WFM guideForecasting

Contact centre seasonality

Q: How do you forecast seasonal contact centre volume?

Seasonal contact centre volume is forecast by decomposing historical volume into its components: trend (long-run growth or decline), seasonality (repeating annual pattern), day-of-week effect (Monday typically 20–30% above weekly average in most contact centres), and noise (random variation). Once decomposed, a seasonal index is calculated for each week of the year — expressing how much volume in that week differs from the annual average (e.g., a Christmas week seasonal index of 1.8 means that week runs at 180% of average weekly volume). The forecast is built by applying the seasonal indices to a trend projection and adjusting for any known events (product launches, regulatory deadlines, marketing campaigns) that will shift the pattern. Seasonal indices should be recalculated annually using the most recent two or three years of data. Using older data risks embedding historical patterns that no longer apply.

Q: How much extra staffing do contact centres need at Christmas?

The extra staffing requirement at Christmas varies significantly by sector and channel. Retail and e-commerce contact centres typically see volume 80–200% above the annual weekly average in the peak 2–3 weeks before Christmas and 150–300% above average in the first week of January (returns, complaints, delivery queries). Financial services (cards, personal loans, insurance) typically see 20–40% above average in December and January. Utilities see a modest seasonal peak in winter (heating queries, billing disputes after December bills) of 10–25% above average. For retail operations, this means staffing plans must flex from a baseline of (for example) 100 agents to 180–300 agents in peak weeks — which is not achievable through overtime alone and requires advance recruitment of temporary staff, typically 6–10 weeks before peak.

Seasonality is the most predictable component of contact volume — and the most disruptive when ignored. A retail contact centre that treats December as an average month will face a queue that collapses service level for six weeks. Seasonal planning requires decomposing historical volume into its components, calculating seasonal indices, and building staffing flexibility that can move with the forecast.

Volume decomposition

Weekly volume = Trend × Seasonality × Day-of-week effect × Noise

Trend

Long-run growth/decline in volume

+3% YoY from customer base growth. Remove before calculating seasonal indices.

Seasonality

Annual repeating pattern

Week 51 (Christmas week) = 1.8× annual average. Calculated from 2–3 years of history.

Day-of-week

Within-week variation

Monday = 125–130% of weekly daily average. Saturday = 60–70% (for Mon–Sat operations).

Noise

Random variation

Weather events, marketing campaigns, competitor actions. Cannot be modelled — plan for it with intraday flexibility.

Seasonal index calculation: For each week of the year, calculate the ratio of that week's volume to the annual weekly average, after removing trend. Average this ratio across 2–3 years to get the seasonal index for each week. A seasonal index of 1.8 for week 51 means: forecast = (trend-adjusted annual volume ÷ 52) × 1.8 × day-of-week factor.

Seasonal patterns by sector

Sector	Peak period	Peak index	Trough	Planning lead
Retail / e-commerce	Pre-Christmas (Nov–Dec) and January (returns)	180–300% of annual avg	February–March (60–75% of annual avg)	6–10 weeks before peak. Temp recruitment lead time is the binding constraint.
Insurance (home / motor / travel)	January (renewal season), June–August (travel)	120–150% of annual avg	October–November (75–85% of annual avg)	4–8 weeks. Mostly overtime and flexible contracts rather than temp hires.
Utilities (energy / water)	October–February (winter billing peak)	115–135% of annual avg	May–August (70–85% of annual avg)	4–6 weeks. Winter staffing often met through annual leave restriction from September.
Financial services (banking / cards)	January (post-holiday statements), April (tax year end)	115–130% of annual avg	June–August (80–90% of annual avg)	4–6 weeks. Most managed through overtime and annual leave coordination.
Healthcare / NHS	January–March (winter pressures), September (back to school)	120–160% of annual avg	May–August (70–80% of annual avg)	6–12 weeks. NHS staffing is constrained by training requirements — temp use is limited; relies on overtime and cross-skilling.
Public sector / local authority	April (new financial year), January (council tax), September–October	115–130% of annual avg	July–August (70–80% of annual avg)	6–8 weeks. Annual leave in July–August reduces staff available during trough — plan accordingly to avoid understaffing during the trough period.

