Your public social media posts, your company's own website pages, and stolen credential dumps are the three richest sources criminals mine first—and they cost the attacker almost nothing. The pattern holds across every industry: the highest-yield intelligence sits in places you published deliberately or leaked by accident, not in exotic hacking tools. Understanding those channels changes how you defend, because most of them are places your security budget never touches.

The ranking below comes from patterns documented by the FBI's Internet Crime Complaint Center, the UK's National Cyber Security Centre, and breach analyses from Verizon and IBM. It answers one question: where does the attacker actually spend the first hour?

Public Social Media Profiles Supply Organizational Charts and Daily Routines

Attackers treat LinkedIn as a free intelligence database. Job titles reveal who approves payments, who administers systems, and who reports to whom. A public profile naming someone as "Accounts Payable Manager" tells a fraudster exactly whom to impersonate in a business email compromise attempt, and the platform's "people also viewed" feature maps the finance team for free.

The FBI's Internet Crime Complaint Center has logged business email compromise as a multibillion-dollar category annually in recent years, and profile harvesting feeds it directly. Check your own exposure today: search your company name plus "finance" or "accounts payable" on LinkedIn and count how many full names and titles appear without logging in.

Out-of-office replies do more damage than most people realize. A message saying "I'm traveling until the 14th, contact Priya for urgent invoices" hands over a colleague's name, a date window, and an organizational relationship in one sentence. The NCSC recommends never naming a deputy or covering colleague in an auto-reply because it creates a ready-made impersonation target.

Instagram and Facebook posts expose physical security detail too. Photos with office badges visible, desk screens showing internal tools, or geotagged conference attendance let someone map when key staff are away. A single photo of a whiteboard can leak project names and system architecture. Post carefully, and strip metadata where platforms allow it.

Corporate Websites Reveal Software Stacks Through Job Ads and Error Pages

Your careers page is an inventory of your technology. A job listing asking for "experience with Fortinet FortiGate firewalls and VMware vCenter" tells an attacker which exploits to try first. The NCSC explicitly warns that recruitment ads disclosing specific product versions shorten an intruder's reconnaissance phase, and many organizations still publish these details without review.

Error messages and page source code leak more. A stack trace on a misconfigured page can display server versions, framework names, and file paths. Web crawlers built for reconnaissance—Shodan, Censys, and commercial equivalents—index these automatically, so a page that errors once can stay searchable indefinitely. Ask your developers whether error output is suppressed in production.

Press releases and investor documents carry strategic timing. An announcement of a merger, a new ERP rollout, or a quarterly filing date gives a phishing campaign perfect narrative cover: "Urgent: action required before the systems migration." The attacker sounds credible because they read your newsroom. Verify who reviews outgoing announcements for operational detail.

Domain registration records and email headers add infrastructure facts. WHOIS data (increasingly redacted, but still partly visible), DNS records, and mail server headers reveal hosting providers and gateway vendors. A "Received:" line showing Barracuda or Proofpoint identifies your email filter by name, which tells an attacker which bypass techniques to test. Check what your headers disclose by sending yourself a test message and reading the raw source.

Data Breach Compilations Put Your Credentials on Sale for Cents

Credential dumps are the cheapest intelligence in existence. Collections like "RockYou2021"—a compilation of roughly 8.4 billion password entries reported by Cyber News researchers in 2021—circulate freely, and criminal markets price individual account records at fractions of a dollar. An attacker doesn't need to hack you if your employee's old password from a forgotten forum breach still works on the corporate VPN.

The economics explain the persistence of this channel. Verizon's Data Breach Investigations Report has repeatedly found that the majority of breaches involve a human element, with credential misuse a leading vector, and reuse is the fuel. People recycle passwords across personal and work accounts, so a leak from a fitness app becomes a key to the office.

Attackers automate the check. Credential-stuffing tools replay leaked username-password pairs against login pages at scale, and the FBI has warned about this technique against financial and retail accounts for years. Defenders should run their own exposure checks: query the Have I Been Pwned service for every corporate email domain and act on the results.

The failure case matters here. Organizations that force a password reset after a third-party breach often miss accounts created through single sign-on aliases or personal email addresses registered on work systems. Audit for those exceptions before declaring the incident closed, because attackers know contractors and alumni accounts survive resets.

Email Threads, Signatures, and Auto-Replies Hand Over Structure and Tone

A single forwarded email thread teaches an attacker your communication style, your internal jargon, your approval chain, and your calendar. Business email compromise succeeds when the fake message reads like the real ones, and criminals build that mimicry from stolen mailboxes or from messages employees pasted into public forums asking for help.

Email signatures are a quiet leak. They often list direct phone numbers, office locations, department names, and sometimes system references like "Submit tickets via ServiceNow." The NCSC advises keeping signatures minimal and avoiding organizational detail that helps an outsider navigate internally. Review what yours says this week.

