Event-driven volume analysis
Summary
- Situation
- Contact volumes rise and fall with the sporting calendar, which makes the base volume for upcoming weeks hard to estimate.
- What I did
- Starting in October 2026: an Excel tool that analyses past volumes around sporting events to estimate base volumes for the weeks ahead.
- Result
- In progress. Once it's live, I'll measure its estimates against actual volumes and the time it saves forecasters.
Before
- Pull past volumes by hand
- Check the sporting calendar separately
- Estimate base volumes manually
After
- Refresh past volumes and the event calendar
- Review the tool's base-volume estimates
- Hand them to the forecasters
- Role
- Sole developer
- Period
- Oct 2026 – Present
- Team
- Built for the forecasting team
- Stack
- Excel · Power Query · Formulas · Pivot tables
- Sheet
- P-01 · rev. Sep 2026
The question
In iGaming, customer contact follows the sporting calendar: big fixtures bring spikes, and quiet weeks sit well below them. That makes it hard to know the base volume for the weeks ahead, the level to expect before event effects are added.
Can past volumes, matched to the events around them, give forecasters a reliable starting point for upcoming weeks?
Data & constraints
- Inputs: contact volumes for voice, messaging and support since 2025, and the calendar of upcoming sporting events that the forecasters keep up to date.
- Tools: Excel (Power Query, formulas, pivot tables).
- Role of the tool: it supports forecasters and doesn’t replace their forecast.
- Confidentiality: no volumes or event data are shown here.
Approach
The build starts in October 2026. The plan:
- Power Query brings in past volumes and the event calendar and lines them up by date.
- Pivot tables and formulas compare volumes on event and non-event days to separate the base level from event-driven spikes, then adjust the estimates in line with the historical increase in volumes.
- The output is a base-volume estimate for each upcoming week that forecasters can start from.
Result
In progress, so no results are claimed yet. Once it’s live I’ll measure:
- how close its estimates are to actual volumes (for example, the average percentage error per week)
- how much time it saves forecasters each cycle
Limitations & next steps
- It assumes upcoming events behave like similar past events. New competitions or unusual fixtures won’t be covered well.
- History starts in 2025, so there’s little data for events that happen once a year or less.
- Excel limits how much history it can handle comfortably.
- Next: build it, measure it against actual volumes, and publish the results here.