How to prepare outsourcing for seasonal volume peaks: forecasting, capacity and a surge plan
A volume peak does not have to cause delays. Combine rolling forecasts, capacity reserve, priorities and cost control.
Document and mailing volumes rarely grow evenly. Month-end activity, annual statements, campaigns, price changes or regulatory duties can multiply demand within days. If client and provider plan around a monthly average, a service that looks sufficient develops a queue exactly when deadlines matter most.
Peak planning is different from business continuity. A disruption is unexpected; a seasonal increase can be forecast and prepared. The current UK Government Project Delivery Teal Book recommends balancing resource supply and demand, allowing for sourcing lead times, using rolling forecasts, scenarios and buffers. ISO 37500 emphasises flexibility as business needs change while outsourcing risk remains governed.
A monthly average hides the real constraint
A process averaging 40,000 items a month may receive 2,000 every day or 18,000 in two days. Planning therefore needs daily and hourly volume, batch count, exception rate, formats, page count, personalisation, dispatch deadlines and input readiness.
Start with 12-24 months of history. Mark repeatable peaks, one-off events and days where delay began in input, approval or production. If the history is missing, use four weeks of measurement and record every assumption openly.
Forecast three scenarios
- Committed volume: confirmed work tied to a specific delivery date.
- Best estimate: the most likely demand based on history, commercial plans and current information.
- Stress scenario: a realistic upper band used to test reserve capacity and escalation triggers.
Use a rolling forecast. The next week may be specified by day and batch, the next month as a range, and the next quarter by season and demand source. Regular updates and confidence levels are more useful than false precision.
Measure end-to-end capacity
Capacity is not the speed of the fastest machine. File validation, personalisation, print, quality checks, inserting, envelope supply, carrier collection or client sample approval may become the constraint. Measure correct output per hour across the complete route and include changeover time between formats.
| Capacity layer | Purpose | Example trigger |
|---|---|---|
| Base capacity | Normal predictable volume | Regular staffing, machines and collection windows |
| Planned reserve | A seasonal peak notified in advance | Extra shift, earlier materials order, larger collection |
| Contingency reserve | Forecast overrun or unavailable resource | Alternative line, approved subcontractor or resequenced work |
Each layer needs an owner, mobilisation time, maximum duration and cost. “Scalable” is not a capacity commitment until it is expressed in units, time and conditions.
Set notice periods and trigger levels
The service schedule should state when the client supplies a forecast and when it becomes binding. A workable pattern might be an indicative forecast four weeks ahead, confirmation ten working days ahead and a final file two days before production. Trigger levels can activate specific actions:
- 80% of available capacity - confirm data and materials;
- 90% - reserve an additional shift and transport;
- 100% - decide on batch split, date change or reserve use;
- above the agreed maximum - approve cost and adjusted SLA separately.
The exact thresholds depend on the process. Their purpose is to create a decision before backlog appears.
Inputs must be ready before capacity is activated
Extra operators and equipment do not help when the file is invalid, the template is unapproved or address batches arrive without identification. Define a ready input: data format, mandatory fields, document version, approved sample, record count, acceptance deadline and decision owner.
Agree a cut-off time. Work arriving later moves to the next window or needs an explicit priority decision. This protects both SLA and quality.
Not every item has the same priority
During a surge, one common queue can delay legally time-bound correspondence behind informational material. Define critical, time-bound, standard and deferrable classes. For each, set turnaround, decision owner and evidence.
Connect priorities to the RACI and escalation route. A provider should not independently decide which client obligation matters more.
Materials and transport are part of capacity
Paper, envelopes, toner, labels, transport containers and collection windows may constrain the process before production does. For long-lead items, set minimum stock, reorder point and an approved substitute. Segregate client materials so another peak does not consume them.
Check the daily carrier limit and availability of extra collections. Finished output waiting until tomorrow is not a completed dispatch.
Pricing should separate readiness from usage
Reserved capacity costs money even when only partly used. Options include a readiness fee, a higher unit rate above the base band, or a hybrid model combining a minimum commitment with actual volume. Agree the rules before the season, not after delivery.
Compare the options with our guide to fixed, unit and hybrid pricing. Include the cost of retaining internal reserve staff, equipment and space in the alternative.
Run a load test
Before the peak, process a representative batch at target speed. Test data ingestion, personalisation, print, inserting, control, reporting and transport hand-off. Measure correct output rate, rework, changeover, downtime and evidence completeness.
A test without criteria does not answer the question. Set minimum throughput, maximum error, tolerated backlog and stop point in advance. If the test exposes a shortfall, there is still time to alter schedule, format or batch split.
A peak-period dashboard
Report more frequently during a surge than in a normal month. Useful measures include forecast versus received volume, capacity utilisation, first-time-right output, queue size, oldest-item age, days of material cover, exceptions and plan variance. Connect them to the main service KPIs.
After the season, compare forecast accuracy, reserved-capacity cost, overtime, delay cost and exception causes. The findings should update the next plan rather than merely enter an archive.
Prepare in six steps
- Collect volume history and mark peak days.
- Build three scenarios and a rolling forecast.
- Calculate end-to-end capacity, materials and transport.
- Agree thresholds, priorities, notice and pricing.
- Run a load test with acceptance criteria.
- Use short reporting cycles and complete a post-peak review.
If the seasonal increase affects correspondence, printing, inserting or bulk mailing, BackOffice Outsourcing can help define volumes, deadlines and delivery options. A reliable quotation begins with a demand profile, not one average number.
Sources
- UK Government Commercial Function, The Sourcing Playbook, updated 17 September 2026
- UK Government Project Delivery, The Teal Book, Chapter 28: Resource management, updated 1 July 2026
- UK Cabinet Office, The Contract Management Playbook, 25 March 2026
- ISO 37500:2014, Guidance on outsourcing
Frequently asked questions
How much reserve capacity should a provider maintain?
There is no universal percentage. Reserve should reflect historical variance, mobilisation lead time, process priority and delay cost. Define it in units and time, not only as a percentage.
Must the client guarantee forecast volume?
Not always. An indicative forecast can be separated from committed volume. Reserving a defined capacity band may, however, require a minimum fee or advance confirmation.
How is a seasonal peak different from a disruption?
A peak is forecast demand growth requiring a capacity plan. A disruption removes systems or resources and activates continuity measures. The two plans should be connected.
When should a load test be run?
After data, materials and configuration are defined, but early enough to fix constraints. The test should reproduce target speed and the full process route.