bastion

Understanding Data Breaches

Build this skill

Distinguish a data breach from a data exposure, identify common causes and warning signs, and order the first steps of a breach response.

The caller who knew Mrs. Gokhale's medicines

On a Wednesday afternoon, Mrs. Lata Gokhale gets a call from a man who says he is from Tulsi Health. He knows her name, her address in Erandwane and that she orders blood pressure tablets every month. He says her last payment failed and asks her to read out the code that has just arrived on her phone. She hangs up and calls the store instead. By evening, three more customers have reported the same call. Somebody outside Tulsi Health has customer data they should never have seen, and that is the heart of a data breach: an event in which data is accessed, copied, changed or destroyed by someone who was not allowed to, or leaked to them.

Mrs. Gokhale 72, Erandwane Incoming call TH Tulsi Health Care unknown number ✕ ✓ Caller "Namaste, Lata ji. I am calling from Tulsi Health. Your Amlodipine order to Erandwane, Pune, had a failed payment. Please read me the code that just came on SMS." ! Real details name, address, order: the sign of a breach ✕ Asks for a code Tulsi Health never asks for one ✓ What she does: hangs up and calls the store on the number printed on her bill

A breach often shows up first as a scam that is too accurate. The caller's real details make the request for the code feel believable.

A door left open, or someone who walked in

Two words get mixed up all the time. A data exposure is data left reachable by people who should not reach it, for example a folder of supplier contracts set so that anyone on the internet can open it. Nobody may have looked yet. A breach is when someone actually did get in and saw or took the data. Think of leaving your flat's door unlocked while you are at college. That is an exposure: bad, and worth fixing at once. If you come home to find your laptop missing, that is a breach.

Data exposureData is reachable by outsidersNo proof anyone lookedFix the setting fastCheck logs: did anyone open it?Data breachSomeone got in or took dataData may be in criminal handsContain, investigate, notifyPeople may need to be warned

An exposure becomes a breach the moment someone who should not have the data opens it. Logs are often the only way to tell which one you have.

The difference changes what you do next. For an exposure you close the gap and then check records of who opened the data. Many cloud services keep logs, an automatic diary of who opened which file and when. If the logs show only Tulsi Health staff opened the folder, it stays an exposure. If they show unknown visitors downloading files, it was a breach all along. Without logs, you often have to treat an exposure as a breach, because you cannot prove otherwise.

Try sorting a few Tulsi Health situations. For each one, ask a single question: is there evidence that someone without permission actually got the data?

How breaches usually begin

Breaches rarely start with clever hacking. Most begin with something ordinary: a password stolen through a fake login page, or reused from another website that was breached earlier. Others start with a misconfiguration, a wrong setting, such as a cloud folder shared publicly or a database left without a password. Software that has not been updated leaves known holes open. People inside the company cause breaches too, sometimes on purpose, such as an unhappy employee copying data before leaving, but more often by mistake. And a supplier who holds your data, like Rathi Softech with access to MedLedger, can be breached instead of you.

Stolen loginsPhishing pagesReused passwordsNo two-step loginWrong settingsPublic cloud folderDatabase with no passwordOld accounts left onUnpatched softwareKnown hole not fixedOld app versionAttackers scan for itPeople and partnersMistakes in emailInsider copying dataVendor gets breached

Most causes are ordinary weaknesses rather than rare tricks, which is good news: ordinary controls stop most of them.

Each cause has a plain answer. Two-step login means a stolen password alone is not enough to get in. Checking sharing settings and switching off accounts the day someone leaves closes the wrong-settings gap. Installing updates promptly shuts known holes before attackers find them. Giving each person and each supplier only the data their work needs limits how much any single mistake or insider can take, which is often the difference between a breach of a hundred records and one of two lakh.

What a breach costs, and who pays

The first people to pay are the customers. With Mrs. Gokhale's name, address and medicines, a criminal can make very convincing scam calls, try to take over her accounts, or embarrass her by revealing a health condition. Unlike a card number, a medical history cannot be cancelled and reissued. Next comes the business: experts to investigate, time spent on the response instead of serving customers, possible penalties, and customers who quietly move to another pharmacy because they no longer trust it.

There are legal duties as well. In India, the Digital Personal Data Protection Act, 2023 requires an organisation that suffers a breach of personal data to inform the Data Protection Board and the people affected, and CERT-In, the national agency that handles cyber incidents, expects many incidents to be reported within a few hours of being noticed. You do not need the details at this stage. The point is that hiding a breach is not an option, and the clock starts when you find out.

How breaches are noticed

The uncomfortable truth is that breaches are often found late, and often by someone outside the company. Customers complain about scam calls, a researcher emails about an open folder, or the data appears for sale on a criminal website. Tulsi Health can find out sooner by watching for the signs: logins at odd hours or from cities where nobody works, one account downloading far more records than its job needs, alerts from the cloud service, and new accounts or settings nobody remembers creating. A breach found in a day causes much less harm than one found in a month.

Found by alertsFound by staff12 daysFound by customers40 days0 days5 days10 days15 days20 days25 days30 days35 days40 days45 daysAttacker gets in

Every extra day lets the attacker copy more, and gives criminals more time to use what was copied. Alerts that a person actually reads shorten the gap.

The first hours of a response

When a pipe bursts in your flat, you shut the main valve before you mop the floor, then you call a plumber, warn the neighbours below, and only later ask why the pipe failed. A breach response follows the same logic. First contain it, meaning stop it getting worse: disable the stolen account, reset passwords, close the public link. Then work out what happened and which data was involved, while keeping records and logs as evidence rather than wiping machines in a hurry. Then notify the people and authorities who must be told, fix the weakness that let it happen, and write down what you learned.

Put Neha's response to the scam-call breach in order. The order matters as much as the steps, because a good step taken too late lets the harm keep growing.

Notice that the warning to customers is part of the response, not an afterthought. Telling Mrs. Gokhale that Tulsi Health will never ask for a code on the phone protects her in a way no change to the servers can.

Follow your curiosity.

↑ ↓ to explore · Enter to open · Escape to close