Christmas spike: the mechanics

For retail and e-commerce operations, the Christmas period is not a single spike but a four-phase event with different WFM requirements:

Phase 1: Pre-Christmas (weeks 47–50)

130–180% of avg

Gift purchase queries, delivery tracking, product availability. AHT close to average — most contacts are simple transactional queries. Requires significant extra headcount but manageable through overtime and temp staff who have been trained for 4–6 weeks.

Phase 2: Peak despatch / Black Friday (week 47–48)

180–250% of avg

Highest volume of the year. Order confirmation, delivery chase, click-and-collect queries. Often combined with promotional campaign handling. This is when all temp hires must be fully trained and operational.

Phase 3: Christmas closure / bank holidays (weeks 51–52)

50–80% (closed periods) / 200%+ (reopening spike)

Volume collapses during closure days. Reopening spike (26 Dec–2 Jan) is often the highest single-day volume of the year — returns, complaints, gift activation, delivery failures. Staff returning from Christmas leave face the highest occupancy of the year.

Phase 4: January returns (week 1–3 of new year)

150–300% of avg

Returns, refund processing, complaints about Christmas gifts, delivery damage claims. January week 1 is often worse than any pre-Christmas week. Temp staff must be retained through mid-January. Many operations underplan January relative to December.

Flexible staffing mechanisms for seasonal peaks

Temporary / agency staff

CapacityCan scale 50–200% above baseline headcount

Lead time6–10 weeks (recruitment, training, systems access)

CostAgency rate typically 25–40% above permanent FTE hourly cost (including agency margin). For a 12-week peak period, net cost premium is typically 8–12% of annual wage bill.

RiskQuality risk during first 2–3 weeks on floor. AHT for temp staff is typically 15–25% above experienced-agent average. Plan for this in peak headcount calculations.

Overtime (existing staff)

Capacity10–20% above baseline working hours (beyond this, burnout and quality risk accelerate)

Lead time1–2 weeks' notice for scheduling

CostTypically 1.25–1.5× standard hourly rate. No training cost. No quality ramp period.

RiskCannot sustain over 6–8 weeks. Occupancy at 88–92%+ for extended periods triggers the 3–6 month attrition lag. Overtime covers short peaks well; it cannot substitute for temp recruitment in a 12-week peak.

Annual leave embargo

CapacityEquivalent to 5–10% headcount increase (bringing back AL absent staff)

Lead timeRequires policy and 8–12 weeks advance notice in employment contracts

CostNo direct cost. Annual leave is accrued and must be taken eventually — embargoed leave must be permitted in the trough. Risk of trough understaffing if not managed.

RiskLegal risk if embargo is imposed without contractual basis. Employment contracts must permit seasonal leave restrictions. ACAS guidance requires reasonable notice. Cannot be used repeatedly without affecting agent morale and attrition.

Cross-trained back-office agents

Capacity5–15% headcount increase from internal flex pool

Lead timeCross-training must complete 4–6 weeks before peak. Cannot train and use simultaneously.

CostTraining cost (typically 2–4 weeks of off-floor time per agent). No ongoing cost premium once trained.

RiskBack-office work is deferred, not eliminated. If back-office flex agents are used for 8 weeks of voice handling, back-office SLAs will slip unless there is also a back-office trough. Requires coordinated capacity planning across both functions.

Seasonality questions

How do you forecast seasonal contact centre volume?

Decompose historical volume into trend (long-run growth), seasonality (annual repeating pattern), day-of-week effect, and noise. Calculate a seasonal index for each week of the year (ratio of that week's volume to annual weekly average, after detrending). Apply seasonal indices to a trend projection. Recalculate indices annually using the most recent 2–3 years of data — using older data risks embedding patterns that no longer apply.

How much extra staffing do contact centres need at Christmas?

Retail/e-commerce: 80–200% above average in the pre-Christmas peak, 150–300% above average in January returns. Financial services: 20–40% above average. Utilities: 10–25% above average. For retail, this means planning temp recruitment 6–10 weeks before peak. January week 1 is often the worst week of the year — most operations underplan January relative to December.