The Importance of Forecasting
The Challenge

The most important data you need to allocate your marketing budget optimally is how well each campaign will perform at different spending levels. If you extrapolate from past performance, you’re ignoring diminishing returns, seasonality, and other factors that may affect future performance. You’re basically driving in while looking in the rearview mirror.

Ad platforms forecasts are incomplete

Google and Meta provide forecasts for how your campaigns will perform at different levels of spend. These visualizations are helpful but incomplete. For one, they are only available as charts in the UI, so it’s hard to build them into a model. But more importantly, neither platform offers forecasts for all of the campaigns.

Of course, to allocate resources accurately across your entire program, you also need to combine the forecasts from individual ad platforms.

Marin AI + ad platform forecasts = 🔥🔥🔥

To provide comprehensive coverage, Marin combines the publisher's forecasts with our proprietary models to simulate future results for each campaign at different spending levels. Our models understand the impressions share and/or audience reach of each campaign to account for campaigns that are at or nearing their maximum potential spend.

We automatically identify day-of-week, day-of-month, and day-of-year patterns in your campaign volumes and conversion rates. Our advanced forecasting AI analyzes the impact of events from your Marketing Calendar and incorporates future events into the simulation. 

Simulations and Algorithms

The output of our algorithms is a simulation for each campaign that shows the change in spend and conversions or conversion value as the campaign target is increased or decreased. This process is repeated for each day in the forecast period. In this chart, the vertical line at 0 represents the campaign's current performance. Campaigns with great sensitivity to changes will have steeper curves.

The next step to generating the forecast for your entire Strategy is to find the optimal allocation to each campaign. To simplify a complex process, we find the best use of each incremental dollar of spend until we have achieved your target. This results in a new recommended campaign target and budget ready to be sent to the ad platforms.

Our goal is to give marketers as much control as possible so we offer a range of rules that can constrain the generated recommendations. The most common of these is to limit the daily change in an attempt to avoid triggering learning mode.

When the optimal allocation is different than the constrained outcome, we show both on the chart, with the solid dot representing the constrained outcome and the dotted line representing the unconstrained. Both values are available for review in the campaigns grid. 

Before and After

The impact of the optimal allocation is then presented in our “before and after” charts. The left side of these charts represents the “status quo” or what we forecast your results will be with no changes. The right side is the results we forecast using the Marin recommendations.

Once you are comfortable with the recommendations, you can set the Strategy to push changes to the publisher. Marin will then be continually monitoring performance and adjusting the targets and budgets to hit your goals.