McDonald's has just bought an unknown Israeli technology company. It's called Dynamic Yield and, according to its website, they do this: "Power a deep personalization with an agile segmentation and a data management system". It's incomprehensible. This is explained below: "Unify customer data between all offline and online contacts and use them to enhance your marketing ecosystem."
They are all euphemisms, language in code. In the acquisition's press release, McDonald's talks about improving "personalization" and "logical decision technology" based on previous purchases, time of day, weather and other minor details. But that does not cost 300 million dollars – which is the figure given by several American media – and Dinamic Yield, in fact, knows how to do more things. For example: "Unchain behavior layers". The translation would be something like this: "Send messages to your mobile phone at a very specific moment to be encouraged to do something you did not plan to do".
McDonald's is a good customer: it sells a tempting product with a bad reputation
McDonald's is a good customer: it sells a tempting product but with bad reputation. It is easy to imagine a consumer wanting and at the same time not wanting a hamburger. What if McDonald's could know when that happens to us and, perhaps, encourage us to go, come, for a day nothing happens? Dynamic Yield can help you.
With all its euphemisms, encourage consumption through mobile in real time by legitimate means -the user gives without being aware of his approval when downloading an app, buying a jersey or visiting a web- is a usual commercial route today. One morning someone leaves your house and surfs the internet while having breakfast. Your pattern of visited websites varies from other days. Look, for example, more webs of the heart, a clue that is associated with sadness. Sadness is a factor that increases the desire to buy. At that precise moment, an email, pop-up or sms with a shelter or travel offer is more likely to be effective.
The cheap price of your data
In 2018 there were 33 zetabytes of data in the world. In 2025 there will be 175, according to a study by the consultancy International Data Corporation (IDC). One zetabyte is 10twenty-one bytes. It is the unit that goes after gigabyte, terabyte, petabyte and exabyte. All the words that have been said by humanity occupy 5 exabytes, not even 1 zetabyte, according to TechTarget. In 2010, there were barely a couple of zetabytes in total.
The growth of data in the cloud of companies, humans, cars and institutions in the cloud will be sidereal. We are alone in the first chapter of the data age. Maybe then the value of our data grows, because today our data is cheap: 1.42 euros per month in Europe, 1.05 in the world, according to the calculations of the publication Strategy + Business. It's not much, but at the moment they generate mostly ads. The dominators are Google and Facebook. The work that companies like Dynamic Yield will do is growing and will turn all companies into technology.
Companies are essentially of two types: some track data and others order them
The data industry is very unknown. The companies are essentially of two types: some track data and others order them. Sometimes there are companies that do both. They are immensely useful: it is not the same to try to convince 1 million people with an ad on TV that 50,000 are interested in your product and right now it is more likely that they will buy it.
Between 2013 and 2018, Facebook had an agreement with at least four major data commerce companies, Acxiom, Epsilon, Experian and Oracle Data Cloud: "This agreement with Facebook has allowed us for the first time in history to see what data data agencies have, but not knowing where they come from," says Oana Goga, a researcher at the National Scientific Research in Grenoble (France).
Those companies had hundreds of information variables on more than 90% of Facebook users in the United States. The researchers They looked at six other countries and the volume was smaller. This information was made public by Facebook for 5 years in its audience tool. Data companies continue to do their work without Facebook.
A proof of the desire that this sector has to remain in the shade is how they hide their activity. The domain rlcdn.com Talk to hundreds of apps, including many media, great services and even porn pages. What does a profiling company want to know that someone sees porn? Or, better yet, how much porn do you see? The domain registration is done in Panama and several proxies and screen names aspire to hide who is at the end of the chain. Attribution is one of the challenges of academics studying the sector.
The data is transmitted linked to a personal identifier: the IMEI of the mobile, the IMSI of the SIM, the email, the MAC address, the telephone number. They are the glue of the rest of the data. From there hang data that alone seem anodyne details but added to life: what apps you have on your mobile, how you type, what you write and delete before posting, the photos you upload, the heart rate, the accelerometer, the location, the times you look at the mobile, at what speed you give "I like you", gestures of your face. The variables are hundreds.
The data is transmitted linked to a personal identifier
There are companies that track this flow of information, closely linked to the mobile. Then there are others -or the same ones- that take that information and study how to measure "sadness" or "enthusiasm". If it is in real time, better. The novelty is not so much the storage capacity but the computing: what do we see if these three variables that give this type of profile we add this one? The results do not need to be scientifically perfect. It is an approximation, not a diagnosis. If someone seems to be sad or asleep the same way he wants a hamburger. And if you fail, it's not serious or expensive.
The American state of Vermont launched a law in February to keep data companies in a public registry. They signed up 121. These companies do not exist for a few years. Acxiom, for example, has been in the sector for decades: "Access to more than 45 years of data and specialization in identity resolution in a completely self-managed way," they say on their website with their euphemistic language. Before the internet, the data came from subscriptions to magazines and super loyalty cards. Now they have a lot more depth.