Falls, which often result from sensory loss that affects balance, are a major cause of injury and death for people over age 65. Researchers have found that vibrating shoe insoles can greatly lower the risk of falls by exposing feet to random sensory noise. This seems paradoxical, since one might expect random noise to impair sensory connection with the ground. But underlying the research is a phenomenon called “stochastic resonance,” wherein noise—small, random deviations from regularities—can amplify weak signals, making them easier to detect. Numerous examples from biology and engineering show how the presence of random noise, counterintuitively, improves performance through stochastic resonance.

In business, noise is generally seen as the enemy of performance: deviation from plans leads to inefficiency; lack of standardization is a recipe for unmanageable complexity; inconsistency in execution is a threat to reliable quality. To combat such outcomes, executives pour billions of dollars and millions of hours into minimizing noise in planning and execution. Consider, for example, the Six Sigma methodology for eliminating defects in manufacturing by reducing variation, or “noise,” to under 3.5 parts per million. Introduced by Motorola in the 1980s, Six Sigma was quickly adopted throughout the world; at Jack Welch’s GE, it became a “way of life” and continues to be practiced today.

Standardization has many benefits—particularly when leaders are navigating stable and predictable contexts. But we argue that because unpredictability and change are unavoidable aspects of business, there are many situations where the shrewd business leader should embrace noise and use it to their advantage, rather than simply trying to eliminate it. Specifically, we will highlight three ways noise is used for advantage outside of business and suggest how companies could harness the benefits in a turbulent context.

Noise enables resilience and adaptability

Within species, there is significant noise in terms of random genetic variation, which manifests most obviously in diverse appearances and behaviors. For example, peppered moths exist in both light and dark variants. In the early 19th century, the lighter variant was predominant in Manchester, U.K. But during the Industrial Revolution, soot from new factories darkened the trees in the area. The light moths became easier for predators to spot, while their darker-bodied relatives could hide more easily. By the end of the 19th century, 98% of moths in the region were the darker variants. Mutation is a form of genetic noise that creates and sustains genetic variation, crucial to adaptation in the face of environmental fluctuations.

Businesses can use similar mechanisms by deliberately embracing noise in their operating models; noise creates adaptive capacity for dealing with fluctuating environments. GitHub, for example, already supported remote work for parts of their employee base before the pandemic, which meant that the company was better prepared than most when society transitioned to online work. Similarly, in 2005, Microsoft began experimenting with agile ways of working in small teams. This led to a cultural shift that is credited with helping Microsoft achieve its remarkable success over the past decade, by transforming the organization from a “giant battleship” to something “more like a flotilla of speedboats […] maneuvering in an orchestrated fashion.” Organizations that have experience supporting different kinds of operating models will be better positioned to integrate new models when they become advantaged in new environments.

It’s also possible to inject noise into daily execution. Companies like 3M, Atlassian, Apple, and Google have, at different times, allowed or even encouraged employees to work on side projects for a certain amount of time each week. At Google, the program has been credited with helping bring about many of the company’s biggest successes, including Google News, AdSense, and Gmail. Fostering a culture of experimentation can enhance innovativeness and the agility with which new opportunities are tackled.

Noise prevents us from getting stuck in suboptimal ways of thinking and doing

Noise is used extensively in the training of AI models. While the natural assumption is that we should use the cleanest possible training data, strategically adding noise (a technique known as “data augmentation” or “noise injection”) mitigates the risk of overfitting. In the latter scenario, the model essentially memorizes the specific details of the training dataset, rather than generalizing the underlying patterns.

Noise also provides benefits in low-resolution audio and visual systems, where “dithering” (adding random noise) can help overcome quantization error, leading to smoother, more intelligible signals, and in electronic and magnetic oscillators, preventing these systems from getting stuck in non-ideal states in the same way that a gentle nudge to a stuck pendulum clock can get it swinging again.

In business, noise can play a crucial role in preventing organizations from getting stuck in outdated or dysfunctional ways of operating, thinking, and making decisions. To gain these benefits, businesses must intentionally use noise to break up standard procedures and mental models. Tiger Tyagarajan, the outgoing CEO of global outsourcing firm Genpact, shared with us that one of the key learnings from his career has been the importance of disrupting successful stable teams, by challenging the status quo or rotating team members in order to prevent stasis and stagnation.

For example, noise can be injected into meetings to help participants entertain novel alternatives to their standard ways of thinking. This can come from appointing “red teams” that are charged with highlighting future scenarios that may be overlooked. IDEO, an innovative global design firm, encourages the use of Magic Circles, where teams explicitly set aside project constraints to brainstorm “outrageous ideas,” since it’s “easier to take crazy ideas and rein them [in]” than vice versa. These ways of introducing “conversational noise” help break existing mental models and prevent groupthink.

