Are You Letting Social Media Build Your Political Profile—and Sell It?

You don’t have to post “I’m a Republican” or “I’m a Democrat” for a platform to get the idea. Your political profile is built from the boring stuff: what you watch, what you pause on, who you follow, what you share “just to make a point,” and which comments you linger on. That profile isn’t just used to shape your feed. It can be used to shape the ads you get, including political ads. Campaigns and advocacy groups also buy data from brokers to sharpen these profiles, and those categories can get weirdly specific.  This guide breaks down how the profiling happens, what common actions quietly leak signal (including the classic “I Voted” post), and the exact settings you can change to reduce political targeting across Facebook/Instagram, X, and TikTok.

How your “political identity” gets inferred (without you saying a word)

Platforms don’t need you to announce a party. They infer political views from behavior—the tiny, “boring” signals you leave behind while you scroll.

1) Engagement signals: your feed is a lie detector (for attention)

Behavioral advertising runs on one simple idea: attention predicts interest. So the system watches what you do, not what you claim.

Common signals that get folded into a political targeting profile:

• Watch time and re-watches (you didn’t “like” it, but you watched it twice)
• Pauses and linger time on a clip or comment thread
• Likes, shares, saves, replies
• What you hide/report (even “I don’t want to see this” is a preference signal)
• Link clicks to news, candidate pages, issue explainers, donation pages

This is why election season can feel like your feed “knows” what topics get under your skin. It’s reacting to what keeps you engaged, because that’s what ad systems optimize for. EFF points out that political campaigns use many of the same invasive tactics as behavioral ads—pulling in data from multiple sources to build a profile they can target .

2) Network signals: who you follow can label you

Even if you never react, your graph talks:

• Who you follow (and how quickly after an event you follow them)
• What groups/communities you join
• Who you DM or interact with often
• Which creators you consistently return to

A single follow won’t “prove” anything. Patterns do. And patterns are what machine learning models love.

3) Off-platform signals: the “stitching” problem (and why it’s so powerful)

Here’s the part most people miss: platforms and advertisers don’t rely only on in-app behavior. Political ad targeting can be strengthened by combining consumer data, smartphone/location data, and public voter info—often through data brokers .

EFF lays out a simplified pipeline that shows how it works in practice:

1. A campaign starts with a voter list (names, addresses, party affiliation)
2. A data broker enhances it with consumer and inferred attributes
3. An ad targeting company delivers ads to your device—sometimes using signals like IP address or phone location

That’s how “political identity” becomes something that can be bought, scored, and targeted—often without you seeing the joins.

4) “Why am I seeing this ad?” is your proof it’s happening

On Facebook, you can click the three dots on an ad and choose “Why am I seeing this ad?” to see some of the targeting logic behind it . It’s not a full confession, but it’s usually enough to spot the inputs: interests, demographics, location, or “similar to” audiences.

Quick reality check: you can’t fully stop political ad targeting

You can shrink the signal, though. EFF’s guidance is blunt: you can’t completely opt out of behavioral ads on Facebook, and TikTok doesn’t let you fully disable personalized ads . The practical win is reducing how confidently you can be labeled, and cutting off the easiest data sources feeding the machine.

The hidden data risks in normal election-season behavior (yes, even the “I Voted” selfie)

Once you accept that platforms infer political views from behavior, the next uncomfortable truth is this: election-season “normal” behavior creates durable signals. Stuff you do casually in October can shape political ad targeting for months.

Harmless actions that scream “target me”

These aren’t shady actions. They’re everyday internet habits that get translated into political interests:

• Reposting hot takes “for discussion”

Even a quote-tweet dunk can tag you as someone who engages with that issue or candidate.

• Reacting to charged content (likes, angry reacts, replies)

Ad systems don’t need your agreement. They need your attention.

• Joining local groups (neighborhood, school board, city politics)

Local groups are gold for political targeting because they imply geography + issue priorities.

• Clicking “volunteer,” “donate,” or “get updates” and using the same email/phone you use everywhere

That contact info acts like a connector between your social profiles, campaign signups, and broker databases.

The “join key” that makes it stick: your email + phone

Your email address and phone number are the closest thing to a universal ID in political marketing. When you reuse them across petitions, campaign sites, and social apps, it becomes easier to connect your activity across sites and devices.

