A marketing model regresses weekly sales (in thousands) on:

• $x_1 =$ advertising spend (in thousands of dollars)
• $x_2 =$ season (coded 0 = off-season, 1 = peak season)
• $x_3 = x_1 \times x_2$ (interaction term)

The fitted equation is: $\hat{y} = 20 + 5x_1 + 15x_2 + 8x_3$

During peak season ($x_2 = 1$), what is the predicted effect on sales for each additional $1000 spent on advertising?