The Consultant's Classic: Delete the Database, File the Ticket
Another week, another masterclass in consultant accountability theater. The story making the rounds — a major consultancy's contractor wiping a production environment and logging it as a "bug" — isn't just dark comedy. It's the inevitable output of an industry that has spent decades perfecting the art of failure without consequence.
The details are almost secondary: a consultant with elevated privileges, a "routine maintenance window," a cascade of deletions, and then the pièce de résistance — a Jira ticket filed solemnly categorizing the self-inflicted catastrophe as a software defect. Not human error. Not process failure. A bug. As if the database had developed a sudden autoimmune disorder and attacked itself.
The Paper Trail as Shield
This is where the consultant-industrial complex reveals its true genius. The ticket isn't documentation; it's armor. By framing catastrophic incompetence as a product defect, the consultancy shifts liability upstream to the vendor, downstream to the client's "inadequate testing," and sideways to "unclear requirements." The consultant becomes a mere observer — the messenger who happened to be holding the loaded gun when it discharged.
I've seen this script before. At a financial services firm in 2019, a Big Four partner's team dropped a production schema during a "low-risk migration." The incident report cited "an unforeseen interaction between the migration tooling and the database engine's constraint validation logic." Translation: they ran DROP TABLE without a WHERE clause. The vendor paid for the recovery. The consultant billed for the "remediation."
The Economics of Impunity
Why does this keep happening? Follow the money. Consultancies operate on a heads-I-win, tails-you-lose model. Success fees flow to the partner; failure costs land on the client's budget line items — "additional scope," "extended timeline," "unforeseen complexity." The consultant who deleted the data? They're already staffed on the recovery effort at $300/hour.
The industry's certification mill shares blame. We've created a caste system where a twenty-something with a Scrum Master badge and three weeks of Terraform training gets root access to payment infrastructure because their firm is "certified." The client's actual engineers — the ones who know which tables are load-bearing — watch helplessly from read-only Slack channels.
Bug Reports as Confession Booths
The "bug report" framing is particularly insidious because it weaponizes the client's own tooling against them. Every enterprise runs Jira, ServiceNow, or some homegrown equivalent. These systems are designed to track software defects — not human malpractice. When a consultant files "BUG-14729: Data loss during maintenance," they're not admitting fault. They're initiating a workflow that ends with a vendor patch, a credit note, or a change request. The consultant's firm gets paid either way.
Real software bugs have stack traces, reproduction steps, and git bisect trails. This had a timestamp, a username, and an audit log showing DELETE FROM customers WHERE 1=1. That's not a bug. That's a crime scene.
The Complicity of Procurement
Clients enable this. Procurement teams negotiate SLAs that penalize downtime but not root cause. A consultancy that causes an outage and fixes it in four hours meets their SLA. One that prevents the outage by saying "this is unsafe" gets billed as "blocking progress." The incentive structure rewards heroics over competence, cleanup over prevention.
I watched a CTO once reject a consultant's request for a read-only replica to test a migration script. "We're paying for experts," he said. "They shouldn't need training wheels." The consultant ran the script on primary. The recovery took three days. The consultancy got a contract extension for "crisis management."
What Accountability Would Look Like
Fixing this requires uncomfortable changes. Consultants who delete production data should be removed from the account, not reassigned to the cleanup. Firms should carry professional liability insurance that actually pays out — and premiums should spike after incidents. Clients should demand "blast radius" analyses before granting write access, not after.
Most importantly: the bug tracker must not be a confession booth. When a consultant files a ticket for their own destruction, the correct response isn't triage. It's an incident report with their name on it, a root cause analysis that leads to a human, and a contract clause that makes "we'll be more careful" legally insufficient.
The consultant who deleted the data and called it a bug didn't break the system. They followed it. The system is designed to absorb exactly this failure mode and convert it into billable hours. Until clients stop paying for the cleanup, the consultancies will keep bringing the matches — and invoicing for the fire department.
Next time you see a "bug" with no stack trace, no reproduction steps, and a consultant's name in the reporter field, don't assign it to engineering. Assign it to legal. That's the only language this industry understands.