“What impact did our Black Friday promo really have on revenue?”
It’s one of the most common questions we hear from our clients, and it’s why we built one of our favorite features at Recast: Spike Modeling.
Traditionally, MMMs have tried to capture the effects of holiday sales using dummy variables. A dummy variable is a simple switch in the model—turning “on” for a holiday like Black Friday and “off” for the rest of the year. While this approach might seem logical, it’s deeply flawed.
Dummy variables:
- Overcredit sales to the holiday event itself without accounting for the impact of marketing.
- Fail to account for consumer behaviors like delaying purchases until a sale (pull-forward effects) or stockpiling during the sale and skipping future purchases (pull-backward effects).
- Treat holidays as isolated events and ignore how marketing spend and promotions interact to drive results.
What does this look like in practice? A typical Black Friday sales pattern has three key phases:
- Pull-Forward Effects
Many consumers delay purchases in the weeks leading up to a sale like Black Friday, waiting for discounts.
For example, someone who usually buys on November 1 might wait until Black Friday to make the same purchase—potentially even buying in bulk. A dummy-variable approach captures the Black Friday spike but misses the pre-sale dip, leading to an inflated view of the holiday’s impact.
- Pull-Backward Effects
Stockpiling behavior is common during sales. Customers might buy more than they need, reducing purchases in the weeks or months after the event. Traditional models rarely account for this, overestimating the holiday’s contribution to overall revenue. - Marketing and Promotion Interplay
Dummy variables treat promotions as isolated events, ignoring how marketing activity drives awareness and conversions during the holiday period.
For example, a strong Black Friday performance might depend on a carefully timed TV campaign or Meta prospecting ads in the weeks prior. If your model doesn’t account for this interplay, it undervalues your marketing efforts.
If your MMM only looks at the sales spike during Black Friday, you’re missing the full picture.
You might think the promotion was a huge success when, in reality, you’ve just shifted purchases forward or backward in time without generating true incremental revenue.
Worse, traditional models often fail to credit marketing efforts and can lead you to flawed logical conclusions like “cut spend during holidays.”
Recast’s Spike Modeling solves this problem by capturing the full complexity of holiday promotions, providing a clearer picture of what’s really driving your results.
Here’s a bit more on how it works:
Why Spike Modeling Matters and How Businesses Use It
Spike Modeling isn’t just a “nice-to-have” feature –it’s critical for accurately measuring the impact of holiday sales. Here’s why it matters and how businesses leverage it:
1. Prevents Overcrediting Promotions
Spike Modeling makes sure you don’t overestimate the value of a holiday promotion by distinguishing incremental sales from shifted purchases. This prevents you from making costly mistakes like:
- Assuming all Black Friday sales are net new revenue.
- Giving away profit on products customers would have purchased anyway.
2. Drives Smarter Media Spend
Recast’s model captures how marketing spend interacts with promotions and helps businesses optimize their budget. For example:
- When should you ramp up Google Shopping ads to drive pre-sale awareness?
- How does Meta prospecting contribute to holiday sales spikes?
- What’s the ideal timing for TV campaigns in the weeks leading up to Black Friday?
By understanding these dynamics, businesses can allocate spend more effectively across channels and timeframes.
3. Supports Better Planning and Forecasting
Spike Modeling shouldn’t be used just for analyzing past performance – it should also help inform future strategy. We’ve seen businesses use it to:
- Forecast revenue for upcoming quarters, factoring in holiday and promotional effects.
- Decide whether to repeat or refine recurring promotions based on their incremental value.
4. Evaluates Customer Quality
Promotions don’t just drive sales—they also acquire customers. But are these customers as valuable as those acquired during non-promotional periods? Spike Modeling helps businesses figure out:
- The lifetime value (LTV) of customers acquired during holiday events.
- Whether promotional-driven customers are likely to stick around.
TLDR: How To Model Spikes
- Black Friday, Cyber Monday, and other holiday promotions are some of the most important—and challenging—events to measure.
- Traditional MMMs often fall short, overcrediting the promotion itself while undervaluing marketing and ignoring changes in consumer behavior.
- Recast’s Spike Modeling changes that. By capturing pull-forward and pull-backward effects, integrating marketing contributions, and providing flexible tools for any promotion, it reveals the true impact of holiday sales. Businesses can use these insights to optimize promotions, allocate spend effectively, and make smarter, more data-driven decisions.
