specific customer, as well as more specific analytics data. A truly colourful and practical idea that could be implemented straight away!

And storming in at first place!

Explorer Quadricolor

The idea that won over the crowd in the end was Quadricolour – a way of creating colour variations in Analytics Suite, using Cascading Style Sheets (CSS). The presentation and idea were simple but game changing – to give Analytics Suite a range of colour themes that can be personalised according to the customer, personal preference, weather and specific public holidays etc. They showed off a wide spectrum of colour variations – fixed themes such as blue, green, orange, pink, red – colour personalised to individual customers – and even various event themes such as Christmas (snow theme), Halloween (pumpkin theme), and St Patrick’s day (Irish theme). The possibilities are endless, and the idea could be put into practice immediately.

The winning team received a tour of Bordeaux’s Cité du Vin complete with wine tasting, followed by dinner at the Halles de Bacalan high-end gastronomic food hall next door. A superb gift for a worthy team!

A big shout to…

…everyone involved who worked so hard to make this event a resounding success. To the events manager for organising everything, the refuelling team for keeping everyone running on all cylinders, the band who kept everyone rocking, and all the devoted and talented teams behind the wealth of projects. Above all, thanks to a certain CEO for making the company such a fun and inspirational place to work!

See you next year!

Article Ship-It Day 2019 – it was a sizzler! first appeared on Digital Analytics Blog.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Is your analytics data GDPR-compliant one year into the regulation?

To mark the one-year anniversary of the implementation of the GDPR, we have released a comprehensive guide with everything you need to know about being compliant.

Here’s a quiz to accompany the guide and get you up to speed with the ins and outs of the regulation. Are you ready to take the GDPR challenge?


It’s a year since the GDPR came into effect. To mark the occasion, we’ve updated our GDPR guide which is available here. To get you in the mood, here’s a refresher quiz – 

The GDPR in 15 questions!

Start Quiz


Your answer:

Correct answer:


You got {{SCORE_CORRECT}} out of {{SCORE_TOTAL}}

Hungry for more? Subscribe to our newsletter!SUBSCRIBE AND GET MY RESULTS

Receive our latest digital analytics content by email: downloadable resources, invitations to our webinar on product and market news, and our latest articles along with news from our Digital Analytics Blog.

Skip this step

Your Answers

Article 15 questions to test your GDPR knowledge first appeared on Digital Analytics Blog.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Getting you up to speed with all the latest in the Catching up

Leaky AdTech – the investigation continues…

The UK Information commissioner’s office (ICO) has described the AdTech industry as “immature” in terms of personal data protection in a recent report. You can read the article here in French.

The ICO stated that: while AdTech has made progress in the collection and management of personal data, targeting technologies, automation, retargeting and auctions, there has been a significant increase of breaches in the protection of personal data. This particularly applies to the use of data in real-time bidding (RTB) where advertisers compete for the purchase of online advertising space.

With the skyrocketing of the global smartphone market, advertising campaigns have unprecedented reach and RTB is a boom industry. The current $6.6 billion value of the RTB market could reach as high as $27.2 billion in 2024.

By comparing RTB processes with the requirements of the GDPR and the British Privacy and Electronic Communications Regulations (PECR), the ICO underlines major concerns globally and systematically, about the level of RTB compliance. Thousands of organisations process billions of auction requests in the UK every week – however there is little or no consideration for the requirements of data protection laws on international transfers of personal data. The profiles created on individuals are extremely detailed and are shared repeatedly with hundreds of organizations for each request to participate in an auction – all this without prior agreement. Individuals also have no guarantee that their personal data will be secure within the ecosystem.

Among the main players in the market are Google, Facebook, Yahoo, AppNexus, BrightRoll, Adobe, Criteo, Salesforce, SpotXchange, TubeMogul, Tremor Video, PubMatic, Smaato, Yandex, Rubicon Project. The battle for our data privacy is far from over…

Weighing up on cookie management

To mark the first year of the GDPR, Empirik has analysed the cookie management policy of the 100 most visited sites in France. They created a barometer to measure the impact of the regulation and look at how companies are faring up after the first 12 months.