Auto-replies and read receipts broadcast availability. A reply-all storm or a "delivery failed" bounce to a public list reveals mailbox structures and naming conventions. Those patterns let an attacker guess addresses confidently—firstname.lastname@company is a gift compared with random strings.

Phishing tests demonstrate the payoff. When security teams run simulated campaigns, the convincing messages borrow exactly these details: a real project name from a press release, a real manager's name from LinkedIn, a real travel date from an out-of-office. Training fails when it teaches people to spot typos rather than to verify unexpected requests through a second channel.

Third-Party Vendors and Supply-Chain Partners Leak Your Playbook

Your suppliers know your systems, your timelines, and your contacts—and their security is not yours. The 2013 Target breach began with credentials stolen from a heating and refrigeration contractor, Fazio Mechanical, according to the Krebs on Security investigation and subsequent reporting; the attackers then moved into Target's payment environment. The lesson stands: a small vendor with weak authentication becomes your reconnaissance asset.

Vendor questionnaires reveal how much partners store. If a facilities contractor holds building access schedules and network diagrams for badge systems, that document is intelligence. Request data-flow diagrams from critical vendors and check whether they store your information beyond the contract period.

Public case studies and customer lists work against you. When a vendor publishes "How ACME Corp migrated 40,000 endpoints with our tool," they've named your infrastructure and your scale. Ask contractually for approval before your name appears in marketing material, and read the vendor's own security page for the versions they run.

Concentration risk amplifies all of this. A managed service provider serving fifty small firms becomes a single reconnaissance point for all fifty, which is exactly why the 2021 Kaseya and 2020 SolarWinds incidents had cascading effects documented extensively by CISA. Map which vendors can see your network, then limit what each can reach.

Dark Web Markets and Forums Package Stolen Data for Resale

Criminal forums operate as a supply chain with reputations, escrow, and reviews. Initial access brokers sell footholds into corporate networks, and their listings typically include company size, industry, security stack notes, and asking prices that have ranged from a few hundred to tens of thousands of dollars depending on access level, according to reporting from threat intelligence firms like Intel 471 and Recorded Future.

Ransomware affiliates buy from these brokers. The chain—broker sells access, affiliate deploys ransomware, negotiator handles payment—means your leaked VPN credential changes hands multiple times before anyone encrypts a server. Each handoff adds reconnaissance: someone checks your backups, someone else checks your insurance.

Monitoring services exist for defenders. Dark web monitoring products from vendors including Have I Been Pwned's commercial tier, SpyCloud, and Kela alert you when corporate credentials appear in marketplaces. They are not perfect; closed forums and private Telegram channels stay invisible, as Europol reporting on encrypted criminal communications has noted repeatedly.

Treat any alert as a breach signal, not a nuisance. A listed credential means someone has already validated it works, usually. Rotate it, review that account's access, and check for lateral activity before assuming the incident is contained.

Search Engines and Public Records Aggregate the Fragments

Google assembles the puzzle. Query operators like "site:company.com filetype:pdf" surface internal documents accidentally published—meeting minutes, org charts, onboarding guides. Attackers run these searches before every engagement, and so should you, because a leaked PDF naming your help-desk phone number and password-reset process is a phishing script waiting to be read.

Business registries expose leadership and ownership. Companies House in the UK, state secretary-of-state databases in the US, and equivalent registers publish directors, registered addresses, and filing histories. These are legitimate public records, but they give an attacker verified executive names and often personal addresses for spear-phishing and fraud.

Code repositories are a special case. GitHub searches for API keys, database connection strings, and internal hostnames regularly turn up live secrets, and GitGuardian's annual State of Secrets Sprawl reports have counted millions of leaked secrets across public repositories. A developer pushing a config file "just for testing" creates a permanent artifact unless it's caught.

Shodan and similar scanners index internet-facing devices continuously. Run a free search for your IP ranges to see which ports and banners are visible—an exposed management interface with a version banner is both a finding and an attacker's starting point. Fix before someone else finds it.

Attacker Effort Versus Payoff: A Comparison of Collection Channels

Not all channels are equal. The table below ranks them by what an attacker must invest and what they get, drawing on the reconnaissance phases described in MITRE ATT&CK's framework, which catalogs techniques like "Search Open Websites/Domains" and "Gather Victim Identity Information" under its Reconnaissance tactic.

Collection channel Attacker effort Typical payoff Typical defense Free or low-cost for attacker?
Public social media Minutes, no tooling Org chart, roles, travel windows Posting policy, minimal profiles Yes
Corporate website and job ads Minutes, search operators Software versions, vendors Ad review, error suppression Yes
Breach compilations Near zero, bulk purchase Working credentials Password reset, MFA, breach monitoring Yes, cents per record
Email signatures and replies Passive, one message Tone, contacts, structure Minimal signatures, no named deputies Yes
Vendor and partner documents Weeks of relationship building Network paths, schedules Vendor risk review, contract clauses Moderate
Dark web markets Money, forum access Validated footholds Credential rotation, monitoring No, hundreds to thousands of dollars
Search engines and public records Minutes, free operators Documents, directors, exposed devices Index cleanup, exposure scanning Yes

The verdict splits cleanly. If you have no posting policy, no exposure scanning, and no breach monitoring, you're losing on the three cheapest channels simultaneously, and fixing those costs process time rather than licensing fees. If you've done the basics, shift attention to vendors and dark web monitoring, where the attacker must invest more and your detection window is longer.

