Chapter 10 — Privacy and data protection: quality and troubleshooting
Lesson map
Lesson 10 published. Map, example, EX046.
flowchart TD
N0["Learning outcomes"]
N1["10.1 A recurring problem"]
N2["10.2 Root cause and fix"]
N3["10.3 Quality checks"]
N4["10.4 Measuring improvement"]
N0 --> N1 --> N2 --> N3 --> N4
N4 --> EX["Worked example"]
EX --> Q["Exercise EX046"]
Logical drawing
flowchart LR
N0["Learning outcomes"]
N1["10.1 A recurring problem"]
N2["10.2 Root cause and fix"]
N3["10.3 Quality checks"]
N4["10.4 Measuring improvement"]
N0 --> N1 --> N2 --> N3 --> N4 --> Q["EX046"]
[Learning outcomes]
|
v
[10.1 A recurring problem]
|
v
[10.2 Root cause and fix]
|
v
[10.3 Quality checks]
|
v
[10.4 Measuring improvement]
|
v
[Worked example] --> [Exercise EX046]
Learning outcomes
You will (1) diagnose a recurring privacy problem; (2) improve process quality; (3) measure improvement.
10.1 A recurring problem
After the deletion job starts, two candidates complain that they were deleted while still in active recruitment processes. Marta investigates. She finds that the job uses the last contact date, which is updated only when a consultant logs a call or email. Consultants often communicate through networking messages that are not logged in the recruitment system. The data were therefore inaccurate: active candidates appeared inactive. The problem is a data quality failure that causes a privacy control to act on the wrong people.
10.2 Root cause and fix
The root cause is that the system's idea of contact does not match real practice. Marta works with consultants to add a simple button, mark active, that updates the date without logging details, and the deletion job now sends a warning to the responsible consultant seven days before deleting a profile. The consultant can confirm or extend with a reason. She notes that Article 5(1)(d) requires accurate data, and that accuracy matters for automated decisions about retention as much as for content.
10.3 Quality checks
Marta adds quality checks: a monthly report of profiles with an active application but a last contact date over three months old, and a sample review of five profiles per consultant each quarter. The first report finds thirty-one such profiles, which consultants update. She explains that quality checks catch problems before they cause harm, such as wrongful deletion or, in the opposite direction, keeping candidates who should have been deleted because someone updated the date without real contact. Consultants receive their own lists.
10.4 Measuring improvement
Over the next quarter, no candidates are wrongly deleted, and the number of profiles flagged by the monthly report falls from thirty-one to six. Consultants report that the warning emails are useful. Marta also monitors misuse of the mark active button by sampling extended profiles for a recorded reason. Ms. Lin observes that good privacy controls depend on good data and good processes, and that troubleshooting often reveals a gap between how a system assumes work happens and how people actually work.
Worked example
Marta traces wrongful deletions to inaccurate last contact dates, adds a mark active button and seven-day warnings, introduces monthly and quarterly quality checks and sees flagged profiles fall from thirty-one to six.
Synthetic case. The problem is synthetic; GDPR Article 5(1)(d) is real.
Errors and limits
New buttons can be misused to keep data indefinitely; sample their use.
Exercises
- EX046 — Cause of deletions. Why were active candidates deleted? Rubric: 3 = last contact date not updated for networking messages;; 2 = mentions date;; 1 = vague;; 0 = none.
- EX047 — Warning period. How long before deletion is the consultant warned? Rubric: 3 = seven days;; 2 = partly correct;; 1 = wrong;; 0 = none.
- EX048 — Accuracy principle. Which article requires accurate data? Rubric: 3 = Article 5(1)(d);; 2 = partly correct;; 1 = other;; 0 = none.
- EX049 — Flagged profiles. How did flagged profiles change? Rubric: 3 = from thirty-one to six;; 2 = partly correct;; 1 = wrong;; 0 = none.
- EX050 — Button misuse. How is misuse of mark active checked? Rubric: 3 = sampling extended profiles for reasons;; 2 = mentions sampling;; 1 = vague;; 0 = none.
Continue studying — lesson 10 (chapter 10 of 12)
← Chapter 09 · Course chapter index · Chapter 11 →
Continue to the next authored lesson.