The results came out with three categories:

1. Companies that have a cookie management policy that seems to respect the GDPR – most of the basic rules seem to be well understood by the top 100 websites in France.

More than 9 out of 10 observed sites have implemented a cookie banner on their site, and almost all of them have a permanent link (often placed in footer) to a page describing the cookie policy. Nearly 95% of sites also present and distinguish cookies by category, and then include at least one description of each of these categories: functional, analytics, social, personalization or advertising.

2. Companies that make refusal and personalisation confusing and challenging for the Internet user.

Many sites put obstacles in the way of Internet users who want to refuse or personalise the collection of cookies. Even experienced Internet users who want to manage their preferences are facing difficulties. Given the limited number of sites that allow you to refuse directly from the banner, the user must make more effort, in time and clicks, to customize his preferences, while acceptance often requires only one…

3. Tracking Analytics is often launched by default – even after the refusal.

Of the 85% sites they observed, tracking analytics starts before clicking on an “I accept” button. Among these sites, 92% start tracking immediately upon arrival at the site. While there is no ambiguity about advertising cookies, obtaining consent is less certain for Analytics cookies. This calls into question cookies from Google Analytics tracking that does not qualify for the CNIL’s consent exemption.

Calling all Mario Kart fans!

Whether you were a Yoshi, Bowser or Mushroom player back in the day, it’s now possible to determine the best Mario Karts character using data science. Using the concept of Pareto efficiency, you can work out the optimum choice for the most effective character in terms of speed, acceleration and kart/tire combination. Read here to never lose a race ever again!

Bienvenue à Bordeaux!

The Glassdoor website, which publishes the ranking of the French’s preferred employers every year, has just revealed which companies have the highest ratings outside the Paris region. This is obviously relevant for the 55% of Parisians aged 28-45 who are desperate to get out of the big smoke and head for the provinces. Find out more here (in French).

Aix-en-Provence tops the list, followed by Lyon and Nantes among others. Storming in at 8th place with an impressive score of 4.3 is a certain digital analytics company based in Bordeaux. You can read the full list of their reviews in English and French here.

The Glassdoor site, which allows employees to assess their working environment anonymously, also points out that many of the companies on the list currently offer dozens of job offers.

Transform your view of the digital economy and society

Transformation Maps, a new platform developed by the World Economic Forum is a completely different sort of visualization. The Swiss non-profit calls it a “dynamic knowledge tool” and it charts the interactions between dozens of industries, countries, and issues that shape the world.

Try out this interwoven map of all the permutations surrounding digital economy and society…

See you next time on the Internets!


World Economic Forum

Article Catching up: leaky AdTech, cookies, Mario Kart and Bordeaux calling! first appeared on Digital Analytics Blog.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

The digital world has evolved beyond all recognition over the last 15 years. The expanse and sophistication of websites has grown exponentially along with the multiplicity of online offerings. This in turn has transformed the online behaviour of consumers, along with their needs and expectations. When trying to measure and understand this behaviour, digital analysts have always used impressions, CTR & conversion rates as their “go to” metrics. However, in today’s complex online environment, classic web analytics metrics no longer make the grade – at best they give you a topline view of your site’s performance.

The latest advanced data processing and storage capabilities allow you to carry out far more complex analyses – providing you with a clearer, more precise, and more in-depth understanding. By going beyond analysing traditional website analytics metrics – page views, click-through & conversions rates, you can create significant value by observing user paths on a site, i.e. how users go from the landing page to the conversion or exit page. Based on this information, it’s possible to make a qualitative assessment which will reveal ways to optimise a website’s performance.

Here are three types of qualitative analyses that can help boost the measurement and effectiveness of your sales funnel.

1.      Visualise your top consumer paths

Gaining a macro understanding of major flows on your website is a great starting place: visualising your top 5 paths or your top-viewed pages following a page of interest will provide valuable insights into how most users interact with your platform.