One real drawback of relying on public-data removal services: they clean search results but not cached copies, archive snapshots, or forum reposts. Removing a document from your site doesn't undo a year of indexing, so prevention at publication beats takedown requests.

What This Means for Defense: Shift Controls Toward the Cheap Channels First

Prioritize by channel cost, not by technical sophistication. Most breach post-mortems describe entry through stolen credentials, phishing, or exposed services—not zero-days—so the controls that matter are multi-factor authentication on every external login, a tested password policy, and patching internet-facing systems. CISA's guidance emphasizes exactly these fundamentals because they close the channels attackers actually use.

Run an exposure audit quarterly. Search your domain in breach databases, query Google for leaked files, review your public job ads for version detail, and check Shodan for your ranges. Document what you find and fix the top ten items. The exercise takes a few hours and mirrors the attacker's first day.

Train people on verification, not vigilance. Teach staff to confirm money transfers and credential requests through a second, known channel—a phone call to a number they look up themselves. Simulated phishing that rewards reporting over clicking measurably improves detection in most organizational data.

Write down what's acceptable to publish. A one-page policy covering social profiles, job ads, press releases, and vendor case-study approval prevents most leaks before they happen. Review it annually, because platforms and job markets shift. The goal isn't secrecy; it's removing the free map.

FAQ

Do most cyberattacks really start with information people published themselves?

Most reported breaches trace back to stolen or weak credentials, phishing, or exposed services rather than sophisticated exploitation, according to Verizon's annual breach analysis. That means information channels like breach compilations and social profiles carry more weight in real incidents than headlines suggest. The self-published details matter because they make the phishing convincing. Assume the attacker has read your website and profiles before the first message arrives.

Is LinkedIn outreach from strangers a real risk, or just sales noise?

Both happen, and the pattern is easy to distinguish. Genuine recruiters discuss roles and never ask for credentials, payment, or software installation. A message asking you to open an attached salary document or log into a portal via an unfamiliar link is an attack, regardless of the sender's profile polish, because profile photos and employment histories can be fabricated. Verify through the company's official careers page before engaging. Report suspicious accounts to the platform.

What's the actual cost of a credential bought on a criminal market?

Pricing varies widely by access type, and current listings should be checked through a threat intelligence provider rather than assumed. Broadly reported ranges run from under a dollar for bulk consumer account records to hundreds or thousands of dollars for validated corporate network footholds sold by initial access brokers. A single working VPN credential to a mid-sized firm has been reported in the low thousands of dollars range, which is cheap relative to ransomware payouts.

Can deleting my old posts and documents make me safe?

No, and that's the misconception worth correcting. Search engines, web archives, and third-party reposts retain copies, so deletion reduces exposure but doesn't reverse it. If a leaked API key or credential was ever public, treat it as compromised and rotate it. Prevention at publication is cheaper than cleanup. Build a review step before anything goes live rather than relying on takedowns afterward.

What should a small business with no security team do first?

Start with multi-factor authentication on email and remote access, then check the company domain against breach databases, then review job ads for software version details. These three actions close the cheapest attack channels and require no new products. Assign one person to run the checks quarterly and keep a simple log. Escalate to a managed security provider if you handle payment data or regulated records.

How do vendors leak our information, and what can we do contractually?

Vendors leak through marketing case studies, support portals, misconfigured file shares, and their own breaches, as the Target contractor incident showed. Ask for approval rights over any public mention of your company, require incident notification within a defined window, and request deletion of your data at contract end. Review these clauses before renewal, because they're easier to add then. Audit what partners can actually access.

Are public records like business registries a real threat channel?

They are legitimate, but they verify facts an attacker needs: director names, registered addresses, and filing dates. That verification powers spear-phishing and invoice fraud that references real officers. You can't remove public records, so reduce the damage by training executives and finance staff on impersonation patterns. Where filings permit, use a business address rather than a home address for directors.

Does monitoring the dark web actually prevent attacks?

It shortens the window between credential sale and use, which is prevention of a sort. Monitoring services alert you when corporate credentials appear in marketplaces, but closed forums and private channels stay invisible, so absence of alerts proves nothing. Treat an alert as confirmation that a credential works until you verify otherwise. Pair monitoring with forced rotation and access review rather than relying on the alert alone.

This article is general educational information about attack reconnaissance and defensive practice, not a security audit, legal opinion, or guarantee of protection. Organizations handling regulated data, payment systems, or critical infrastructure should engage a qualified local security professional and check obligations with their legal counsel. If you suspect an active intrusion, contact your national computer emergency response team or law enforcement immediately.