What it means
The effect describes a measurement problem rather than a management technique. When a group knows it is being watched, output, accuracy and compliance often rise regardless of what has actually been changed, so the measured improvement mixes the real effect of the change with the effect of the attention itself.
It matters commercially because a great deal of business decision-making rests on pilots. A new incentive scheme, a picking system, a sales script or a safety procedure is trialled with one team, the numbers improve, and the improvement is used to justify a full rollout costing many times more than the pilot did.
The trouble is that the attention does not scale. When the same change is deployed across 40 sites with no researchers, no daily check-ins and no sense of being special, the uplift shrinks or disappears, and the business has budgeted for a benefit that was never really there.
The standard defences are straightforward to describe and harder to apply. Run a control group that receives the same attention without the change, extend the measurement well past the end of the observation period, and compare against a baseline drawn from data collected before anyone knew a trial was coming.
There is a useful nuance about how the original studies are usually retold. The findings were mixed and the effect is smaller and less consistent than the popular version suggests, so the practical takeaway is caution about pilot results rather than a belief that attention alone reliably lifts productivity.
The effect also shows up outside pilots entirely. Performance during an audit week, call quality when agents know their calls are being reviewed, and safety compliance while an inspector is on site all reflect the same mechanism, which is why sampling should be unannounced where that is fair and practical.
In practice
Real-world examples.
Example
A contact centre pilots a new call script with one team of twelve agents while a quality manager sits in daily. Average handling time falls by 18% and customer satisfaction rises. When the script is rolled out to 200 agents with no on-site observer, handling time improves by only 4%, and the operations director revises the savings forecast accordingly.
Example
A hospital trials a hand hygiene protocol on one ward with observers recording compliance. Compliance reaches 94% during the trial and drops to 61% once observation ends. The trust responds by building in periodic unannounced audits, treating the observation itself as part of the intervention rather than as a neutral measurement.
Example
A software company runs a four-week productivity experiment with a volunteer engineering team, tracking output daily and holding weekly review sessions. Velocity rises sharply, then falls back once the experiment finishes. The engineering manager reruns the test with a control team that gets the same review sessions but no tooling change, and finds most of the original gain came from the reviews.
Formula
Calculation
There is no formal equation, but the useful calculation separates the temporary observation effect from the durable one:
True effect = Post-observation performance - Baseline performance, and Hawthorne component = Observed performance during the trial - Post-observation performance
A warehouse team packs a baseline of 40 orders per picker per day. During a six-week trial of a new picking route, with an analyst on site taking daily measurements, output rises to 48 orders per picker per day, an increase of 20%.
Three months after the analyst leaves and the attention stops, output settles at 42 orders per picker per day. The durable improvement is 42 - 40 = 2 orders per day, or 5%, and the Hawthorne component is 48 - 42 = 6 orders per day, meaning 6 of the 8 extra orders, or 75% of the measured gain, was temporary.
The financial difference is what makes this worth checking. At a contribution of $25 per order across 250 working days, budgeting on the trial figure implies 8 x 250 x $25 = $50,000 of annual benefit per picker, while the durable figure supports 2 x 250 x $25 = $12,500, so the business case would have been overstated by $37,500 per picker.Case study
Seen in the real world.
Larkfield Assembly is an invented manufacturer used purely for this illustrative example. It piloted a new shift handover procedure at one of its five plants, with a continuous improvement lead present every day for eight weeks recording defect rates, changeover times and downtime, and the results were strong enough that the board approved a $1,400,000 rollout across the group.
Twelve months later the group-level numbers had barely moved. The improvement lead revisited the pilot plant and found that its own gains had largely faded once daily observation stopped, which suggested the procedure had contributed far less than the attention had.
Larkfield redesigned its approach to trials rather than abandoning the procedure. Future pilots ran with a matched control site that received identical management attention but no process change, measurement continued for six months after observers withdrew, and business cases were built on the post-observation figure. The illustrative point is that the company did not stop piloting; it stopped believing pilot numbers that had never been separated from the effect of being watched.
Watch out
Common mistakes.
- Building a rollout business case on the peak numbers from an observed pilot, without measuring what survives once the observation stops.
- Running a pilot with no control group, which leaves no way to separate the effect of the change from the effect of the attention.
- Concluding that the effect makes measurement pointless, when the answer is to design the measurement better rather than to stop measuring.
Questions
People also ask.
How long does the Hawthorne effect last?
It typically fades within weeks to a few months of the observation ending, which is why measurement should continue for at least a quarter after a trial finishes.
Can the effect be used deliberately?
To a degree, since visible attention and feedback genuinely do lift performance, but it decays without something durable behind it and should not be mistaken for a process improvement.
Does it apply to financial audits?
Yes in a limited sense, because controls are often applied more carefully during an audit period, which is one reason auditors use unannounced sampling and test transactions from across the whole year.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%Related