The democratisation of mobile phones has facilitated the appearance of a large number of applications in such a way that people today can personalise their cell phones to match their tastes and interests. In 2021, for example, there were 143.6 billion app downloads. The interesting thing about this growth is the amount of information that can be collected from the apps so that companies know their prospects better; these insights give rise to what is known today as fragments.

‍

Fragments are insights obtained from metadata coming from mobile phones. The difference between fragments and metadata is that metadata is the raw data, while fragments are the concepts and ideas obtained after that data has been collected and analysed.  

‍

Fragments allow for a more accurate understanding of customers and can traverse an entire organisation providing useful information to different sectors. Thus, it can help the risk team approve a credit or insurance, and the marketing team make better promotions, more tailored offers, and product recommendations. For example, an insurance company can see what kind of apps a person uses and determine whether or not they are interested in topics related to sports, health, safety, and make an offer in pursuit of satisfying these needs.

‍

Some of the fragments that credolab uses are:

• Installed applications: Allow for a better prediction of a person’s personality. For example, knowing if prospects are users of digital banking or traditional banking if they make investments and where, if they use competing products and what their source of income is (self-employed, in a dependency relationship, among others).
• Device information such as brand, operating system, model: These allow to perform technographic segmentations, predict tech savviness, and improve micro-segmentations
• Detection of calendar events: The habit of scheduling upcoming activities may be an indicator of how organised a person is, how well she plans 
• Analysis of contact behaviour: Having more than just names and phone numbers on the contact list may indicate a higher degree of perfectionism.
• Detection of velocity: Analysing the uniqueness of each device and the speed at which some customers repeat loan applications provides additional insights to help prevent fraud
• Analysis of typing behaviour and UI interactions: Detecting the number of times the customer completes or changes a certain field, uses an autocomplete functionality, uses copy/paste/delete in filling out an application form can be used to improve the user experience and eliminate friction

‍

Data collection based on artificial intelligence

‍

Fragments are a product of the appearance of smartphones, the increased internet penetration, the rise of app adoption, and the development of new technologies based on artificial intelligence and machine learning. These technologies are helping overcome the lack of traditional information systems behind credit bureau scoring models. That is, previously information was collected from past actions: credit or debit information in the case of the financial system, or samples obtained at a certain point in time to predict segments through statistics in the case of market research. 

‍

In contrast to traditional information models, modern algorithms allow us to overcome the past and move from a static data collection system to a dynamic one. Thus, we can now find fresh, real-time data and update it to create behavioural models and obtain fragments from the present. In this way, companies can instantly learn about users and identify changes in their behaviour to make decisions faster and more effectively, in the moment.

‍

Technology has not only made possible to improve the user experience by expanding the variety of devices and software, it has also managed to overcome the limitations of old data collection systems, providing companies with better insights and fragments of information for a greater and more precise understanding of each customer.