Why data matters to a social service
Social services run on information. A feed that suits you, a suggestion for someone to follow, a recommended video and an ad for something relevant all depend on the service knowing something about you and about the content. Personalization is the result: an experience adjusted for each person instead of the same for everyone.
Practices differ by service, region and legal setting, and they change. What follows describes the commonly reported categories, not any specific company’s behavior.
Kinds of data commonly gathered
Information you give directly. Name, email address or phone number, birthday, profile text and photos, interests you select, and any details you fill in. This also includes what you post, comment on, message and upload.
Information about what you do. Which posts you view, how long you stay on them, what you like, share, save, search for or hide, whom you follow, and which links you click. Many services treat this behavioral data as especially useful because it shows what you actually do, not what you say you like.
Information about your device and connection. Device type, operating system, app version, language, time zone, screen settings and an approximate location derived from your network. Precise location may be collected if you allow it.
Information about your contacts and connections. If you choose to upload an address book, or if others tag you, the service may learn about people connected to you, including people who are not users.
Information from outside the service. Some services receive data from other websites and apps that use their tools, for example when a shopping site tells an ad service that someone viewed a product. Businesses can also upload customer lists to match against users. Whether this applies depends on the service and your settings.
Inferred information. From the above, systems may draw conclusions, such as likely interests, age range, or whether you are likely to buy something. Inferences can be wrong, and people often cannot see them.
How data is used to personalize
- Feed and recommendations. Signals about your behavior help predict what you will want to see, as the article on how feeds decide what you see describes.
- Suggestions. “People you may know” and group suggestions often rely on contacts, shared connections and location.
- Search and autocomplete. Past searches and popular queries can shape suggestions.
- Advertising. Data can help choose which ads to show you, and lets advertisers define groups to reach, a subject covered in the explanation of audience targeting. The related practice of retargeting ads, which follows interest from one site to another, is a source of annoyance for many people.
- Safety and security. Data about devices and patterns of use helps detect fake accounts, spam and unusual logins, such as those involved in an account takeover.
- Product improvement. Aggregated usage data guides which features to build or change, and is often used to test variations.
Benefits and concerns
Personalization has real benefits. It can reduce clutter, surface relevant material and make a service feel convenient. Free access to many services is funded in part by advertising that relies on data.
The concerns are also real and worth weighing:
- Privacy. Detailed profiles can reveal sensitive facts, such as health, relationships or beliefs, even if you never stated them.
- Security. Stored data can be exposed in a breach, or accessed by the wrong people.
- Transparency. It is often hard to know exactly what is collected and how it is used.
- Narrowing. Heavy personalization may show you more of what you already engage with, which links to concerns about echo chambers.
- Manipulation. Detailed knowledge of preferences could be used to nudge people in ways they would not choose if they noticed. How serious this is in practice is debated.
- Fairness. Automated inferences can lead to different treatment of different groups, whether intended or not.
Rules and your rights
Many places have laws on personal data that may give people rights to know what is held, to correct or delete it, to object to some uses, and to receive a copy. The details differ widely; the overview of platform regulation touches on these approaches. Whether you have a particular right depends on where you live.
Practical controls
Most services provide some controls. Look for them in settings, and review them from time to time because options and defaults change.
- Review privacy settings. The guide to audience controls explains the basics, including who can see your posts.
- Check ad preferences. Many services let you see or limit categories used for advertising, or opt out of some forms of personalization.
- Manage permissions on your device. Location, contacts, microphone, camera and photos can be set to “never,” “while using,” or “ask each time.”
- Limit contact uploads unless you need them.
- Delete old activity where a service allows it, and request a copy of your data if you are curious what it holds.
- Think about what you post. A reasonable habit is to consider your wider digital footprint before sharing sensitive details, since posted content can be copied or remembered.
- Use strong account protection, such as a password manager and two-step login, to reduce the chance that your data is reached by others.
Be aware that some options reduce personalization without stopping data collection altogether; wording in settings can be vague. Reading the service’s privacy notice, though often long, helps clarify.
For businesses and creators
If you run an account, you may receive aggregated information about your audience. Use it responsibly, avoid collecting more personal information than you need from customers, tell people plainly what you do with their details, and check the rules that apply to you.
Keeping perspective
Data collection is neither purely harmful nor purely harmless. It is the engine that makes tailored services possible and also the source of many privacy worries. A practical stance is to share deliberately, use the controls available, and stay alert to changes in policies, since services do revise them.