Resolve the Conflict Between Efficiency and Resilience

Operational efficiency is critical for both financial success and customer satisfaction. Efficient systems, characterized by minimal buffers and idle time, tight schedules, and maximum asset utilization, allow organizations to do more with less, thereby boosting revenue and appealing to time-sensitive customers. However, such systems often lack resilience, increasing an organization’s vulnerability to operational disruptions.

The tension between efficiency and resilience is especially visible in the airline industry. A resilient airline network can absorb disruptions, protect passengers from severe service failures, and recover quickly without incurring excessive costs. But airlines also face constant pressure to offer faster itineraries and maximize the use of costly resources, such as their fleet of aircraft and flight crews. Meanwhile, passengers strongly favor efficiency as well, in the form of shorter travel times with minimal layovers. Those preferences can lead to itineraries with little to no time buffer to absorb the delays or cancellations common to air travel, leaving passengers frustrated and stuck waiting in terminals. Moreover, such disruptions propagate throughout interconnected networks, affecting passengers, flight connections, crew schedules, and aircraft positioning. These ripple effects result in significant financial and reputational damage for airlines.1

This challenge is not unique to airlines, however. Supply chain managers need to balance inventory costs against the risks of stockouts. Health care systems strive to optimize patient flows and increase throughput while maintaining quality of care. Despite the operational differences across these contexts, the fundamental challenge is the same: How can organizations design operations that are both efficient and resilient?

Our analyses of millions of flights and airline passenger journeys in several academic studies reveal why managers do not need to treat efficiency and resilience as opposing goals. We identify three actionable strategies that enable organizations to achieve both objectives, whether managing flight schedules, patient flows, call center operations, or global supply chains.

Strategy 1: Measure What Matters to Customers

Traditional operational performance metrics often do not reflect customer experience, leading to perverse incentives that can weaken actual service quality and system resilience. For example, in the U.S. airline industry, the Department of Transportation (DOT) publishes its Air Travel Consumer Report each month. The report includes statistics for all major carriers on the percentage of flights that arrive within 15 minutes of the scheduled arrival time. Known as on-time performance (OTP), this metric serves as a proxy for service quality and reliability in the DOT’s and others’ rankings of airlines.

At first glance, publicizing OTP appears to benefit consumers: One might expect that being judged on this metric incentivizes airlines to reduce flight delays. However, as is often the case with KPIs, organizations give in to the temptation to game the metric: Airlines often add time to the flight durations they publish, to improve their chances of being “on time” — a practice known as schedule padding.2 This helps airlines boost their OTP statistics without meaningfully improving their reliability.

Putting the airlines’ strategic gaming behavior aside, it is questionable how useful a measure OTP is for customers. Non-stop-flight passengers are likely to experience a 20-minute-late arrival as only a marginal difference, and they’re unlikely to be concerned that their flight missed the scheduled arrival time by 14 minutes but was “on time” within the DOT’s 15-minute cutoff. However, for a connecting passenger with a tight layover, a 14-minute delay may be long enough to cause a missed connection — and hours spent in the airport waiting to get onto another flight.

Given that OTP does not accurately capture what matters to passengers, in a recent study we published, we urged the DOT to release more informative passenger-level statistics, such as the proportion of passengers reaching their final destinations within defined delay time intervals (within 15 minutes, 15 minutes to one hour, one to two hours, two to three hours, and more than three hours of the scheduled arrival time).3 In the long run, shifting the focus to passenger travel times (including total flight times, layover times, and potential missed connections and flight delays in each leg) can incentivize airlines to pay more attention to passengers’ travel experiences. Both efficiency and resilience could improve as a result, because airlines would have a stronger incentive to create more efficient routes and itineraries, with shorter total travel times that are less susceptible to missed connections and long delays.

Similar issues arise in other industries, where the operational performance metrics an organization uses can negatively affect efficiency and reliability. For example, in health care, hospitals often track operating room utilization rates as a key efficiency metric. While high utilization sounds desirable, it can incentivize hospital administrators to create extremely tight surgery schedules, with back-to-back procedures. This practice leaves little buffer for unexpected complications or overruns, causing subsequent patients to experience long delays or even cancellations and negatively affecting efficiency, care quality, and patient satisfaction.

