We are entering into an era where data is King, and where our every move, our every emotion and our every contact can be tracked. With the increasing analysis of social media, there is often very little that can be done about our lives that can be hidden from organisations wishing to push customized content to us, to try and understand how we live our lives. If an on-line company can drop a cookie onto our machine, they can sustain a long-term tracking of our activities, and this now includes understanding how we react to advertising material, especially on what material has made us click on the content, and increasingly they are learning our behavior.
The need to gain ethical permission, in the same way that research teams require when they involve human participants, is slowly eroding, and perhaps, in some cases, a natural extension of existing practices, where advertising content is focused on target groups.
One of the major changes within this Big Data era, is that users often freely offer their data to the Internet, and where it can be used in ways that are often unexpected to the user. For example, a tweet on a local event will time stamp where a person was at a given time, and even reveal information around their movements, and perhaps who they had contact with.
Mining for sense and emotion
With so much on-line data, it is key for advertising agencies to understand the emotions of messages posted on-line, and where studies that would take months or even years, can now be done within minutes. Like it or not, we are all part of an on-going experiment which is mining our data on a continual basis, and then pushing content our way, and monitoring even how we use the customized content.
SInce advertising began as an industry, researchers have thus been trying to mine large populations for the emotions, and the challenge within social media is to make sense of large amount of comments and to mine for the sentiments shown with them. This is fairly easy with a tweet of:
I am so happy that the sun is shining today :-)
but by placing a different emoticon on it, changes the sense:
I am so happy that the sun is shining today ;-)
and then is changed completely with the dreaded exclamation mark:
I am so happy that the sun is shining today!
which gives the impression that someone is very unhappy about the weather.
Part of a great experiment
Facebook took the emotion research one step forward in January 2012, when their data scientist Adam Kramer conducted a two week experiment on 689,003 Facebook users, in order to find out if emotions where contagious within social networks. It basically came of the almost obvious conclusion that users feel generally happy when they are fed good news – the economy is looking good and the weather is nice – and depressed when they get bad news – a bomb has gone off injuring many people and it looks like snow is on the way.
For users the mining of the data is generally fine, and Facebook, and many other Internet-focused companies, especially Google and Amazon, extensively mine our data and try to make sense of it. For this symbiotic relationship, the Internet companies give us something that we want, such as free email, or the opportunity to distribute our messages. What was strange about this one is that the Facebook users were been treated in the same way as rats in a laboratory, and had no idea that they were involved in the experiment. On the other hand it is not that much different from the way that affiliate networks have been created, and which analyse the user, and try to push content from an affiliate of the network, and then monitor the response from the user (Figure 1).
We increasingly see adverting in our Web page accesses, where the user is matched to their profile though a cookie, and where digital marketing agencies and affiliate marketing companies try and match an advertising to our profile. They then monitor the success of the advertising using analytics such as:
- Dwell time. This type of metric is used to find out how keen the user has been before clicking the content.
- Click-through. This records the click-though rate on content. An affiliate publisher will often be paid for click-throughs on advertising material. This can lead to click-through scams, where users a paid to click on advertising content on a page.
- Purchases. This records the complete process of clicking-though, and the user actually purchasing something. This is the best level of success, and can lead to higher levels of income, and in some cases to share a percentage of the purchase price. Again this type of metric can lead to fraud activity, where a fraudster will use stolen credit card details to purchase a high price item through a fraudulent Web site, and use this to gain commission from an on-line purchase (which is then traced to be fraudulent at a future time).
The key areas which are relevant to monitoring of user activities are:
- Transaction verification. This involves protecting users by understanding their activities, especially around the types of purchases they make.
- Brand monitoring. This involves understanding how brands are used within web pages, and how they are integrated and if key messages are picked-up.
- Web-traffic analytics. This involves understanding how users search of pages and navigate around web sites.
- Affiliate platforms. This tries to match users to affiliates, and integrate targeted marketing.
- Campaign verification. This uses analytics to verify that campaigns are successful in their scope.
One of the most successful uses of targeting the user and monitoring their actions is in affiliate marketing, where businesses reward affiliates for each ‘customer’ brought about by the affiliate’s own marketing efforts. This is a booming market:
- Of projected global online sales of nearly $780 billion by 2014, ~ $90bn will be driven by affiliate marketing
- $4.62bn sales driven by affiliate marketing in the UK in 2010.
Figure 2 shows an example of targeted advertising, where a previous page involved a search for a Microsoft Surface Pro, and Ad-Choice (which is maintained by Criteo) has integrated an advert for it within another page. In this way Ad-Choice has decided that this is a good advertisement for us, and if we click though, the click will be remembered, and the host site will get some form of payment for the click. If the user actually follows-through and purchases the goods, the host site could gain a part of the commision.
Thus we are being monitoring and mined all the time, and the content which is pushed to us is focused on us. Increasely the content for us is being customized with advertising messages. There is generally no need for informed consent, at the present, for this type of push advertising, as users generally feel that there an acceptable level in intervention for their Web content, but perhaps forget that there is a whole of lots matching and analyzing going on in the background.