Operating Models for Uneven Demand: Designing for Peaks, Lulls, and Mix
- RESTRAT Labs
- 14 hours ago
- 15 min read
When demand fluctuates - think weekend restaurant surges or slow weekdays - most businesses struggle. Why? Many rely on average-based models that assume steady demand, which rarely exists. Peaks overwhelm systems, lulls waste resources, and mixed tasks stretch capacity. Here's how to fix that:
Stop Using Averages: Systems built on averages fail during spikes or slumps.
Buffers Work: Protect critical areas, like kitchens or crews, with extra resources.
Separate Tasks: Pre-plan routine work during slow periods to handle spikes better.
Flexible Staffing: Use part-time workers or cross-train teams for quick adjustments.
The key takeaway? Design systems to handle variability by preparing for extremes, not averages. This protects profits, prevents burnout, and ensures smoother operations. Let’s dive deeper into how to make this work.
C3 Variability is the Villain in Operations and Must be Managed
Why Models Built for Averages Break Under Variability
Relying on averages to build operating models assumes a level of stability that rarely exists in reality. W. Edwards Deming once pointed out that variation isn't a flaw in execution - it's simply how systems naturally behave. When businesses design their operations around averages, they end up preparing for a scenario that almost never happens. This disconnect becomes glaringly obvious when real-world fluctuations hit.
Let’s dig into why this approach falls apart, using the concepts of non-linear queueing effects and shifting constraints. Queueing effects, in particular, behave unpredictably under strain. Donald Reinertsen’s research on flow economics shows that as demand rises, wait times don’t just increase steadily - they can skyrocket. A system that runs smoothly under moderate demand can grind to a halt when pushed closer to full capacity. Even a small spike in demand can lead to massive delays, creating a domino effect that overwhelms the system. This is why systems designed for average demand often collapse when things get even slightly unpredictable.
Eliyahu Goldratt’s Theory of Constraints adds another layer to the problem. In systems with variable demand, bottlenecks don’t stay in one place. For example, a contractor might find crew capacity to be the limiting factor one week, only to face permitting delays the following week. Similarly, a restaurant might hit a kitchen bottleneck during a busy Friday night, while on a quieter Sunday, the constraint could shift to the front-of-house staff. Average-based models assume a static bottleneck, leaving businesses scrambling when the real constraint moves unexpectedly.
"Analyzing these systems is not straightforward, because standard queueing theory focuses on the long-run steady-state behavior of stationary models." - Linda V. Green, Peter J. Kolesar, and Ward Whitt [1]
This "steady-state" approach creates weaknesses. Without flexible resources - like cross-trained staff or adaptable processes - businesses can’t shift capacity to where it’s needed most. This results in a frustrating paradox: idle resources in one area and backlogs in another, even though the system’s total capacity might seem sufficient.
Peaks Create Bottlenecks and Burnout
Demand peaks are notorious for creating bottlenecks and burning out employees. Take a hospitality business during the December holiday season. A restaurant designed for average demand might suddenly face a surge in weekend reservations, especially during peak dining hours. The kitchen becomes a bottleneck, and the non-linear queueing effect amplifies delays throughout the system. Even a slight increase in demand can lead to disproportionate wait times.
The fallout isn’t just about lost revenue from turning away customers. When systems can’t handle peaks, the burden falls on employees. Managers may work overtime, and owners might step in to fill gaps. While these stopgap measures can help the business survive the rush, the long-term costs - like staff burnout or higher turnover - can be devastating.
Lulls Leave Fixed Costs Idle
On the flip side, demand lulls create their own set of challenges. Fixed costs, like maintaining a full crew or operating a facility, don’t disappear just because demand drops. For instance, a trade contractor might keep a team ready for peak construction activity, only to have them sit idle during delays caused by permits or bad weather. Unlike manufacturing, where inventory can smooth out demand fluctuations, service-based operations can’t "store" unused capacity.
The instinct during lulls is often to cut costs by reducing capacity. But this can backfire. Layoffs or reductions in skilled staff might save money in the short term, but they can lead to the loss of valuable expertise and higher rehiring costs when demand picks back up. The root issue is that average-based models don’t account for the need to adjust resources dynamically, treating all capacity as equally necessary all the time.