In supply chain management, companies typically use high inventory turnover ratios as a proxy for operational efficiency. However, an excessive focus on this metric can prompt businesses to cut safety stock too aggressively, heightening the risk of shortages, delaying downstream production, and leaving customers with long wait times for the products they want to purchase. And in customer service, when call centers measure agents’ performance by the number of calls they handle, that may motivate agents to rush their interactions with customers. That, in turn, can lead to a reduction in problem-resolution quality and an increase in repeat calls, thereby prolonging the total time it takes to resolve an issue. These examples illustrate how narrowly defined metrics can distort incentives, prompting companies to optimize for the metric rather than the actual service experience.

Defining the right performance metrics can be challenging because metrics must balance an organization’s efficiency objectives with other managerial considerations (such as service reliability, customer satisfaction, and fairness). One way to attain this balance and ensure that high efficiency does not come at the expense of customers is to treat performance measurement as a paired system, with one metric that tracks operational efficiency and another that measures what matters to customers. For example, in our recent study, we developed two metrics — one to track efficiency and the other to assess resilience — to ensure that an airline’s focus on efficiency (measured by short scheduled travel times for passengers) would not lead to long travel delays due to missed connections.4 Similarly, hospitals can measure both operating room utilization and the delays surgical patients experience when extremely tight surgery schedules are disrupted.

Appropriate performance metrics should be easy for stakeholders to understand while accurately capturing an organization’s performance objectives and customer satisfaction levels. Defining such measures requires input from key stakeholders, including customers. Misaligned incentives pose another risk, especially when metrics are tied too rigidly to employee performance evaluations: Such correlations may incentivize employees (or even an entire organization, as documented in the airline industry) to game the system rather than improve service quality. Thus, designing incentive structures that reward long-term service quality based on a combination of efficiency and resilience rather than short-term metric gains is another important step. Finally, periodic reviews and stakeholder input can help ensure that metrics evolve with changing customer expectations and operational dynamics.

Strategy 2: Avoid a One-Size-Fits-All Approach and Deploy Buffers Strategically

Measuring what matters to customers can motivate an organization to be more proactive in preventing major service or supply failures and improving its resilience to disruptions. In the airline industry, major service failures often take the form of flight cancellations and long delays, which hurt airlines’ current and future financial performance.5 While some disruption triggers, such as severe weather, are beyond companies’ control, airlines can influence how disruptions propagate through their networks by strategically designing flight schedules.

Strategic scheduling involves building sufficient flight and ground time buffers to reduce the risk of delays cascading from one flight to the next. Traditionally, some airlines have relied on simple rules of thumb, such as using historical data to estimate the average time to complete a flight and then adding a fixed buffer. For example, if a flight averages one hour, adding a 15-minute buffer results in a scheduled flight time of one hour and 15 minutes. However, this one-size-fits-all approach can negatively affect both efficiency and resilience: A buffer may be unnecessarily long for some flights, reducing efficiency, but insufficient for others, increasing an airline’s vulnerability to disruptions.

A more strategic and data-driven approach considers both the likelihood and consequences of a disruption. In the airline industry, our analyses revealed that airport congestion levels, layover times, weather conditions, and the time of day predict the likelihood of a disruption.6 Moreover, the operational consequences of a disruption vary by flight: A delayed flight with many connecting passengers can create severe ripple effects, whereas a delayed aircraft on its last flight of the day, carrying no connecting passengers, will have minimal impact on the network. Allocating larger buffers where disruption risk and impact are high, and smaller buffers where they are low, can improve efficiency and resilience simultaneously.

The same principles apply to other industries. In health care, setting the duration and sequence of surgeries based on patients’ risk profiles and the consequences of potential delays to subsequent procedures can improve overall system performance.7 In call center operations, staffing schedules can be determined by considering not just the expected length of each call but also the likelihood of follow-up interactions resulting from unresolved issues. In supply chain management, rather than applying uniform inventory policies across all products, companies can tailor their safety stock levels and replenishment intervals to account for supply disruption risks and their downstream consequences, such as production delays due to material shortages. In project management, allocating contingency time to tasks with high interdependencies or critical-path activities can prevent small delays from cascading into major schedule overruns.

