Piglet In Bed: Advertising Forecasting For Sales Periods
What are Statistical Models For Forecasting
An accurate forecasting tool/model enables you to both forward plan (“I need to do X to achieve Y KPI”), and understand performance retrospectively.
The best models can accurately predict incremental revenue from marketing, considering demand, discounting, spend, seasonality etc.
Statistical models for forecasting allow you to:
- Plan budgets: get a concrete answer to how much you can spend & scale revenue, with various contextual factors like seasonality, discount % etc, in order to achieve X KPI.
- Understand how much you might need to raise from investors and give deeper clarity on how it’s allocated.
- Know and confidently assert whether performance has been good given the day of week, time of year, appeal of promotions & sales.
- Invest the right budget splits across advertising channels to achieve X KPI (profitability, revenue growth etc) when increasing overall spend, considering diminishing returns.
Forecasting for Piglet in Bed
Piglet wanted to plan how much they revenue they could drive with spend in Q4, dependent on the level of discounting they offer over Black Friday.
We provided them a growth model which predicts marketing performance based on time of year and the level of discount. This enabled them to effectively plan the optimal budget pacing for the Black Friday sale and get ahead on the required stock with a high degree of accuracy. The model also enabled Piglet to objectively assess whether they were improving the effectiveness of their marketing activity across sale and BAU periods alike, even outside of Q4.