Noise can also be introduced by establishing temporary cross-functional teams or project-based work, which systematically disrupt established routines and team compositions to foster novel perspectives, approaches, and solutions. An alternative is embracing managerial rotation: For example, Alibaba implemented a program in 2012 that rotated their top 22 managers across their numerous business units. By instituting regular change in leadership, the program injected noise into the operations of each business unit, which ensured that they would not get stuck in peculiar or outdated ways of working, while also increasing the organization’s ability to adapt to new situations.

Noise helps uncover previously undetectable information

To see how stochastic resonance—the phenomenon that occurs with the vibrating insoles—works, imagine that we wanted to extract details from an overexposed photo like (a) in the series below. To do that, we can add random amounts of darkening or lightening to each pixel. When making all pixels randomly lighter or darker, the details that were hidden in the overexposed parts are more likely to become dark—and visible—than places where there was no detail to begin with. Thus, the picture becomes progressively clearer as noise is added (b, c).

Figure 1: Overexposed photo of BCG’s Founder Bruce Henderson with progressively more random noise added

Stochastic resonance is a well-understood phenomenon in physics, chemistry, engineering, and biology. For example, it enables crickets to more effectively detect predators and is routinely used in signal processing to uncover weak audio signals. Meanwhile, the potential uses of the phenomenon in business are relatively underexplored—but offer many promising possibilities.

A recent model shows how stochastic resonance can be used to detect weak signals in financial markets. This research suggests the possibility of exploiting fluctuations of asset prices to uncover new market information. Similar techniques can likely also be used to quickly detect trends and shifts in other kinds of quantitative data, including data changes in demand, consumer behavior and preferences, operational efficiency, and competitor activities.

Large Language Model (LLM) technologies might also be able to exploit noise to amplify weak signals in qualitative data, such as text. This could unlock the ability to more easily anticipate customer needs and preferences, find what’s important in product usage stories, understand what’s going on under the hood of employee feedback surveys, distinguish an organization’s message among its competitors, and quickly digest the salient aspects of new research and whitepapers.

Finding the right balance

While noise has it benefits, it is not a panacea. Too much noise in operations can increase errors and reduce reliability, leading to management burden and decreasing efficiency. Too much noise in your signals will muddle perception. So how should businesses balance the benefits of noise against the potential downsides?

One way that biological systems calibrate noise is by finding the “optimal mutation rate,” which balances the potential for beneficial changes against the risks of deleterious mutations. The optimal mutation rate is species- and situation-specific and depends on many factors, including environmental stability, how long individuals survive, how many offspring they have, and the sensitivity of these factors to disruption. For example, some species switch between self-fertilization and sexual reproduction depending on environmental conditions. Analogously, any given business situation has an optimal level of noise—which depends on the stability and predictability of the business environment, the risks of missing signals or getting stuck in outdated ways of operating, and the company’s ability to support noise, both financially and managerially.

Given the complexity of the factors, it’s impossible to give an exact formula for how to optimize the impact of noise; but, inspired by biology, we can start to understand when noise is helpful. During the early part of an idea lifecycle, for example, when you’re searching for new insights and perspectives, dialing up the noise can help you break free from conventional thinking patterns, expose hidden assumptions, and encourage the serendipitous connections that can lead to innovative breakthroughs. In turbulent environments, noise can help you get ahead of changes more quickly, better detect and adapt to emerging trends, and identify opportunities for innovation that might otherwise be obscured by the status quo. And when efficiency is high and the environment is more predictable and stable, turning down the noise can help you streamline processes and focus resources on the most promising initiatives while making sure that your rate of innovation is sustainable.

How to make your organization (enduringly) noisier

How can organizations lay the groundwork for embracing noise to their advantage?

Mind the noise: First, a mindset shift is required. Business leaders must stop seeing noise as purely an obstacle and start embracing it as a ubiquitous phenomenon that may have strategic value. In many organizations, this will need to be accompanied by a cultural shift—one that demands a departure from the traditional quest for uniformity and predictability and replaces it with an appreciation of variability and unpredictability as potential sources of long-term advantage.

Encourage diverse ways of thinking and doing: Embracing this mindset requires more than just tolerating noise. Companies must encourage, take seriously, and learn from new or unusual ways of thinking and doing. This doesn’t mean throwing away current mental and working models, but it does mean not getting stuck in them. This might involve seeking out interdisciplinary insights by engaging with experts in different fields and industries, allowing ideas and approaches to cross-pollinate. It might also involve fostering a culture of curiosity, risk-taking, and continuous learning. And of course, this will only work if employees feel empowered to express unique ideas and perspectives.

