Peak season is bearing down on every travel desk in Bihar, and with it comes the yearly pile-up: booking forms, scanned ID proofs, card details phoned in over WhatsApp, driver rosters, refund spreadsheets. Most small operators guard all of that with one shared password and good intentions. A DataSecOps routine gives you something firmer. The word is young — the name entered circulation on 4 August 2021, courtesy of Ben Herzberg, who was then Satori Cyber’s chief scientist, for an agile way of protecting data while the data and its users keep changing — and it never turned into an official standard; the formal label most policy bodies still use is data security governance. The label matters less than the habit. A tour operator’s data shifts daily: bookings land, seasonal staff rotate, taxi partners appear and vanish. The ten checks below take one afternoon, and they should be done before the first tourist bus of the season rolls out of Patna.
Why travel desks get hit without noticing
Data leaks in this trade rarely announce themselves. Nobody breaks a window; a login just keeps working long after the person who used it has left for a job in Delhi. In one published worked example of a data-protection review, 310 of 1,900 access grants — 16.3% — had sat unused for ninety days or more, and 212 of 940 user accounts, 22.6%, could reach more data than their jobs required. Scale those ratios to a small office with three shared inboxes and a decade-old laptop, and the picture is uncomfortable. The point of a checklist is to find those numbers in your own operation before someone else finds them for you.
Seasonal churn makes it worse. A counter clerk hired for the winter rush needs full access for four months, then none. Handled by hand, that handover always slips — which is why mature teams treat access like inventory rather than like a favour. One pattern worth copying from engineering teams: the asker states which dataset is needed, what for, and until when; the owner signs off before the grant exists; and an expiry date rides along automatically. It sounds bureaucratic until you count how many ex-employees could still open your booking sheet today.
The ten DataSecOps checks to run this week
- Write down every place customer data lives. Booking sheets, email threads, phone galleries, USB sticks, the old laptop in the store room. Pass mark: an inventory document that someone has updated this month.
- Mark the sensitive fields. Government ID numbers, payment details, medical notes attached to pilgrim tours. Automated scanners such as Google’s cloud classifier or Microsoft Presidio help, but none of them is fully accurate, so a human confirms anything tagged sensitive. Pass mark: a written list of fields, not a feeling.
- Count who can read what. Every account, every folder, every shared drive. Pass mark: no logins shared between two people, anywhere.
- Switch on multi-factor authentication. Password-only logins are on their way out across the industry — Snowflake, for one, is retiring them through 2026. Pass mark: no account anywhere that opens with a password alone.
- Put an expiry on every grant. Seasonal staff should lose access on their last day automatically, not whenever someone remembers. Pass mark: nothing older than ninety days without fresh approval.
- Watch the exits. Bulk exports — the moment data leaves your systems — deserve one named route and a log. Pass mark: you can say who sent the last big export, and why.
- Mask before sharing. Demo copies, vendor samples and training files should carry fake names and shifted ID numbers. Pass mark: nobody outside accounts can open a copy with real customer rows in it.
- Write rules as code, not as dashboard clicks. Policies clicked into a console never survive a migration, and copies made by hand drop the tags that carry them. Pass mark: rules stored in a file that can be re-applied elsewhere.
- Read the access log monthly. Not skim — read. Pass mark: one named person signs off on the month’s reading, every month.
- Hold a review that revokes. A quarterly meeting that only re-confirms everyone is theatre. Pass mark: at least one access right removed per cycle.
Two incidents this DataSecOps list is built to prevent
The first is the credential-stuffing wave of 2024, tracked as UNC5537, which swept through customers of the Snowflake data platform. Close to eight in ten hit accounts — 79.7% — had already shown up in infostealer records, multi-factor authentication had never been switched on, and some passwords had not been rotated in four years or more. Roughly 165 organisations had to send notifications to their own customers. Snowflake’s own systems stayed intact; the thieves simply walked the data out with plain COPY INTO export statements run on stolen logins. Every one of those failures is caught by checks four, five and six above.
The second failure is quieter and self-inflicted: rules that live only in a management console. They feel like protection, but they are invisible to any migration, any manual copy, any CREATE OR REPLACE rebuild — all of which quietly drop the tags and leave the data bare. This is why check eight insists the rules live in code.
Scoring yourself and planning the first month
Eight or more passes means you are season-ready. Six or seven means fix the gaps this fortnight. Anything lower, and the 30-day starter plan is the right response: write the classification standard, run a discovery pass over every store, give the top fifty tables — or your entire inventory, if it is small — a proper look, and get multi-factor authentication onto every account. The maturity ladder behind this runs from ad hoc to classified, then enforced, then evidenced, then adaptive — the last being the state in which any newly created sensitive table is gated before the first person is granted to it, without exception. Exiting level two takes coverage above half your tier-one tables; level three wants better than ninety percent and zero console-only policies. The full loop behind these steps — find, label, gate, watch, review — is laid out in the DataSecOps entry in xOps School’s Ops Atlas, and it doubles as the syllabus for training a new office manager.
Where the discipline meets the itinerary
None of this replaces the work of running tours; it just keeps the back office from undoing it. The same care you apply to customer records, you already apply to trip planning — the family travel guide for Bihar exists because someone verified every stop on it. Verification is the same instinct that confirms a scanner’s guess about a sensitive column. And when guests ask what documents they must carry, your office already follows the habit of clear checklists, the way rental pickup requirements are spelled out in advance. Data protection is that same checklist instinct, pointed inward. Run the ten checks, keep the evidence, and let the season’s chaos stay outside the ledger where it belongs.
