- Dec 7, 2018
- 13 min read
How to create personalised content in an ethical way?
In the previous chapters, you learned how and why to monitor and track consumers’ online behaviour data. We were talking to you about the art of listening and about monitoring. But in this last chapter, we wanted to explain to you how you can use the collected data from your listening and monitoring to personalize advertisements. We will get deeper into how much data are collecting so we will explain to you a lot about the Big Data. Of course, we will touch hot topics concerning the privacy of people. Therefore, the aim of this chapter would be to make you know how to create personalised content in an ethical way.

"I think that you get we are about to finish this ebook. The team YouPro made its job as they explained to you a lot as to make you create a successful startup. However, this last chapter is necessary: you know now how and why to collect data through listening and monitoring your network, but you still really don’t know how to use this collected data. You might have already guessed, but I think that the team YouPro has still some insights and interesting things to explain to you. Let’s read this last chapter then!"
How to personalise content thanks to the data collected?
What is Online Behavioural Advertising (OBA)?

As a startup, it is important to reach the right audience. We think you get that now. Your marketing budget might not be that much important so it is of extreme importance to reach the right people. This is why OBA becomes so important. OBA means the collection of data from computers and devices over time and the use of this data to predict web user preferences or interests (Alliance; 2016). This gathering of information has been extremely useful for marketers and companies. The usage and gathering of data have also sparked a huge debate about the privacy of consumers. In this part, we are going to discuss how you can use OBA.
“Online Behavioural Advertising means the collection of data from a particular computer or device regarding web viewing behaviours or mobile app use over time and across multiple web domains and/or mobile apps not under common control for the purpose of using such data to predict web user preferences or interests to deliver online advertising to that particular computer, mobile app or device based on the preferences or interests inferred from such web viewing or mobile app use behaviours.” (Alliance, 2016)
During the day people use a lot of applications and media. Thanks to their mobile phones and devices they are always connected to the internet. By doing so they leave a digital footprint. Thanks to the chapter about monitoring (chapter 5), you are aware of the tools you can use to analyse the data but you don’t know anything yet about the tracking tools. For instance, a cookie is a small file or part of a file stored on a World Wide Web user's computer, created and subsequently read by a website server, and containing personal information (such as a user identification code, customized preferences or a record of pages visited) (Merriam-Webmaster dictionary, unknown). A web beacon or web bug is one of the various techniques used on web pages or email, to unobtrusively (usually invisibly) allow checking that a user has accessed some content. (Wikipedia, unknown). Pixel-based tracking is the process of using a 1x1 pixel transparent gif to track a visit or event on a webpage, to track ad impressions or to track opening for an email. (digitalmarketing-glossary, unknown). Thanks to those tracking tools, you can use the gathered online data to determine people’s interests. But what kind of information does OBA gather? How much of this data is collected?
Luckily organisations are not allowed to gather personal detail such as names and addresses thanks to privacy laws that are already in place. So to gather their data and keep some personal data safe, everyone is given an online ID number. Some examples of data that is linked to this number are the age or gender. However, as people are always connected, everywhere and at any time, they are sharing a lot more information which can be collected. It can be from pages you visit, to post your comment or even when you were born. Because even if you've got an ID number, it is still possible to collect a lot of data about your online behaviour and, therefore, to really know you.
In 2016 Facebook CEO Mark Zuckerberg had to explain how Facebook gathered their information and who it shared this information with. Facebook is a free-to-use application, but it still makes over 80 billion dollars a year (Wagner 2018). Facebook is an interesting example of how advertisers gather your information. During the trial, Zuckerberg explained that outside businesses can only share your information with third-party apps if you grant them permission. Apps like Uber and Spotify require you to use your Facebook account. What you can understand about it is that you need to be aware of what can be collected from you and you don't have to give too much data if you don't want businesses to know everything about you.
What do people think about OBA?

People lack knowledge about what OBA is but start to know that some of their data are collected to personalise ads. But as they are not really aware of what can be collected and how the data collected can be used afterwards, they are becoming afraid. We now talk about the "allergic consumer" who is trying to prevent himself from having is private data collected online. Therefore, people are calling for transparency about the collect, the use and the sharing of these data. That is why nowadays, new laws had been created. For instance, cookies have to be accepted by users. He can manage his cookies preference on each website: he can decide to accept certain types of cookies like the necessary one but not accept the marketing cookies for instance.

