Events, promotions and price

Events and prices, modelled before they happen.

Horizon clears holidays, stock-outs and promotions out of your history automatically, then estimates how each promotion, price change or launch will move volume, revenue and mix. Once approved, the impact goes straight into the forecast, with the assumption still visible.

Start with Horizon InstantBook a demo
EventsHistoryPrice
Deep Sensing
Example · SKU-2207 at Retailer North
Promotion
2+1 offer · Retailer NorthWeeks 14–15 · 3 similar promotions in event memory
Promotion+38% volumeDip after, week 16W10W14W15W19
BaselineWith the event
Volume
+4,200 units
Revenue
+€31k
Mix
2 related items
Impact into the forecast, with the assumption visibleApproveIn the forecast
3event types

promotions, price changes and launches, each modelled as an event

PromotionPrice changeLaunch
Before

every event modelled before it happens, in volume, revenue and mix

VolumeRevenueMix
Auto

holidays, stock-outs and promotions cleared out of the history

HolidaysStock-outsPromotions
Event memoryEvery measured effect informs the next event1Plan the eventPromotion, price change or launch2Model the impactVolume, revenue and mix3ApproveThe key calls reach planners4Into the forecastAssumption still visible5MeasureWhat really happenedFrom plan to forecast, and back into memory
Eventfrom plan to memory
Plan the eventPromotion, price change or launch
Model the impactVolume, revenue and mix
ApproveThe key calls reach planners
Into the forecastAssumption still visible
MeasureWhat really happened
Event memoryEvery measured effect informs the next event

What you can do

Cleaner history, sharper events.

History cleanup

One-off spikes out of the history

  • Holidays, stock-outs and promotions cleared out of historical data automatically, so one-off spikes don’t distort the next forecast.
  • Noise from events, promotions and price changes cleaned before it reaches the demand forecast.
  • Stock-out periods recognised, so lost sales aren’t read as low demand.
History · SKU-2207
Example · 52 weeks
PromotionStock-outHolidayRaw historyCleaned historyHISTORY CLEANUP · SKU-22073 one-off effects cleared before forecastingStock-out weeks no longer read as low demand

Impact analysis

Every event modelled before it happens

  • Horizon estimates how a promotion, price change or launch will move volume, revenue and mix before it happens.
  • Causal models handle drivers such as launches and prices, and Deep Sensing’s event memory recalls how similar events played out.
  • Once approved, the event’s impact goes straight into the forecast, with the assumption still visible.
Event impact · weeks 14–15
Example
Promotion · 2+1 offer at Retailer NorthWeeks 14–15 · SKU-2207 and 2 related itemsModellingVolume+4,200 units+38% during the promotionRevenue+€31kAfter the discountMix2 related itemsSmall halo on SKU-2208Deep Sensing · Event memory3 similar promotions at this customer recalled, with their measured effectApproved · impact in the forecast, assumption visibleIn forecast

Price impact

What a price change does to volume

  • Estimate how a price change will move volume, revenue and mix before it goes live.
  • Price changes modelled as events, next to promotions and launches, on the same forecast.
  • Approved price impacts flow into the forecast and on to supply planning.
Price change · SKU-2207
Example · from week 20
VOLUME AGAINST PRICEToday €2.40+5% · €2.52Price →VolumeVolume−3.1%Revenue+1.8%MixShift to 2 value packsPrice change from week 20, modelled as an eventReview

Learn more

Learn more about events and price impact.

What is demand sensing?How to improve forecast accuracyHow does AI improve demand planning?What is demand planning?Demand planning

FAQ

Events and price impact, answered.

What is event planning in demand planning?

Event planning captures the things that move demand away from its normal pattern, such as promotions, price changes and launches, and models their effect on the forecast before they happen, instead of correcting the plan afterwards.

Why clean promotions and stock-outs out of the sales history?

Because one-off spikes and dips distort the next forecast. A promotion peak would inflate the baseline, and a stock-out would make lost sales look like low demand. Horizon clears holidays, stock-outs and promotions out of historical data automatically.

How does Horizon model the impact of an event?

Horizon estimates how a promotion, price change or launch will move volume, revenue and mix before it happens. Causal models handle drivers such as launches and prices, and Deep Sensing’s event memory keeps every earlier event with its measured effect, so similar events inform the next one.

What happens once an event is approved?

The event’s impact goes straight into the forecast, with the assumption still visible, so everyone can see why the forecast moved.

Can Horizon model price changes?

Yes. Price changes are modelled as events, next to promotions and launches, with their effect on volume, revenue and mix.

How fast can we start?

Start with Horizon Instant: your ERP connected and real demand planning on your own data, up and running in less than 24 hours.

See your events modelled on your own data.

Horizon Instant: up and running in less than 24 hours.

Start with Horizon InstantBook a demo