Mixed Demand Types Overwhelm Shared Resources
Even when overall demand stays steady, variability in the type of demand can wreak havoc on systems built for averages. Consider a service contractor juggling emergency repairs and routine maintenance. Emergency work requires immediate attention, while scheduled maintenance needs careful planning. When both types of tasks rely on the same resources, urgent jobs can disrupt planned workflows, causing delays across the board.
This issue - referred to as "floating bottlenecks" [4] - happens because the system’s constraint shifts depending on the current mix of work. For instance, a plumbing service might seem adequately staffed overall but hit a wall when specialized equipment is needed for multiple emergency calls at once. Average-based models fail here because they assume resources are interchangeable and demand is uniform - assumptions that rarely hold up in practice.
Understanding these flaws in average-based models is the first step toward creating systems that can handle both surges and slumps effectively.
How to Design Operating Models That Absorb Demand Swings
Building operating models that can handle fluctuations in demand without breaking down requires thoughtful planning. Donald Reinertsen’s research on flow economics highlights the balance between capacity, speed, and cost, while Eliyahu Goldratt’s work on constraints emphasizes the importance of using buffers to protect workflow. Combining these ideas helps create systems that manage variability effectively without relying on last-minute, high-stress fixes.
The goal isn’t to add slack everywhere - that would be inefficient. Instead, it’s about strategically placing buffers, decoupling demand from execution, and creating adaptable capacity in areas most prone to strain. Tools like buffers, decoupling points, work classification, flexible staffing, and clear policies can help businesses handle peaks and lulls more effectively.
Where to Place Capacity Buffers and Why They Work
In service operations, capacity buffers often take the form of "service inventory", a concept described by Sunil Chopra and Martin Lariviere. This involves completing preparatory work - like staging or documentation - before the customer arrives.
"Service inventories allow firms to buffer their resources from the variability of demand and reap benefits from economies of scale while also providing customers with faster response times." – Sunil Chopra and Martin A. Lariviere [5]
Dell provides a great example by adjusting its "push-pull boundary" to reduce mismatches between forecasted and actual demand [5].
Cross-trained staff can also act as a buffer. Research on "skill chaining" shows that even minimal cross-training can make a big difference in serial production lines. For instance, a restaurant might temporarily assign a server to assist in the kitchen during a dinner rush or have a prep cook help out front during slower periods. These shifts in capacity ensure smooth operations during demand surges [3]. The cost of maintaining this flexibility is often far lower than the potential losses from rushed work, missed revenue, or employee burnout.
Decoupling Points: Separating Demand from Execution
Decoupling demand from execution is another way to stabilize operations. Decoupling points separate tasks completed ahead of demand from those done after it arises. This "push-pull boundary" [5] allows part of the work to be planned and optimized in advance (the "push" side), while the rest is handled responsively (the "pull" side).
For example, a trade contractor might handle permitting, material staging, and crew scheduling ahead of time (push work), while installation waits until the customer is ready (pull work). This separation reduces the burden during peak periods since fewer tasks need to be handled in real time.
Decoupling also opens the door for automation and self-service options. Tasks on the "push" side can often be automated or even completed by customers. For instance, a hospitality business might automate reservations to free up staff during busy times. In industries with fluctuating demand, refining staffing models to account for time lags can further ease peak congestion [1].
Classifying Work by Demand Profile
Once tasks are decoupled, classifying work by its demand profile helps allocate resources more effectively. Not all tasks are created equal - some are predictable and can be scheduled during slower times, while others are unpredictable and need immediate attention. By categorizing tasks as either steady-state or surge work, businesses can better manage their resources and avoid overwhelming shared capacity.
Steady-state tasks, like maintenance, administrative work, or process improvements, can be planned in advance. Surge tasks, such as emergency repairs or rush orders, require immediate attention. By tackling steady-state work during lulls, businesses can keep fixed costs productive while reserving capacity for high-demand periods [5].
This classification also prevents bottlenecks. Assigning dedicated resources or policies to each type of work ensures that surge tasks don’t disrupt steady-state processes, keeping the system running smoothly.