However, there is a big paradox because people call for transparency and respect for private life but accept all the cookies without reading what data is collecting from them. As an example: have you already seen a window like it pop up from a website? Because we often did

But did you read it carefully? Did you click on Privacy Policy or User Agreement for details? We think this answer would be in most of the cases “no”.
“It would take a person approximately 201 hours per year to read all the privacy statements for the websites he or she visits” (McDonald and Cranor, 2008).
As a matter of fact, it is a scary thought that we share all this information and we don’t know what it will be used for. There is a growing call for privacy from the consumers. And as a startup, you cannot ignore this call. OBA is completely legal. It is however not 100% ethical. The European association for online advertising has set some guidelines that can be used when using OBA. So if you are considering using OBA make sure to take into account the following:
- Make sure you have a clear agreement with the advertising organisations about any handling of complaints and questions consumers might have.
- Make sure your consumers know that you have used their data and you might sell it through a third party. Transparency is more important than ever.
- Keep measuring the results of your online campaigns. Are you getting the users you wanted? Are they indeed purchasing and using your products and services?
OBA: the future of advertising?
Through OBA, a level of personalization is determined. This level is based on (a) the types of personal data that are used to target the ad (e.g., browsing data or search history) and (b) the amount of information that are used (e.g., just one search term or a combination of browsing data and search history (Boerman, Kruikemeier, Zuiderveen Borgesius; 2017). The following scheme shows the interactions between the advertiser-controlled factors, the consumer-controlled factors and the outcomes.

- The advertiser-controlled factors which include (a) the ad characteristics, or the factors which are part of the ad itself and which can differ among different online behavioural ads, and (b) the forms of transparency which advertisers use to communicate that an ad is a base on online behaviour.
- The consumer-controlled factors which include (a) a cognitive aspect, including people's knowledge and abilities with respect to OBA; (b) an affective aspect, including people's perceptions of OBA in general or of a specific ad; and (c) personal characteristics, such as a person's age or desire for privacy.
- The outcomes which include consumers’ responses to OBA with respect to (a) the actual advertising effects, such as purchases and click-through rates, and (b) the degree of which people accept or avoid OBA.
"Several studies demonstrated that the level of personalization in OBA influences click-through intentions and clickthrough rates. Tucker (2014) found that Facebook ads targeting a person’s interests (e.g., a celebrity of whom a person is a fan) led to higher click-through rates than ads targeting background characteristics (i.e., the college a person is attending)” (Boerman, Kruikemeier, Zuiderveen Borgesius; 2017).
To conclude about OBA, we just wanted to compare it to classical advertising to get the benefit you can get from OBA. In 2009 EASA conducted a research into the effects of OBA in comparison to the classical network-run advertising. This research found three key results:
1. Advertising rates are significantly higher for OBA-targeted ads. In 2009 an OBA campaign was about 2.7 times more expensive than classical network advertising.
2. OBA has proven to have a much higher conversion rate.
3. This gathering of information has another benefit. As discussed, the cost of using OBA is higher than the classical networking advertising. A lot of companies, however, decide to sell this data to third-party networks, creating an additional revenue stream. But what about ethics?
OBA is, therefore, a very useful advertisement solution for you as an upcoming startup if you opt for a moderate level of personalization as it is the most successful level. (Boerman, Kruikemeier, Zuiderveen Borgesius; 2017). However, you really need to respect the privacy of people and it is not something really easy. As to have a better grasp of the ethical and privacy outcomes listening, monitoring and OBA can create, we prepared for you a last best practice.