It’s also useful to add a segment on specific users to your analysis and observe how these top paths differ according to the various clusters. For example, by comparing converting and non-converting users, or by segmenting based on engagement figures, you can understand the varying needs of your users and tailor your services accordingly. They may have different interests and are looking to consume different content: rich media vs. written content for example. The complex consumer interactions
with a brand’s platform is unique to each brand –
and must be customised. It’s only by segmenting
your consumer base that will you be able to analyse their individual and unique needs.

2.      Identify synergies & areas of improvement

Most analysts already have a decent understanding of which user flows are critical to paths for their business – if not, a macro view described above will show you which of your user paths are

important. With shrinking attention spans (!), an effective way to drive optimisation is by shortening them and making the user experience as straightforward as possible.

It’s also important to pay specific attention to each step along the paths and optimise those with high users churn rates – a conversion funnel is only as good as its weakest link! You can take this further by distinguishing between users who drop off the funnel and leave the website, and users who drop off but stay on the site. Their intention will be different depending on their respective category.

3.      Peripheral paths: recognising patterns outside your key conversion paths

Although most of your time should be spent optimising your most valuable conversion funnels, don’t forget to keep an eye on recognising patterns outside your key conversion paths. If you’re too focused on one specific part of your site, you may be missing out on a smaller (yet important) part of the traffic. By identifying correlated content on the website, you can determine if users who read blog article A and more likely to convert than users who read blog article B.

AT Internet

At AT Internet we want you to be able to extract as much value as possible from our solution.

The Navigation tool seamlessly integrated into Explorer allows you to extract qualitative information. This provides our clients with a solid and genuine understanding of visitor behaviour to be used on top of classic web analytics metrics.

The Navigation tool gives you the ability to quickly identify your top consumer paths. You can then take this thought process further with the tree-view analysis – an interactive tool to explore and reveal insights
and enable you to solve your user flow issues over the long term.

Article 3 tips for optimising performance analysis: above and beyond click-through & conversion rates first appeared on Digital Analytics Blog.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

With its latest ITP system update for iOS and macOS, Safari is blazing the trail for increased privacy among the major browsers. Earlier this year Firefox also updated its privacy policy, and even Microsoft and Google are set to give users more control over their cookie options. Are we witnessing the start of a major new trend for web browsers?

ITP has been around for a while now. Introduced in June 2017, Safari’s privacy feature began cancelling any 3rd party cookies from being deposited on a browser by default. In March this year, Firefox introduced new privacy rules with the release of 66.0, and Safari came out with ITP 2.1 which also restricts 1st party cookies.

With 14 and 15% respectively, the combined hits generated by Firefox and Safari account for nearly a third of the internet. Blocking 3rd party cookies by default could therefore have a seismic impact on companies that rely on cross-site traffic measurement and retention analyses.

The technical consequences

ITP doesn’t only affect cross-site tracking ­– by drastically limiting the lifetime of first party cookies, it potentially prevents a user from being recognized from one session to the next.

This leads to distorted figures with an inflated number of unique visitors, audiences that do not reflect reality, and a complete fragmentation of conversion paths. In practical terms, on the basic indicators, session and new visitor figures are artificially inflated. In terms of allocation, ITP and especially ITP 2.2 have a bias in favour of free channels (less likely to include parameters in URLs) and contact points at the end of the several-day process (favoured by the limited cookie duration).

Firefox 66 has just introduced a pre-checked box in the new browser settings that automatically prevents the deposit of 3rd party cookies; while it’s private-browsing mode cancels all hits. Safari already blocked 3rd party cookies by default – its new rule now means that all remaining cookies that are not ITP compliant are deleted after 7 days.

Safari’s requirements state that cookies must be http-only, i.e. server-side, as opposed to client-side management (JavaScript). They must be explicitly set with a secure flag and only on https – secure http. Their collection domain also needs to be the same as the website domain, i.e. 1st party.