Flexible Staffing and Vendor Arrangements
Fixed staffing models assume consistent demand, but flexible staffing models are better suited for variability. This doesn’t mean constantly hiring and firing - it’s about building relationships with vendors, part-time staff, or contractors who can step in during peak periods without adding full-time costs during slower times.
For instance, a trade contractor might maintain a core team for regular operations while relying on subcontractors during busy periods. Similarly, a hotel might use on-call staff during holidays or weekends. The idea is to avoid overcommitting to fixed costs while ensuring enough capacity to handle spikes in demand.
Studies on structural flexibility show that even small adjustments - like limited cross-training or multipurpose roles - can make a big difference, especially in systems with parallel queues [3]. Organizations don’t need to make every resource fully flexible; targeted flexibility in key areas is often enough to manage most demand swings. Establishing vendor relationships well in advance of peak periods ensures scalability without the chaos of last-minute arrangements.
Clear Policies for Surge Versus Steady-State Work
Without clear policies, teams may treat all tasks as equally urgent, leading to inefficiencies, burnout, and constant disruptions. Defining how to handle surge versus steady-state work helps streamline operations and protect employees from overload.
For example, a policy might route emergency calls to a dedicated team while scheduling routine maintenance during slower periods. Another policy could set service-level agreements that prioritize rush orders (e.g., same-day service) over standard ones (e.g., batched and processed regularly). These policies reduce cognitive load and ensure that resources are allocated appropriately.
Clear policies also help maintain profitability. Surge work can often command higher prices, while steady-state tasks benefit from predictable, efficient pricing. By clearly distinguishing between the two, businesses can ensure stability and avoid the need for constant firefighting. Teams can focus on their work with confidence, knowing that the system is designed to handle variability effectively.
SMB Studio Examples: Contractors, Hospitality, and Owner-Led Businesses
Small and medium-sized businesses (SMBs) often face unpredictable demand swings. Whether it’s trade contractors dealing with weather delays, hospitality businesses navigating weekend peaks, or owner-led ventures struggling to manage everything personally, these fluctuations can wreak havoc on margins and team morale. The solution lies in carefully designed operating models that can handle these shifts without breaking the system.
Trade Contractors: Managing Weather and Permit Challenges
In construction, weather and permit approvals are two major disruptors. Imagine a construction crew scheduled for exterior work, only to be sidelined by rain. Without a plan, those workers sit idle, even though there are plenty of other tasks waiting. On the flip side, when permits finally clear after weeks of delays, the sudden demand for action can overwhelm teams that aren’t ready to scale up.
A practical way to address this is through cross-training. For example, a crew trained in both exterior and interior work can pivot to indoor tasks like prep, staging, or finishing when outdoor work halts. This strategy, often called skill chaining, ensures that workers are never completely idle and helps smooth out bottlenecks when demand spikes [3][4]. The idea isn’t to make everyone a jack-of-all-trades but to create enough overlap in skills to keep things moving when plans change.
Another useful tactic is dynamic staffing. Instead of hiring for an average workload, contractors can align staffing levels with the delays between permit approvals and actual project starts [1]. This approach reduces the chaos of sudden surges and prevents burnout, ensuring teams are ready when the workload hits its peak.
Hospitality: Tackling Weekend, Holiday, and Event Surges
For hospitality businesses, demand swings are often predictable but intense. Weekends, holidays, and local events bring in crowds, while weekdays or off-seasons leave staff underused. The challenge is to maintain top-notch service during busy times without overspending on labor during slower periods.
Service inventory is a game-changer here. By completing certain tasks ahead of time - like food prep, room staging, or administrative work - hospitality operators can better handle demand spikes [5].
"Service inventories allow firms to buffer their resources from the variability of demand and reap benefits from economies of scale while also providing customers with faster response times." – Sunil Chopra and Martin A. Lariviere [5]
Shifting the push-pull boundary is another effective strategy. For instance, restaurants could prep ingredients earlier in the week to ease the pressure during Friday dinner rushes. Similarly, hotels could stage rooms during quiet hours to ensure quicker turnovers on busy weekends. This proactive approach reduces customer wait times and keeps staff productive during slower periods.