Best practice
Ethic
This last best practice will be completely different from the others in the ebook but it will really be useful for you: you will get some useful insights and tips about the Big Data, privacy and ethics.
”Everyone wants to go, to make the Big Data a success but don't go because they are afraid" (Desbordes, 2018). This shows there is a real question, debate, around Ethic and Big Data. In today's world, big data is the new oil. Every company gather every single data from their user, ultimately to serve it for formulating strategies and also to understand their customer better. As technology grows exponentially, so does the use of big data and hyper-personalization. However, there are still many discussions going on about the ethical considerations in using it, due to the fact that technology has developed faster than the support around it. How can we, as practitioners, deal with the Big Data without intruding people's life, regarding ethics? To answer this question, we interviewed various persons, we collected insights and we made a video. But before watching the video, we prefer you to read what is following as to understand the main terms and issues discussed in the video.
Big Data, Fast Data, Thick Data: what is the difference?
In the video you will see, you will find some important terms you need to understand. That is why we decided to explain to you a little about Big Data, Fast Data and Thick Data.
Thanks to Big Data, we can now do things we couldn’t before. Big data are actually only large amounts of data. A large amount of data is referred to as big data if it is too large or too complex to be processed manually. This is especially true for data that is constantly changing. Big data, that could be harmless data from climate research. But data about people is also collected: Communication behaviour, consumer behaviour or surfing behaviour of Internet users. You can see the effects of big data analysis every day on the Internet. A typical example is personalized advertising.
Fast Data is the application of Big Data Analytics to smaller amounts of data to solve a problem in real time or near real time or to create business value. The goal of Fast Data is to capture and analyse structured and unstructured data as quickly as possible so that appropriate measures can be taken.
While Big Data collects a large amount of a certain type of information on a certain topic, Thick Data describes the ideas, needs and habits of a few people around that topic. Thick Data is generated from human experience, whereas Big Data relies on machine learning and statistical methods. Thick Data shows the social context between the connections of data points, often creating surprising insights, exciting causalities and new connections. (Wang, 2017)
How should be data to be relevant?
Data are merely facts, figures or statistics that are collected. These collected data is stored and it doesn't mean anything without any further interpretation or analysis. So, in order to get data that is relevant, you need to take time to analyse, to confront, to contrast and so on. It is not about speed, but to take time to look after and have the right filters for what you need.
The most important aspect in the component of data quality is data-accuracy, which means data needs to be accurate to enable a good analysis. Next, to that, it is crucial to have a good interpretation skill after you gathered the right data. People need to be able to process the context of data because data will be useless if you misinterpret it, or even worst, it might lead to a bad decision. (Desbordes, 2018) So definitely, data need to be well analysed in order to be productive. Data analysis is considerably more challenging than simply locating, identifying, understanding, and citing data. There is a multiple-step pipeline required to extract value from data. (Labrinidis, Jagadish, 2008)
Therefore, critical thinking is important in analysing your data. It is essential to develop those critical thinking skills first, before starting to do any analysis of big data. Facts need to be analyzed carefully and from different perspectives, because cause-effect is not the only way to reach conclusions, correlations are not so simple to understand. (Etlinger, 2014)
In essence, interpretation is the key because data can mean anything, so educate yourself around big data first before you can use it (Etlinger, 2014). By doing so, you will be able to produce relevant information out of the data that you have gathered.
What is the problem of the Big Data?
We can say that the main problem of the Big Data is that people wanted to go too fast, as always. And nowadays, people are afraid about it: what about privacy? How far we will go? Furthermore, investing in Big Data is easy but using it is hard: over 73% of Big Data projects aren't even profitable (Wang, 2017). There is a real paradox which is: I really want to use the wealth of the Big Data but I'm too afraid. For instance, this paradox is omnipresent in France and it is more and more real. There was a catalyst in the press a year ago which started to create the fear in people's mind: when two artificial bits of intelligence succeeded for the first time to communicate in their own language (thanks to collecting and combining a lot of data), a language their developer could not understand. Important French politicians are even refusing the Big Data and try to make people sign an ethical chart saying: "Be careful. We are going too far, we will lose control" (Desbordes, 2018). We can even observe this in Europe with the GDPR which is the new European reference text concerning the protection of personal data. Its principal goals are to grow the protection of the personal data and to empower the actors which process this data. As practitioners, we need to respect it and to be transparent with our customers, if we want to process data in an ethical way.
However, people have also the right to decide what they want to share or not; people are 24/7 online. They just need to be aware of what personal data can be used from them and that all the steps they are following on the web are monitored. And actually, Big Data raises a lot of suspicions because people are getting more aware of the fact people are using their data. People are also becoming more aware of how much data is being gathered. Our smartphone even records our heartbeat! It is starting to look a lot like an intrusion. But it is also a matter of whether you have given permission to an application to share your information. (Harbers, 2018).
This points out the other problem of the Big Data: the lack of skills around it. We are just at the beginning of the Big Data era and honestly, we are not very good at handling all the data that we can now collect. (Cukier, 2014) It's like the challenge that was faced by primitive men and the fire: this is a tool that unless we are careful, we will burn ourselves (Cukier, 2014). It is essential to develop critical thinking skills first, before attempting to do any analysis of big data. Moreover, we see that a good understanding of data skills is crucial for a well-formed marketing plan. Knowing the data of your users means you can fine tune how you use your tools. Finally, we need to start learning more about the consequences of the usage of Big Data (Harbers, 2018) (Debordes, 2018).
Big data is going to transform how we live, how we work and how we think. It has already started to do so. We need, as practitioners, to have a good grasp of what it is and to know how to use it well before really using it. We hope you now have some keys to better understand the Big Data and what it implies. We let you watch the video we made for you as a complementary part of this chapter. This video contains insights from 3 different persons, from different countries: Nicolas Desbordes, consultant in the French startup Klaxoon; Tricia Wang, a Chinese technology ethnographer and sociologist of data who helps corporations grow by discovering the unknown about their customers; and Susan Etlinger, an American industry analyst and expert in digital strategy who makes independent research on data, ethics and AI. Those 3 personalities are embodied by members of the team as it was not possible for us to film them. That is why Bas and Chandra both embody Nicolas Desbordes, Michelle embodies Tricia Wang and Morgane embodies Susan Etlinger.
→ VIDEO ETHICS - “YouPro Fast & Curious: the Big Data”.
As a conclusion, the insights are really similar. We are not really surprised about that as we already realised that Startups are more a global spirit, a general state of mind, than something which can really change between the countries. However, we are pretty sure that those insights will really help you. If you want to read the entire interviews we have made, you can click here.

Last but not least, we hoped this chapter helped you to understand how to use the data collected regarding ethics and people's privacy. We know that it is not something easy and as you might have seen in the previous part, the solutions are not all found. What you have to do, as a startuper, is to really focus on the data skills of your team: from the listening and monitoring to the use of the data. It is really something you should do. We also hoped you liked this all ebook. Don’t hesitate to give us feedbacks in contacting us on Facebook, Linkedin, Twitter or our website. It would be a pleasure for us to have comments from you and to continue to get in touch with our community. Have a good continuation and we wish you all the best.
Team YouPro