It also doesn’t always stop with one campaign. EFF notes that campaigns may share information with other campaigns, and points to campaign privacy policies that explicitly describe sharing with aligned organizations or groups . They even cite a real example where an email list was treated like a valuable asset—Hillary Clinton’s campaign gave its email list to the DNC as a contribution valued at $3.5 million .

Yes, even petitions and “I Voted” energy can have a paper trail

If you’ve ever signed a ballot-measure petition, that “small civic moment” can become public record. EFF points out that in many states, petition signature pages remain part of the public record, and the information provided may be used for mailings or targeted political ads .

That’s the vibe with “I Voted” posting, too: you’re not revealing your vote, but you’re confirming engagement, timing, and sometimes location context.

The bigger pipeline (and why the ads suddenly get creepy-specific)

Most people only feel this system after it hits them: one week you casually engage with a few election posts, the next week your feeds are packed with hyper-specific political ads that seem to “guess” what you care about.

Under the hood, campaigns can start with voter files—which may include public data like voter registration, party registration, address, and participation (whether you voted, not who you voted for) —then enhance it with broker data and push ads through ad delivery systems .

And those broker categories can get weirdly specific. EFF describes political data brokers that focus on campaign-useful data—like i360, TargetSmart, and Grassroots Analytics—and shows examples of inference categories from a broker file such as “Qanon,” “Rightwing Militias,” “Right to Repair,” “Inflation Fault,” “Climate Change,” and even workplace-related labels . They also note model-style attributes like “Voter Fraud Belief” and “Ukraine Continue” .

That’s why “I’m just scrolling” isn’t neutral during election season. You’re feeding a profiling system that’s built to label you, segment you, and sell access to you.

Do this now: reduce political targeting on Facebook + Instagram (Meta)

If you’re getting that “how did they know?” feeling on Meta, it’s usually a mix of ad preferences plus off-platform tracking. The second one is the quiet culprit.

1) Use “Why am I seeing this ad?” like a receipt

This takes 10 seconds and tells you what Meta thinks it knows.

1. On a Facebook/Instagram ad, tap the three-dot menu
2. Tap “Why am I seeing this ad?”
3. Screenshot or note the reason (interest, age range, location, “advertiser’s audience,” etc.)

Do this on 3–5 political ads. You’ll usually see the same themes repeating, which tells you what to change next.

2) Audit Meta’s Ad Preferences (Facebook + Instagram)

EFF’s advice here is direct: even though you can’t fully opt out of behavioral ads on Facebook, you can review ad preferences and opt out of what’s available .

What to look for in Ad Preferences (names can vary slightly by app/version):

• Interests / topics that look political (remove what you can)
• Advertisers you’ve interacted with (hide/remove)
• Ad settings that allow personalization based on your activity

Treat this like cleaning your closet. You’re not deleting the closet, you’re removing the loud labels.

3) The missing piece: disable access to off-site activity (Off-Facebook Activity)

Meta can learn about you from activity happening outside Meta (think: other websites and apps). EFF explicitly recommends disabling access to off-site activity .

Checklist:

1. Go to Facebook settings and find Off-Facebook Activity (sometimes shown as “Your activity off Meta technologies”)
2. Review what’s being sent back to Meta
3. Disable access / disconnect future activity

This one change often reduces how fast political targeting “snaps back” after you clean up interests.

4) One habit shift that matters: stop giving Meta easy join-keys

Meta’s strongest “glue” is anything that stays stable over time—your email and phone number.

• Don’t reuse your main email/number for campaign signups, petitions, and random advocacy forms.
• If you want clean separation without juggling extra inboxes, tools like Cloaked can help by generating masked emails and virtual phone numbers, so a one-time political signup doesn’t become your forever identifier across platforms.

That’s the goal here: not invisibility. Just making political ad targeting less confident, less sticky, and less profitable.

Do this now: reduce political targeting on X and TikTok (and what you can’t control)

Meta gives you a lot of knobs. X and TikTok are more of a mixed bag. You can still cut down the signal—just don’t expect a clean “off switch.”