Cross-training also plays a big role. A server who can pitch in as a kitchen assistant during a rush, or a prep cook who can help at the front desk, creates a more adaptable team. Using methods like Pointwise Stationary Approximation (PSA) to calculate staffing needs based on real-time demand instead of daily averages ensures enough coverage during peak hours without overstaffing during lulls [1].
While hospitality businesses focus on proactive planning and cross-training, owner-led businesses face a different kind of challenge.
Owner-Led Businesses: The Burden of Absorbing Demand
In many small businesses, the owner often becomes the go-to solution for handling demand fluctuations. While this might work in the short term, it’s not sustainable. Constantly stepping in to manage crises or rush orders can lead to burnout and stifle growth.
The key is to shift from personal intervention to structural solutions. For example, service inventory can help owners tackle tasks during slow periods, avoiding a pile-up when things get busy. An HVAC contractor, for instance, could use downtime to handle permits, organize materials, and schedule crews, leaving only the actual installations for peak periods [5]. This repositioning of the push-pull boundary reduces the pressure of real-time demand.
Owners should also avoid planning based on average demand. Instead, they should prepare for actual spikes by implementing chaining strategies. These strategies allow resources and processes to shift between different tasks, ensuring that no single area becomes a bottleneck [4]. This not only protects the owner’s time but also creates a more resilient system.
The ultimate goal is to build an operating model that absorbs variability structurally, freeing owners to focus on growth rather than constantly putting out fires. These examples highlight how SMBs can navigate demand swings effectively, setting the stage for actionable design strategies.
Design Principles for Operating Models That Handle Variability
What separates an operating model that falters under pressure from one that thrives is the intentional design work done ahead of time. These thoughtful choices transform variability from a disruptive force into something the system can handle effectively.
Buffers Protect Flow and Margin
Buffers aren’t about overstaffing - they’re about safeguarding the most critical parts of your system. Eliyahu Goldratt’s Theory of Constraints emphasizes that buffers should protect bottlenecks to ensure they don’t run out of work during high-demand periods or cause delays that ripple through the system. The key is placing them where congestion actually occurs.
Take a restaurant, for example. While customer arrivals might peak at 7:00 PM, the kitchen might not hit its breaking point until 7:45 PM when orders pile up. If staffing is based solely on the arrival peak, the real bottleneck - the kitchen - remains unprotected.
The size of the buffer matters less than its strategic placement. As Xavier de Groote puts it:
"Flexibility is characterized as a hedge against diversity" [6].
Buffers should address two types of flexibility: volume flexibility (handling overall demand shifts) and mix flexibility (handling changes in the types of work). For instance, a contractor might need different buffers for delays caused by weather versus those caused by permit issues. Each scenario calls for a tailored response.
Decoupling Reduces System Strain
Decoupling separates when demand arrives from when it’s fulfilled, creating a buffer between the two. This shift aligns supply with demand, avoiding costly mismatches. One way to achieve this is through service inventory - completing work or preparations in advance so you’re ready when demand spikes.
For example, an HVAC company that stages equipment, secures permits, and schedules crews during slower periods can respond faster when emergency calls flood in. Similarly, a catering business that preps ingredients earlier in the week can handle a surge of Friday events without overwhelming the kitchen.
The goal is to identify which steps can be completed ahead of time and shift them forward in the process. This not only smooths out peaks but also keeps workflows steady.
Work Classification Improves Resource Allocation
Not all tasks behave the same under pressure. By classifying work based on its demand profile - such as predictable vs. volatile, short vs. long duration, or instant-response vs. flexible-deadline - you can allocate resources more effectively [1][5].
Standard queueing theory, which assumes steady demand, often falls short in environments with fluctuating workloads [2]. Models that account for time-varying demand patterns allow for smarter staffing decisions.
For example, a property management company might separate emergency repairs (instant-response) from routine maintenance (flexible-response). This lets crews spread out tasks over the week, balancing urgent needs with less time-sensitive work. Similarly, understanding how different task types interact can help avoid unexpected profit losses [7]. These classifications lay the groundwork for building capacity that adapts to demand swings.