User analysis and the ITP

To ensure that their traffic is not significantly affected by ITP, analysts today need to be able to rely on a powerful and high-performance analytics solution. Analytics tools will always be impacted by web browser evolution, especially with the increasing shift towards data privacy. However, it is possible to be fully privacy-compliant and still maintain a clear and detailed overview of your traffic. CDDC (Custom Domain Data Collection) is an AT Internet solution that collects and sends data directly via your own domain name (instead of AT Internet’s domain) so that tracking scripts are not disabled, making it possible to retrieve the majority of traffic lost due to ad blocker usage.

Another key to your analysis is identified visitors. By using identified visitor ids between your sites, you avoid spreading their information to external sites – you can therefore carry out cross-site analysis without affecting user privacy. This is the time to get visitors to subscribe to your sites.

In terms of 3rd party cookies, their days are numbered. For anonymised visitors, the domain structure will eventually need to be updated not just for analytics but for all resources. To stay ahead of the game, you can avoid the impact of ongoing restrictions by adjusting your measurements to ITP standards. Analytics solutions such as AT Internet already offer its customers ways to analyse anonymised visitors with ITP activated on the browser.

If your users rely on ITP 2.2 and try to access one of your decorated links, your site’s cookies generated by JavaScript will be limited to one day storage.

At the leading edge of the digital analytics market, AT Internet’s solution is fully aligned with the evolution of the major browsers. And CDDC is your perfect ally for handling ad blockers and matching the latest browser privacy requirements. Some of our customers have already recovered a significant amount of their traffic using CDDC – no need to worry about technical infrastructure on your end – AT Internet teams will put everything in place.

Please don’t hesitate to get in touch with our support teams for more information!

Article Intelligent Tracking Prevention – redefining the cookie ecosystem first appeared on Digital Analytics Blog.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Jon Snow approves this edition of Catching Up.

Today’s AI

People often to look to the future (especially sci-fi) to talk about Artificial Intelligence. But what does it mean in real terms today? How will automation affect the working world in 2019? This study looks at two practical examples. In Agtech (connected agriculture), super productivity boosted by big data is struggling to produce results. There are logistical challenges such as the need for sensors that require a large investment to redesign agricultural facilities – not forgetting that the countryside often has poor internet connection. Farmers interviewed in this study criticised the time they had to spend in their offices instead of their fields. Data interpretation is also a problem. The tools that are supposed to save time obviously require time to learn and understand, when they are actually used (the software embedded in John Deere tractors is often unused). Technology is therefore moving faster than the development of the profession. In the retail sector, the automation of cash registers is changing the way work is organised. In terms of the benefits, opinions differ between employees and managers (lack of job stability vs. profitability). In reality, automatic cash registers require considerable time and effort (monitoring, on-hand assistance) by actual employees, who in turn contribute to the success of today’s automation. Imagine if there wasn’t a human being on hand to unlock the cash register. Today’s AI is meeting the real world head-on.

Follow the voice

What are the challenges of audio for digital marketing (and analytics tracking)? Consultant Nicolas Hinternesch looks at the changes brought about by the arrival of voice assistants in the digital ecosystem, and predicts major shifts in the future. The way you interact with your voice (vs. the screen) is already a profound behavioural change in itself. Searching by voice can only offer one possible answer. This therefore takes away results pages along with access to the websites of brands and content producers. This new dynamic will potentially transform advertising models and complicate consumer engagement and retention strategies on a website. Brands will naturally have to adjust, just as the BBC has intelligently adjusted its service offer (design and content) to audio practices. In terms of analytics, the challenge is to integrate this new voice data source into the current scope of analysis. To do this, it is essential to carry out personalized tracking. While the major manufacturers offer various metrics, they are often broad and limited. The ideal solution is to work with an analytics provider that allows sufficiently flexible tagging customisation.

Long speeches

#cicero (French link here) is a platform for analysing the language of political speeches. It aims is to provide people with a clearer view above the mass of information reported by the media as well as understanding the positioning of our elected officials on major issues. By typing in a keyword (ecology, tax, etc.) the user can access candidates’ statements (video, radio, social network formats) and compare their positions. With many political speeches failing to address the main issues in depth, it is important to be able to look at specific subjects in detail. The concept is called meta-discourse (link in French) and is only used to identify issues. This project was initiated by the IDHN (Institute of Digital Humanities) and aims to provide a more objective political voice to reduce the lack of trust in politicians and put people back on the path to the polls.