Make your talent strategy noisy: It’s no surprise that diverse ideas come from diverse people, so one step in using noise to your advantage is having talent with a broad array of backgrounds and perspectives. Diverse workforces can spark novel approaches to day-to-day tasks and foster an environment where questioning assumptions is welcomed.

But adding noise to your talent strategy goes far beyond championing diversity. It also means recognizing the value of mavericks, outliers, and employees who take unconventional approaches. A noisy approach to talent involves keeping an open mind about what kinds of people might successfully fill which roles, which includes loosening your grip on strict experience and background requirements and sometimes experimenting with atypical hires. A noisy approach to talent also means recognizing that sometimes it doesn’t make sense to hold everyone in the same role to the same strict evaluative standards, since different people can provide different kinds of value. When you’re balancing the benefits of noise with standardization and efficiency, some talent will provide value by being highly productive in a traditional sense while others might provide value by putting pressure on the traditional sense of productivity itself.

See the noise: Businesses often rely on averages and aggregates for reporting and performance monitoring, since they’re simple and straightforward. But such metrics hide noise: They can’t distinguish between every salesperson performing at similar levels in similar ways, and 1% of people producing 90% of the revenue. In addition to averages, it’s important to see the variability. Looking at all the data can be unmanageable, so instead you could consider box-and-whisker plots, other representations of the full data (like range and variance), and simulations, which can help you see the noise and its impact on outcomes.

When it comes to qualitative noise, because numbers can’t always capture what’s interesting, it’s vital to focus on and understand performance outliers with unconventional strategies. New tools like those from the collective intelligence startup CrowdSmart can also help. These tools use AI to display real-time visualization of ideas and conversations (including alignment and disagreement) to help identify hidden patterns and diverse perspectives.

Make your operating processes a little noisy: For many organizations, the goal of operating processes is to lay out a maximally efficient series of steps for employees to follow. While it’s good for processes to be efficient, processes that are too strictly standardized get into trouble when they face anomalies in execution, or the assumptions that made them efficient no longer apply. Processes that build in noise and variability can more easily handle anomalies and adapt to changing conditions, making them better for the company in the long run. Instead of giving strict recipes, define which steps must be executed precisely and leave room for variability in other parts of the processes. Injecting noise or variability into process inputs can also provide real-time assessment of whether a process has potential to become even more efficient, or whether the optimum has already been reached.

Capture the noisy learnings: For most organizations, knowledge management focuses on capturing paradigmatic cases and best practices. But doing so ignores the noise. In addition to the typical, also use your knowledge management systems to capture interesting anomalies, alternative approaches and mental models, emerging ways of doing things, and learnings from experimentation. The atypical may not be of particular use today, but it’s only by capturing the learnings from your noisy systems that you’ll be able to use them in the future.

Monitor and readjust your noise strategy over the long run: Once you’re set up to harness the benefits of noise, experimentation is key to calibrating the right amount. The challenge is that measuring the outcomes can be tricky. While the impact of using stochastic resonance to uncover signals can easily be measured, it’s tougher to quantify the benefits of breaking free from old ideas and increasing adaptability and resilience.

As a result, to assess the impact of noise, you’ll need to focus on long-term business health metrics such as innovation rate, and measures of long-term growth potential such as vitality, rather than one-off successes or failures. It can help to regularly reassess a scorecard that includes measures of efficiency (ROI, margins, and productivity) and the productive impacts of noise (products developed, ideas generated in brainstorming sessions, impressions from customer stories, employee survey findings about organizational flexibility, and ability to recover from disruptions).

Once you have a sense of the correlations between noise levels, efficiency, and your organization’s overall goals, refine your noise strategy by adding or reducing noise in specific areas based on the observed impact, or adjusting your response thresholds to account for its presence. Since the factors that affect the ideal balance, such as business environment stability, risk tolerance, and ability to support noise, are constantly in flux, you’ll need to regularly reassess and rebalance the strategy. And since your noise strategy will benefit from seeing early signals, not getting stuck in unproductive ways of thinking, and being resilient and adaptive, it makes sense to include noise in your noise strategy as well.

Instead of treating noise as an ever-present nuisance, businesses should embrace noise as a powerful, omnipresent truth, and a potential ally. By using noise strategically, your business can become more resilient and adaptive, break free from outdated patterns, and enhance its sensitivity to emerging trends—all of which can serve as significant sources of advantage in our constantly changing business environment.