Forecasting
Create forecasts that leaders can trust when decisions on service, cost and staffing matter
Forecasts shape decisions before customers feel the consequences. They influence staffing, budgets, schedules, recruitment, outsourcing, expected service performance, productivity targets and commercial planning.
When the forecast is trusted, leaders can make these decisions with confidence. When it is regularly challenged, planning becomes reactive and attention shifts from deciding what to do to debating whether the numbers are right.
Weak forecasting can lead to missed service levels, avoidable staffing costs, lost revenue opportunities and pressure on colleagues. Atlantic WFM helps organisations develop a more accurate, relevant and trusted view of future demand.
When forecasting support can help
- forecasts or their underlying assumptions are regularly challenged
- actual demand repeatedly falls above or below forecast
- forecast accuracy is unclear or measured at the wrong level
- the same forecast is used for decisions requiring different horizons or levels of detail
- WFM system forecasts are accepted without sufficient testing
- business changes and demand drivers are incorporated inconsistently
- adjustments depend heavily on the knowledge of one or two people
- leaders need a clearer view of uncertainty and the range of possible outcomes
It can also help capable forecasting teams test their current approach, evaluate alternative methods and strengthen the way forecasts support decisions.
Forecasting should start with the decision
A forecast is useful when it supports the decision it informs. A budget needs a long-range view. Capacity planning may require weekly forecasts by workload type. Scheduling needs interval-level demand. Real Time management needs an updated view of the rest of the day.
Each decision requires an appropriate time horizon, level of detail, update cycle and method. A single forecast is rarely sufficient. We begin by establishing which decisions the forecast needs to support, then design the forecasting approach around them.
What good forecasting requires
Reliable forecasting customer operations depends on a set of principles working together. The list below highlights some of the factors that matter most when forecasts are being used to support operational decisions:
Reliable forecasting begins with accurate data and a proper understanding of variation. Historical data needs to be clean enough to use, but it also needs careful interpretation. Trends, seasonality, level shifts, unusual events and changes in customer behaviour can all distort a forecast when they are not recognised. Useful charts help reveal patterns, anomalies and changes that summary measures alone may hide.
The forecast must also be aligned with the decision it is intended to support. Budgeting, capacity planning, scheduling and Real Time management require different time horizons, levels of detail and update cycles. A forecast that is suitable for one decision may be too broad, too late or too unstable for another.
Candidate methods should be tested against actual outcomes and compared with simple benchmarks. This establishes whether any new approach genuinely improves performance. Statistical and machine learning methods can add value, but complexity must justify the additional effort, explanation and maintenance.
Reliable forecasting also depends on skilled analyst judgement. Analysts need to select suitable methods, interpret results, challenge assumptions, understand uncertainty and explain what the forecast means for the decisions being made.
Cross-functional input ensures that forecasts reflect known changes outside the historical data. Marketing activity, digital change, operational decisions, products, finance, policy, customer behaviour and frontline insight can all alter future demand.
What clients receive from our work
We review the current forecasting process, including data, methods, assumptions, adjustments, accuracy measures and the decisions each forecast supports. This identifies where accuracy, credibility or usefulness is being lost.
Where improvements are needed, we test alternative approaches against historical outcomes and simple benchmarks. The chosen method reflects forecast performance, operational value, explainability and the organisation’s ability to maintain it.
Outcomes and benefits
Better forecasting gives leaders a stronger basis for operational and commercial decisions. It can help the organisation:
- increase confidence in forecasts, assumptions and planning decisions
- reduce the risk of overstaffing, understaffing, avoidable overtime and late recruitment
- strengthen capacity, outsourcing and budget decisions
- improve customer experience by aligning resource more closely with expected demand
- protect revenue where availability, response times or backlogs affect conversion and retention
- improve colleague experience by reducing avoidable workload pressure and disruptive replanning
- spend less time debating the forecast and more time deciding what action to take
Why Atlantic WFM?
Forecasting expertise demonstrated through research, competition and practice
Atlantic WFM combines practical forecasting experience across customer operations with formal study in Operational Research, statistical model testing and direct responsibility for operational forecasting teams.
Philip Stubbs has developed forecasting approaches for live environments where data quality, changing demand, system constraints and planning deadlines all matter. He has also delivered forecasting consultancy, training and decision-support modelling across a wide range of sectors.
Forecasting credentials and experience
- Master’s degree with Distinction in Operational Research from Lancaster University
- Published research on short-term forecasting in the Journal of the Operational Research Society
- Co-winner of the 2024 VN1 forecasting competition
- Third place in the 2025 VN2 forecasting competition
- Led forecasting and planning teams within large organisations
- Delivered forecasting consultancy across a wide range of sectors
- Devised forecasting training, including best practice work for The Forum
- Spoken at webinars and events on forecasting best practice
Client perspective
Claire Hill, Customer Operations Director
Frequently asked questions
What kind of forecasting do you support?
We primarily forecast demand and related operational indicators across contact centres, back-office operations, field teams and branch environments. We also support selected marketing analysis and forecasting, as well as product-demand forecasting.
Do you only help when forecast accuracy is poor?
No. Capable teams may also want to test their current approach, evaluate alternative methods, improve decision support or strengthen internal forecasting capability.
Will this add value if we already forecast in our WFM system?
Yes. System forecasts can be tested against actual demand, simple benchmarks and alternative models. This helps establish whether improvements are available through better data, configuration, assumptions or forecasting methods.
Do you use advanced models?
Yes, where they add sufficient value. The choice considers accuracy, operational benefit, explainability, implementation effort and ongoing maintenance.
Will our team be able to use the approach afterwards?
Yes. The work can include documentation, training, model explanation, process design and knowledge transfer so the approach can continue to be used and improved internally.
Request a conversation about Forecasting
If you need forecasts that leaders can trust, or want to understand where your current forecasting approach could be improved, we can discuss the issues you are seeing and what would help.