Video standard

4 media measurement bodies (Médiamétrie, Barc, Numeris and Video Research) have come together to form the international alliance GAMMA, an initiative to standardize video audience measurement worldwide. The aim is to develop common standards and processes to align all video audience measurement. In France, the current Computer Video Audience reference measurement is based on a hybrid sampling methodology (16,000 individuals) augmented by certified site-centric data (ACPM). We therefore expect to see a range of developments in the coming months. You should also remember that any good Web Analytics tool can provide you with accurate behavioural and KPI metrics on video consumption (playback time, time spent buffering, full playbacks, playbacks with displacement, etc.) with the huge advantage of putting this analysis in perspective with the performance of your entire digital device.


These interactive visuals show how ideas, information or even viruses spread across a network. Adjust parameters such as the rate of transmission or the degree of immunity and watch the infection (or rumour) spread.

See you next time on the Internets!



Article Catching up: AI in 2019, audio/video measurements and virality first appeared on Digital Analytics Blog.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Following the release of our digital analytics guide for the media industry, we are running a series of offshoot articles that focus on specific challenges in today’s rapidly evolving online publishing world. If you haven’t read them already, please check out our blogs on working with ad blockers and optimising your audience retention strategy. In the third instalment, we delve into the challenges faced by media publishers when they try to monetise their content via the use of paywalls.

Part 3: creating a paywall strategy

Consumers have enjoyed access to free online newspapers and magazines since the birth of the internet. However, over the last few years, an increasing number of online media publications have been shifting away from advertising-funded journalism and are charging for their content – and more people are ready to pay for what they read online. In this blog, we look at different approaches toward subscription and paywalls and how AT Internet can help you to optimise audience interaction with your online publications.

What is a paywall?

Paywalls are systems that prevent online users from accessing specific website content unless the user pays a fee. They can restrict access to a specific site, story or publication and there are varying levels of restriction.

Soft paywalls such as metered access provide readers with a number of free articles before asking them to become a member and subscribe in order to access more content. By adopting data walls, businesses can collect more detailed information about their audiences and develop their customer profiles. This personalises the reader experience as it shows them the type of content they can find on a site and whether it’s valuable for them. The effectiveness of metered paywalls has been repeatedly called into question.  ‘Metered’ business models can be off putting for readers who have a limitless range of other media sources at their disposal. They also cause user churn as they “reward future traffic rather than the authority and prestige that come from years of honest, serious reporting.”

The Telegraph abandoned its metered paywall for a freemium model in 2016 as it was impossible to differentiate between subscribers and those who were reading a story for free. They were therefore unable to target their content appropriately and were failing to retain their audience. They now run a payment system where free and premium coexist to fuel the subscription part of the business.

Nevertheless, metered paywalls can be effective revenue models. The New York Times reportedly spent $40 million over 14 months to build the infrastructure required to collect necessary data and present their readers with a metered paywall. Although it involved a large investment and significant marketing budget, the online publication reached 4.3 million subscribers at the end of 2018 and this continues to expand. COO Meredith Levien stated their main focus was the “breadth of reporting” and the goal was to improve their overall value by adding journalists and extending their product range.

Freemium content paywalls restrict access to select “premium” articles, while the reader can view all the available free articles. By subscribing, the reader pays a recurring monthly amount for unlimited access to all of the website’s content. Freemium allows content providers to maintain their advertising revenue while boosting conversion and retaining the most volatile readers. Although they can enhance audience loyalty to a particular website, they can lead consumers to change their habits and switch to alternative news sources, search for loopholes or other creative workarounds.

The Guardian operates a freemium model where visitors can access certain articles for free but have to view adverts. If they subscribe to the daily edition, the ads are removed, and the user can access premium articles. The publication also runs a support message at the bottom of each article – where it politely requests that users pay a monthly fee that will allow them to continue producing free articles. Online publisher Digiday has taken freemium paywalls a step further – as well as offering premium research articles, they provide their members with exclusive site access and a copy of their print and digital magazine.

