Why there is no clean number for the payoff
Most of the acoustic research on open offices is aimed at the sound, not the person. Studies compare partition layouts, measure how much speech leaks between workstations, and recommend spacing and orientation to protect privacy [1]. Others test which masking sound and volume work best, using listeners' ratings of speech comprehension, annoyance, and pleasantness [5]. These are real, useful results, but they stop at the acoustic and perceptual level. None of them follow workers to see whether their concentration improves or their mistakes drop [1][5].
The same gap appears in work on glazed partitions. That research weighs sound isolation against lighting, cost, and carbon, and describes voice-privacy concerns and acoustic discomfort, but it does not measure focus, errors, memory, or stress symptoms [4]. In other words, the interventions are well studied; the cognitive and stress consequences are not. That is why any specific percentage improvement would be invented rather than measured.
The closest evidence points in the right direction but stays general
The work that comes nearest to the question is a review of how noise in open-plan offices affects cognitive performance and subjective reactions. It treats speech privacy as one aspect of architectural privacy and examines masking sounds as a way to improve it, and it discusses stress symptoms and memory performance in the context of moving to a modern open-plan office [3]. That framing supports the idea that speech privacy and cognitive outcomes are linked, but a review of this kind does not hand you a clean effect size for a single acoustic change.
A study of water features as masking in two open-plan offices reports on cognitive performance and subjective satisfaction alongside speech privacy [2]. This is the kind of design that could, in principle, show a cognitive benefit from an acoustic change. But the available summary does not give a magnitude, so it supports the direction of the effect rather than its size.
What the acoustic studies do tell us about getting privacy right
Even without cognitive numbers, the acoustic findings show that the size of any downstream benefit depends heavily on design choices. Partition layout matters: parallel and crossed-rod layouts gave the best speech privacy at the lowest partition height tested, but those differences disappeared once partitions were tall enough [1]. In a field experiment at 1.5 m height, the face-to-face layout gave the least privacy [1]. The practical implication is that spacing workers apart and avoiding face-to-face seating are the levers that most directly protect speech privacy [1].
Masking sound is the other main lever, and it is not simply "more is better." In listening tests with 30 adults, white noise was rated the most unpleasant, brown noise at 49 dB was judged the most appropriate for masking, and pleasantness peaked at 46 dB [5]. Those are perceptual judgments about the masking sound itself, not measures of focus or stress, so they tell us how to set up a privacy intervention—not how much it will improve anyone's work.
How to read claims about focus and stress gains
If you see a specific figure—say, a percentage drop in errors after a privacy upgrade—treat it with caution unless the study actually measured errors before and after the change. The acoustic studies here measured speech privacy, not error rates [1][4][5]. The review and the water-feature study connect acoustics to cognitive and stress outcomes, but at a general level rather than with a quantified gain [2][3].
What the evidence does support is a reasonable expectation: improving speech privacy is a sensible step, and the design details—partition height and layout, masking sound type and level—are well enough understood to implement it thoughtfully [1][5]. The size of the resulting improvement in focus, errors, memory, and stress is the part that remains to be measured directly.
Sources used in this answer
Influence of partition layouts on speech privacy in open-plan offices
Partition layout affects speech privacy, with parallel and crossed-rod layouts best at low height and face-to-face worst, though layout differences vanish at greater partition heights.
The use of water features for improving speech privacy and cognitive performance in open-plan offices
Water features used as masking in two open-plan offices were examined for speech privacy, cognitive performance, and subjective satisfaction, though the summary gives no effect magnitude.
Subjective reactions to noise in open-plan offices and the effects of noise on cognitive performance-problems and solutions
A review links open-plan office noise to cognitive performance and subjective reactions, discussing masking sounds, speech privacy, stress symptoms, and memory performance.
The cost of transparency: Balancing acoustic, financial, and sustainability considerations for glazed office partitions
Glazed partitions involve a trade-off between sound isolation and speech privacy on one side and lighting, cost, and carbon on the other, without measuring cognitive or stress outcomes.
Investigation of the appropriate sound level of masking sounds with different sound sources in open-plan offices
In listening tests, white noise was rated most annoying, brown noise at 49 dB was judged the most appropriate masking sound, and pleasantness peaked at 46 dB.