X (Twitter): check these 3 screens

On X, your goal is simple: reduce ad personalization and clean up any inferred interests tied to identity signals.

1. Ads personalization

Look for settings that control:

• Personalized ads (turn off if available)
• Ads based on your activity (turn off)
• Anything that sounds like “inferred” or “deduced” interests (review and remove)

2. Data sharing / identity-based matching

If X offers toggles for matching you via:

• Email
• Phone number
• Off-platform activity

turn those down. Those are the same “join keys” that make targeting sticky across sites.

3. Your privacy and safety basics

This isn’t just about ads. Fewer data connections means fewer ways to classify you.

• Tighten who can find you by email/phone
• Limit location exposure if you don’t need it for your use case

Keep it boring. Boring wins.

TikTok: set expectations (and still make it harder)

TikTok is blunt here: you can’t fully disable behavioral ads . So your best move is reducing how much data TikTok (and ad tech around your phone) can use to guess who you are.

In-app steps worth doing

• Run a quick privacy checkup (TikTok changes labels often, but the options usually live under Settings and privacy)
• Reduce personalization where it’s offered (topics/interests/activity-based toggles)
• Limit permissions you don’t need (location is the big one)

Device checklist (this helps across all apps)

EFF’s guidance is practical: kill the phone ad identifier and tighten location permissions .

iPhone (iOS)

1. Settings → Privacy & Security → Tracking
2. Turn off “Allow Apps to Request to Track”
3. Settings → Privacy & Security → Location Services

• Remove location access for apps that don’t need it
• Prefer “While Using”
• Turn off Precise Location when it’s not necessary

Android

1. Settings → Security & Privacy → Privacy → Ads
2. Tap “Delete advertising ID”
3. Settings → Location → App location permissions

• Set to “Allow only while using the app” where possible
• Disable precise location where it’s not needed

What you can’t control (but should plan for)

• TikTok-style personalization won’t disappear, even after you flip settings
• If an advertiser can reach “people like you,” you can still land in that bucket

The win is making those buckets fuzzier. Less certainty. Less clean labeling. Less political targeting that feels like it’s reading your mind.

A practical ‘election-season privacy’ routine you’ll actually keep

If you try to “fix privacy” once and forget it, the system wins. Election-season targeting changes fast, and the settings drift. The routine below is built for real life: 15 minutes a month, plus a couple of small habits that cut down political ad targeting.

The goal (keep it simple)

You’re not trying to disappear. You’re trying to reduce how confidently you can be labeled and sold.

That matters because political groups spend serious money buying targeting inputs. OpenSecrets found that in 2020, political groups paid 37 data brokers at least $23 million for access to services or data, and those brokers pull from cookies, web beacons, mobile phones, social platforms, and more .

The monthly 15-minute checklist (set a recurring reminder)

1. Quick ad-settings audit (10 minutes)

• Open each platform’s ad settings and remove interests that look political or overly specific.
• For any political ad you see, hit “Why am I seeing this?” before you react. That one click keeps you from feeding the machine emotionally, and it tells you what’s being used to target you .

2. Kill unnecessary location access (3 minutes)

Location is still one of the most valuable signals to advertisers and campaigns .

• iPhone: Settings → Privacy & Security → Location Services

Remove access for apps that don’t need it, use “While Using,” and disable Precise Location when it’s not required .

• Android: Settings → Location → App location permissions

Set “Allow only while using the app,” and turn off precise location where possible .

3. Reset your ad identifier (2 minutes)

EFF points out your phone has an ad identifier that makes it easier for advertisers to track and collate what you do, and recommends disabling it .

• iPhone: Settings → Privacy & Security → Tracking → turn off “Allow Apps to Request to Track”
• Android: Settings → Security & Privacy → Privacy → Ads → Delete advertising ID

The two habits that do the heavy lifting

1. Stop reusing your real email/phone for petitions and campaign signups.

Those identifiers connect everything.

2. Use a “politics-only” contact layer.

EFF recommends creating an alternate email and phone number for campaign stuff . If you want that separation without managing extra accounts, Cloaked is a practical option: masked emails and virtual phone numbers let you sign up without handing over your long-term identifier.

This routine won’t eliminate political ads. It just makes the profiling sloppier and the targeting less precise, which is the point.

‍