Hard paywalls typically only display an article title and a few paragraphs before demanding that the reader pay. They are becoming increasingly common, particularly among professional and financial ‘niche’ publications. Although renowned for causing a sudden ‘drop off’ in the digital audience in the short term, they can be effective over the longer term – subscriptions have the effect of aligning a content provider’s goals with that of their readers. Creating accounts increases the connected user base and generates a higher average income per user.

The Times (London) introduced a hard paywall in 2010 and reportedly saw a 90% dip in visitors. It has since managed to recover a large proportion of its lost audience over time and through the quality of its reporting and reputation. It now boasts well over half a million daily readers and hails the success of its paywall strategy. Despite experimenting with alternative methods of gaining revenue from its online media, The Times is resolute in keeping its hard paywall. According to Digital Director Alan Hunter. “The [online audience] goes elsewhere to read news to be informed, but they come to us to get well informed. They can go deeper beyond the story to find out what’s really going on.”

The Financial Times will not allow readers to see anything other than the title and subheading of their articles and they have to choose from a range of subscription fees. However, their paywall model has worked as its 900,000 digital subscribers represent two-thirds of its total revenue (overtaking income from advertising). CEO John Ridding commented that “A lot of the industry was too quick to dismiss the ability to charge for content. If you have something that differentiates you, something that makes you special — it could be a brand identity, it could be a columnist, it could be a sector of coverage — you have the ability to charge.” The FT defends hard paywalls by saying that it wants to “achieve the habit in digital that people used to have in print” – and that metered models are counter-productive as they “ration” content.

Dynamic paywalls – the solution to monetising your online content

Dynamic paywalls address readers by customizing content and targeting users according to their level of commitment. With personalized segments, publishers can reach readers based on their specific profile. By collecting, interpreting and actioning data around each individual reader’s online behaviour, it’s possible to assess their likelihood to subscribe.

The Wall Street Journal implemented a dynamic paywall in 2018 based on a propensity score. It measures more than 60 signals such as whether the reader is visiting for the first time, the operating system they’re using, the device they’re reading on, what they chose to click on, and their location etc. They then used machine learning to create a flexible paywall that provides the user with precise targeted content and payment options.

Poool is a French startup specialising in intelligent paywall solutions. Founded in 2015 and an AT Internet partner, the company has brought a fresh approach to the paywall market. AT Internet’s digital analytics data can be combined with Poool’s paywalls to help you understand audience behaviour, identify key segments, determine the most effective paywall scenarios, and measure KPIs.

Read on to find out more about how you can benefit from advanced web analytics and dynamic paywalls with AT Internet and Poool.

Digital analytics: the key to understanding and segmenting your audience

When deciding which types of content and how much should go behind a paywall, it is important to understand your audience’s behaviour – and what drives them to purchase an article or subscribe to a site. By using digital analytics data, it is possible to segment your non-subscribers and offer them customised paywall options.

Reader visitors can essentially be grouped into three types:

  • ‘Fly-by’ are one-off visitors that have viewed less than three pages on your site and are yet to engage with your brand. Allowing them to access your content for free, e.g. with a metered approach with complementary articles is a way you could convert them into occasional or regular visitors.
  • ‘Occasionals’ are those that know your brand but don’t regularly consume your content. Depending on your site, one approach is to ask them to sign up for a free account to continue reading your articles.
  • More regular and loyal visitors are known as ‘regulars’ and they access your content on a frequent basis – they often already have an account. Using their data is useful to analysing what type and how much content they consume and set out a ‘pay-per-article’ approach or personalised paid subscription offer.

The profile of a likely subscriber can be summarised as follows:

  • Reads 5+ articles a month
  • Subscribes to a newsletter or has provided their email address
  • Follows your brand on social media
  • Lives in your market area and reads local news
  • Reads multiple categories of content
  • Accesses your content across multiple devices

Quality digital analytics can also give insights into visitor interactions, their multi-device usage, and frequency of visits, etc. It can give you more information on the categories of pages and types of articles viewed on a site – and help you to assess if your paywall is effective across your full range of articles, e.g. news, sport, or just one type.


