Optimization Mechanics: Marketing Fundamentals
Differences in the timing of investment can lead to substantially different outcomes, even if the total investment level is constant. There are three drivers of this outcome.
Law of Diminishing Returns
The response to investment is non-linear. Doubling the investment in one week will result in less incremental return than spending that same amount over two weeks.
What About Minimum Thresholds for Breakthrough/Awareness?
We often get asked about minimum thresholds; wherein, after a certain level of investment, the channel becomes more effective. Keen optimizations typically do not support this theory, as the rate of return is always decreasing as you invest incrementally – take Facebook impressions for example:
Threshold Belief: You need 2M unique impressions per week breakthrough reach.
The threshold's premise is that impressions will be more valuable after 2M than before. We should expect 1.99Mth impression to have returned slightly less than the 2.01Mth impression.
The threshold belief implies a dependency between all previous impressions and incremental impressions; however, in the market, all ads are viewed independently of one other.
There's no reason to believe an impression's impact on a user's behavior is related to other activity in the market.
However, it may still be appropriate to reflect a desired flighting pattern for certain channels in your Plans, based on competing reach/awareness goals you may be juggling against financial goals.
Continuity Dependency
There is a dependence between investment in weeks – investment in the first week coming off a dark period will return less than the same investment after investing in a previous week. These last two points mean that more continuity is better than pulsing or event-based spending. Continuity performs better because Keen acknowledges that there are sales to be driven by investment across the year. The optimal investment does depend on seasonality. However, even with seasonal products, the demand is often not so seasonal that it dictates being dark at times of the year. Furthermore, because the response relationship is non-linear doubling the investment in one week will result in less in return than spending that same amount in two weeks. Finally, there is a dependence between investment in weeks – investment in the first week coming off a dark period will return less than the same investment after investing in a previous week. The confluence of all these factors translates to solution where investing more continuously over time is optimal.