Deploying buffers strategically is easier said than done. It requires granular data, advanced analytics, and coordination across multiple business functions (such as marketing, network planning, and an airline’s ground operations teams). Historical averages and one-size-fits-all approaches are simple and familiar, which makes them hard to replace with dynamic, context-specific, risk-based approaches. Moreover, managers often face resistance when buffer adjustments appear to reduce efficiency in the short term, even though they can improve both efficiency and resilience in the long run. To overcome these challenges, organizations should start small by simulating what-if scenarios and piloting data-driven scheduling in high-risk areas within a division. They should also invest in predictive analytics to identify disruption patterns and communicate the long-term benefits of resilience to stakeholders. Embedding analytics into planning processes and getting buy-in from stakeholders can ensure that buffer allocation becomes a strategic lever rather than an ad hoc decision.

Strategy 3: Curate Personalized Customer Options to Maximize System Performance

In many organizations, operations teams prefer to limit the number of product or service options presented to customers to keep processes streamlined. Sales and marketing teams often advocate for more options to increase the likelihood of meeting diverse customer needs. However, offering a broad choice set significantly increases operational complexity, which can compromise reliability.

In the airline industry, itineraries with layovers highlight the risks of offering too many options to passengers. For instance, consider a passenger who planned to travel from Knoxville, Tennessee (TYS), to Greensboro, North Carolina (GSO), on Sunday, Jan. 11, 2026. Upon searching for flights on a popular travel website, we found that American Airlines offered 13 itinerary options via its hub in Charlotte, North Carolina (CLT), including one with only a 30-minute layover, where the first leg would arrive at CLT at 1:38 p.m. and the second leg would depart for GSO at 2:08 p.m. Similarly, Delta Air Lines offered six options through its hub in Atlanta (ATL), one of which allowed just 40 minutes between flights, with the first leg arriving at ATL at 5:13 p.m. and the second leg departing from ATL at 5:53 p.m. Given the sizes and congestion levels of CLT and ATL, both itineraries posed a significant risk of a missed connection. More broadly, airline reservation systems routinely display such options as long as they meet the minimum connection times published in the International Air Transport Association database, even when they leave little margin for actual delays.

In our recent study, we used a proprietary passenger-level data set provided by Southwest Airlines to simulate the impact of tight layovers on system performance.8 Our analysis found that identifying itineraries with short layovers and preventing passengers from booking them can significantly enhance resilience by reducing missed connections. Notably, curating the choice set by removing risky itineraries does not materially deteriorate efficiency, given that switching from an itinerary with a short layover to one with a slightly longer layover typically increases the total scheduled travel time only marginally (by 10 to 15 minutes, in many cases). Indeed, the actual travel time of an itinerary with a slightly longer layover may be significantly shorter due to the elimination of a missed connection.

The concept of curating customer choices to enhance resilience can be applied to many industries. In health care, to reduce the risk of cascading delays, hospitals could limit scheduling options for elective surgeries when operating room capacity is constrained. To avoid missed commitments, retailers might restrict delivery time slots during periods of high uncertainty in supply chains. By strategically shaping the choice set rather than leaving all options open, businesses can mitigate vulnerabilities without materially compromising customer experience, especially when the curated choice sets impose only minor trade-offs in efficiency.

Implementing curated choice sets is far from straightforward. It requires predictive models that accurately assess risk under varying conditions, as well as real-time systems to update available options dynamically. Internal tensions naturally arise when multiple functions, such as marketing, operations, and IT, must agree on where to curtail customer choices. In particular, marketing teams may worry about lost revenue from eliminating certain options. It is thus essential that organizations quantify both the revenue potential of risky options and the actual costs of disruptions that arise as a result of offering them. (For airlines, paying for stranded travelers’ hotel accommodations, meal vouchers, and rebooking expenses erodes profitability.) Organizations should first pilot curated options in high-risk contexts to validate their benefits. They can use the same analytical insights that helped them curate options to explain to customers that their choices have been limited to safeguard reliability.

In an era of rising customer expectations and increasingly complex service systems, overcoming the traditional efficiency-resilience trade-off is a critical operational skill that can define the trajectory of a service organization. Companies that can deliver both speed and reliability will not only meet customer demands but also differentiate themselves in highly competitive markets. The key managerial takeaway from our research studies collectively is that organizations should proactively design their operations to build resilience into their systems rather than relying on reactive, ad hoc fixes after disruptions occur. Such proactive design requires that they choose performance metrics that reflect customer experience, build systems that can absorb variability, and shape customer choices so that the organization continues to run reliably when conditions become challenging. As disruptions become the norm rather than the exception, reconciling efficiency and resilience by design rather than reaction can separate companies that cope from those that succeed.