Founded in 2015, Poool specialises in the implementation of intelligent paywalls for its customers to optimise revenue from access to online content.

The company’s mission is simple: to encourage teams to regain control of their paywall strategy. Then to enable them to rapidly test and optimise their approach, whether it’s for acquisition, engagement or monetisation.

Poool allows publishers to set up specific rules to access content and charge for it, depending on the profile of each reader. By being able to control content access rules, marketing teams can offer audiences more relevant experiences and by working on each lever, average revenue per user can be optimised – for subscriptions, ad impressions, one-off payments, or data collection.

The solution is highly accessible and adaptable according to customer needs. It is scalable and is implemented gradually via a ‘test and learn’ approach.

AT Internet’s digital analytics data can be combined with Poool’s paywall solution to help you understand audience behaviour, identify key segments, determine the most effective paywall scenarios, and measure KPIs.

Case study – Poool in partnership with Boursier.com

Boursier.com is French online publication that specialises in economic and financial news. The site had been operating with a freemium paywall model, reserving premium content for its subscribers, and were looking to optimise the way it interacted with its audience. Boursier.com’s main market advantage is its expert content that has a strong appeal to its audience. In terms of its monetisation process, it asked users to register on the site to become a ‘member’ before giving them subscription options.

By working with Poool, they developed a strategy to take advantage of their open content to boost the membership base. It also aimed to avoid any impact on the number of viewed pages which could disrupt the guaranteed advertising revenues.

After 3 months of testing, the results were highly positive and led Boursier.com to deploy Poool on 50%, and then 75% of the audience consuming free content. Their objective is to continue to increase account creation as well as developing new ideas: optimization of advertising revenues, anti-ad block monetization, and the differentiation of routes according to devices. With Poool, Boursier.com has increased its membership base by 25%.

Paywall trends

The perception of paywalls has changed significantly over the last few years. Back in 2015, many publishers were convinced that consumers were unlikely to ever pay for online content and were increasingly blocking ads. Fast forward to 2019, and all signals point to a future where considerably less revenue will come from advertising and the ad business model will be supplanted by subscription-based paid content. The editor of the Spectator went as far as to say “Paywalls are the only future for journalism: if the quality of writing is good enough, then people will pay”.

With more and more advertising revenue going to the tech giants who can offer advertisers unduplicated reach, targeted ads and low overheads, an increasing number of publishing companies have launched digital subscription models – including the New York Magazine, Bloomberg Media, The Atlantic and the Condé Nast magazines Vanity Fair, The New Yorker and Wired.

This has also come about through a shift in the audience approach to consuming paid content – primarily thanks to subscription models on platforms such as Netflix, Amazon Prime and Spotify. A recent Deloitte study predicted that by the end of 2020, revenue from subscription will be on an even par with advertising for digital publishers – as recently as 2012, revenue from ads covered about 90% of the market. These days audiences are not necessarily opposed to paying for the media they consume – they are often looking for accessible, reliable, centralised and credible content – with an eye on supporting local news and journalism.

There are also a range of other factors that have caused a shift in reader behaviour – the numerous fake news scandals that have called into question the trustworthiness of major social media platforms, and the overwhelming dominance of the tech giants in the digital advertising market. Audiences are increasingly inclined to pay for quality reporting from a source they are familiar with and trust. This has inevitably fuelled the growth of subscription paywall and premium offer financial models.

By using AT Internet’s digital analytics combined with Poool’s dynamic paywall solution, you can create an online ecosystem that’s as open and efficient as possible and implement the most effective monetisation strategies. 

If you’d like to learn more about orienting your paywall strategy, don’t hesitate to download our new comprehensive e-guide – “Winning the data game – Digital analytics tactics for media groups”.

Article Paywall strategies for the media – understanding your audience behaviour first appeared on Digital Analytics Blog.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Separate tags by commas
To access this feature, please upgrade your account.
Start your free month
Free Preview