Chapter 17 — Control and feedback basics
Lesson map
Lesson 17 published. Map, example, EX049.
flowchart TD
N0["Learning outcomes"]
N1["17.1 Feedback loops"]
N2["17.2 Designing timely feedback"]
N3["Lab — Simulating reporting with and without feedback"]
N4["17.3 Unstable or punitive loops"]
N5["Synthesis"]
N0 --> N1 --> N2 --> N3 --> N4 --> N5
N5 --> EX["Worked example"]
EX --> Q["Exercise EX049"]
Logical drawing
flowchart LR
N0["Learning outcomes"]
N1["17.1 Feedback loops"]
N2["17.2 Designing timely feedback"]
N3["Lab — Simulating reporting with and without feedback"]
N4["17.3 Unstable or punitive loops"]
N5["Synthesis"]
N0 --> N1 --> N2 --> N3 --> N4 --> N5 --> Q["EX049"]
[Learning outcomes]
|
v
[17.1 Feedback loops]
|
v
[17.2 Designing timely feedback]
|
v
[Lab — Simulating reporting with and without feedback]
|
v
[17.3 Unstable or punitive loops]
|
v
[Synthesis]
|
v
[Worked example] --> [Exercise EX049]
Learning outcomes
You will (1) apply feedback loops to security behaviour; (2) design timely feedback for reporting; (3) avoid unstable or punitive feedback.
17.1 Feedback loops
In control engineering, a feedback loop measures an output and adjusts an input to keep the system on target. Aiko applies this idea to behaviour. When staff report a suspicious email, they act; if they hear nothing back, the loop is open and they may stop reporting. If they receive a quick thank-you and later learn what the email was, the loop closes and reporting is reinforced. She sketches Greyford's reporting process as a loop and finds that reporters currently receive no response at all.
17.2 Designing timely feedback
Aiko designs feedback with three parts: an automatic thank-you within a minute, a short message when the security team confirms the email was malicious or safe, and a monthly summary showing how reports helped. Timeliness matters: feedback weeks later does little to reinforce behaviour. She estimates that automation can deliver the first two parts without extra staff time. The design mirrors control principles: measure quickly, respond quickly and keep the loop simple. She writes the three messages in plain language and asks five staff whether they would find them useful, adjusting the wording after their comments.
Lab — Simulating reporting with and without feedback
# Illustrative simulation of monthly reporting rate with and without feedback (synthetic dynamics).
def run(feedback, months=12, r=0.30):
out = []
for _ in range(months):
target = 0.60 if feedback else 0.15
r += 0.35 * (target - r) if feedback else 0.12 * (target - r)
out.append(r)
return out
for name, fb in (("no feedback", False), ("quick feedback", True)):
s = run(fb); print(f"{name:14} month 3 {s[2]:.0%}, month 6 {s[5]:.0%}, month 12 {s[11]:.0%}")
Run with python3 lab_feedback.py. Verified output:
no feedback month 3 25%, month 6 22%, month 12 18%
quick feedback month 3 52%, month 6 58%, month 12 60%
17.3 Unstable or punitive loops
Feedback can also harm. If staff who click a simulation receive harsh messages or are reported to managers, they learn to hide mistakes, and the loop drives the wrong behaviour. Excessive feedback, such as daily emails, causes people to ignore all messages. The lab simulates reporting under no feedback and quick feedback, showing that with feedback the reporting rate rises and stabilises, while without it the rate drifts down. Aiko notes the simulation's assumptions are illustrative, not measured. She recommends a monthly maximum for feedback messages.
Worked example
Aiko models reporting as a feedback loop, designs three-part timely feedback and simulates its effect while warning against punitive or excessive loops.
Synthetic case. Simulation parameters are synthetic.
Errors and limits
Simple loop models omit social influences. Automated thanks can feel insincere.
Exercises
- EX049 — Open loop. What happens when reporters hear nothing back? Rubric: 3 = the loop is open and reporting may decline;; 2 = decline;; 1 = vague;; 0 = none.
- EX050 — Feedback parts. What are the three parts of the feedback design? Rubric: 3 = instant thank-you, confirmation of verdict and monthly summary;; 2 = two;; 1 = vague;; 0 = none.
- EX051 — Lab outcome. How did reporting differ in the lab after 12 months? Rubric: 3 = 60% with feedback versus 18% without;; 2 = higher;; 1 = vague;; 0 = none.
Synthesis
Closing the feedback loop quickly and kindly sustains reporting, while punitive or excessive feedback destabilises it.
Continue studying — chapter 17 of 50
← Chapter 16 (draft) · Course chapter index · Chapter 18 (draft) →