Adaptive Capacity Beats Fixed Efficiency
When demand fluctuates, a fixed efficiency model - designed for averages - can’t keep up. Peaks overwhelm the system, while lulls waste resources. Adaptive capacity models, on the other hand, rely on strategies like cross-training, flexible vendor agreements, and surge policies to maintain performance across varying conditions [2].
As Linda V. Green, Peter J. Kolesar, and Ward Whitt explain:
"Standard queueing theory focuses on the long-run steady-state behavior of stationary models"
and doesn’t account for the predictable daily swings in demand [2]. Adaptive models use time-based performance metrics to proactively adjust capacity. For small business owners, this could mean implementing structural solutions like skill chaining, pre-completed service inventory, or tiered response times to handle demand spikes without constant intervention.
Conclusion: Stability Through Design, Not Effort
Uneven demand is here to stay. Whether it’s a sudden weather delay or a weekend rush, fluctuations are part of the landscape. The real question is: Can your operating model handle these shifts without overloading your team or draining your profits?
Experts like Deming, Goldratt, and Reinertsen emphasize that the answer lies in structural design, not reactive effort. Instead of planning for an imaginary "average", businesses need to design systems that can handle both the highs and lows effectively. This has been the central focus of our discussion.
Key Takeaways
Relying on average-based models sets you up for failure when real-world demand fluctuates. The solution isn’t working harder - it’s designing smarter. Here’s how:
Buffers: These act as shock absorbers, maintaining flow even when congestion spikes.
Decoupling: Aligns the timing of demand with fulfillment, reducing bottlenecks.
Work Classification: Keeps unpredictable tasks from overwhelming routine operations.
Adaptive Capacity: Prioritizes flexibility over rigid efficiency, ensuring performance holds steady during shifts in demand.
"Process flexibility, whereby a production facility can produce multiple products, is a critical design consideration in multiproduct supply chains facing uncertain demand" [4].
This idea isn’t limited to supply chains - it applies to everything from construction teams dealing with weather surges to restaurants managing weekend crowds. The goal is clear: Create systems that absorb demand swings through design, not by overworking people.
Looking Ahead: Designing for Volatility
As market conditions grow more unpredictable, businesses must adapt their operating models to thrive. Industries across the board are feeling the effects of rising seasonality and volatility. Companies that design for these realities will protect their margins, reduce employee burnout, and sustain growth. Meanwhile, those clinging to outdated approaches will struggle with the constant cycle of overload and inefficiency.
Traditional queueing theory, which assumes steady-state conditions, doesn’t cut it in today’s world. Instead, adopting nonstationary models - which account for time-based demand shifts - enables smarter resource allocation during peaks and cost control during quieter periods [1][2]. It’s a forward-thinking approach that positions businesses to navigate volatility with confidence.
FAQs
How do I find my real bottleneck during peaks?
To find the real bottleneck during peak times, pay attention to where queues, delays, or backlogs start to form. These visible signs show where demand outpaces capacity, revealing weak spots in your system. Observing these areas helps you identify constraints, echoing Goldratt’s emphasis on protecting flow and managing buffers. Instead of relying only on forecasts, watch how your system performs under stress to uncover the actual limiting factor.
What work can I pre-do to reduce peak-time chaos?
To ease the stress of peak times, aim to balance workloads and establish buffers that can absorb fluctuations in demand. One effective strategy is to build up inventory or semi-finished goods ahead of time, ensuring you're ready to meet higher demand without overloading resources. For instance, creating response-time and lead-time buffers allows you to adjust capacity in advance. A practical example: in the hospitality industry, prepping stock for busy weekends or holiday rushes helps maintain smooth operations and prevents last-minute pressure.
How can I add flexibility without overstaffing?
To create more adaptability without overloading your workforce, consider crafting operational models that include capacity buffers and flexible resource allocation. This could involve approaches like flexible scheduling, cross-training employees to handle multiple roles, and utilizing resources that serve multiple functions. These methods allow you to handle shifts in demand more smoothly.
Another useful tactic is introducing decoupling points at critical stages. These act as buffers to absorb variability, minimizing the need for constant changes in staffing levels. Together, these strategies help maintain stability, manage demand fluctuations, and prevent inefficiencies or employee burnout.
