🗺️ Plans
The Plans feature allows you to build, view, forecast, and optimize marketing plans, all within the Recast platform. These comprehensive Plans allow you to specify your marketing budgets, spikes (promotions and holidays), non-spend variables, and contextual variables. This makes it easier than ever to plan, forecast, and optimize marketing budgets using Recast.
The Plans Tab
The Plans tab is the homepage for your team's Plans. Here you can see all the Plans that have been created, filter to current or future Plans, and build new plans (coming soon!).
The Default Plan
Your Plans page will always include a Recast created, default Plan. This Plan is automatically updated each time your model updates. The spend is predicted using Recast's "business as usual" forecasting approach that assumes your future spend will follow the same general pattern as your historical spend. This Plan can be used as a starting point for optimization and to give you a rough idea of forecasted revenue (or conversions). The Default Plan will cover the time period between your earliest last model date (if you have multiple models) and up to 730 days after that. The Default Plan will include all channels in any of your models. Spikes used in the Default Plan will match what is planned for the future in your models. Any contextual variables will carry forward their last value for all dates in the future.
On initial release, the Default Plan will be the only Plan available for use. New tools for building Plans are coming soon.
Viewing a Plan
Clicking "Open" on the Plans page takes us to a page to view a specific Plan. Here you can see the components of the Plan (spend, spikes context variables), as well as Recast’s forecast of performance for each KPI if you follow that Plan.
The header shows the Plan name, the time range it covers, the total budgeted amount, and the KPIs the Plan is compatible with. It also gives you the ability to export the Plan for use in other places.
To be compatible with a KPI a Plan must provide all spend, non-spend, contextual variables, and spikes necessary to forecast that KPI.
[Coming Soon] If a Plan is incompatible with a certain KPI, clicking the KPI will walk you through what information you need to provide in order to make it compatible.
Spend
The Plan Summary section shows a breakdown of the planned spend over time and in aggregate.
Spikes
The spikes tab allows you to see and control spikes that happen in your Plan. The Default Plan will always use the spikes present in the underlying models, and cannot be edited. Custom Plans (coming soon!) can use the model’s spikes or a custom set of spikes provided by the one building the Plan.
The spike calendar will show you what spikes were in the past, what are in the future, and if there are discrepancies between what’s planned in the statistical models vs. what’s planned in the Plan.
If a historical spike was in the Plan, but never added to the model, it will be crossed out to show that it was not used in the models.
If a historical spike was in the model, but not in the Plan, it will be added to the spike calendar with a special tag to show that it was not planned for.
If a spike is planned in the future in the models but is not part of the Plan, it will not be used for forecasting until the spike is within 30 days of the last modeled day in the model. This is because once a spike is within a month of happening, it can start to influence the historical dates in the model. At that point, we cannot eliminate the effect of the spike using the forecaster. If you want to remove the effect of a spike, you will need to have the model re-run after removing the spike from the model. Your data scientist should be able to help with this.
Only one spike can happen on a given day in a given model. You will be prevented from adding additional spikes if one is already present.
Non-spend Channels
The non-spend tab shows the planned values for any non-spend predictors in your models.
Contextual Variables
The contextual variables tab shows the assumptions we’re using for your contextual variables for the life of the Plan. By default, we use a “carry-forward” approach meaning we will assume the contextual variable stays flat at whatever the most recent value was. When building a new Plan you can instead provide different assumptions to use regarding the contextual variables.
The values provided here will be used when forecasting or optimizing the Plan.
Lower Funnel Channels
The Lower Funnel Channel Settings allows you to control how we treat your Lower Funnel channels by default when Forecasting and Optimizing.
Each channel has four options:
Uncapped (default) - this means Recast will treat your Lower Funnel channels as an outcome instead of an input. The amount of lower funnel spend will be predicted based on upper funnel activity. Spend will not be stored as part of the plan in this scenario.
Capped - Recast will treat your Lower Funnel channels as an outcome up to a certain limit and then forcibly prevent additional spend above the cap. The cap you provide is for the entire Plan period (past and future). If you provide a cap for a Plan that is already partially in the past, we will subtract the actuals from the cap and use the difference as the new cap for forecasting/optimizing the future. If actuals are more than the cap already, the cap will be set at 0 (no additional lower funnel spend allowed). If the Plan starts in the future, when we forecast by combining with the Default Plan, we will scale the Forecast cap up so that the cap during the Plan period is approximately what you indicated, and the cap prior to the Plan period is proportionally the same amount.
Off - This will turn the Lower Funnel spend in this channel off completely.
Provided - Recast will use a fixed daily amount for spend in this channel. There are two ways to provide an amount: (1) CSV upload, telling us exactly what you want to spend each day and (2) A total amount for the Plan period, which Recast will then take and spread over the period using our business as usual forecasting methodology. If the total amount you enter is less than the total that has already occurred since the Plan start date, the channel will be assumed to be off for the remainder of the Plan.
Once you save these settings, they will be used in all future Forecasts. Optimizations will use these settings by default unless they’re overridden.
Forecast Results
The Forecast Results section shows a forecast of each KPI if you follow this Plan. It includes all the results you’re used to from our Forecasting tool.
Below the forecast results, you’ll get a recommendations on how to change the Plan to increase the forecasted KPI.
Forecasting Plans that start in the future
If a Plan starts in the future, we need to make assumptions about what will happen between the most recent historical data and the Plan start date. This is because your spend leading up to the planned time window will affect the forecast in the planned time window (due to time shift effects). In order to automate this process, we merge your Plan with the Default Plan to obtain the full budget used in the Forecast. If your model has data through the end of May, and your Plan is for July, the following values will be used when forecasting:
Date Range | Data source |
|---|---|
< May 31st | Historical data in the model |
June 1 - June 30 | Default Plan |
July 1 - July 31 | Plan |
The Forecast will be trimmed to only show the expected KPI during the month of July. If you have KPIs supported by multiple models, we will follow the same approach, always using as much historical data as possible and only using Default Plan data when nothing else is available.
Re-running stale Forecasts
A Forecast can go stale either because the Plan is updated (spikes change, spend changes, etc) or because a new model is deployed to replace the old model. When either of these things happen a message will show inviting you to re-run the Forecast.
Clicking “Rerun Forecasts” will launch all new Forecasts that will replace the existing Forecasts.
Coming soon: A historical view of all Forecasts that have been run in connection with a Plan.
KPIs
Your Plan can be used to forecast and optimize specific KPIs. There are two types of KPIs in Recast:
Single model KPIs - these KPIs have a direct relationship to a model built by Recast
Aggregate KPIs - these are weighted sums of single model KPIs that are aggregated into another KPI. For example, if you have an DTC revenue model and an Amazon revenue model, these can be summed (using weight 1) to an "All Revenue" aggregate KPI.
For your convenience, Recast has created a single model KPI for each of your models. If you desire to create aggregate KPIs that would be meaningful to your business, please reach out to your model building team. Once created, Plans will check for compatibility with the KPI and make forecasts for each KPI of interest.
Plan Compatibility Fixing
When a Plan is created it must be compatible with at least one KPI. Over time, however, changes to KPIs or the underlying model may lead to incompatibilities. The Compatibility UI and Fix Page help ensure your plans remain usable.
When a model underlying one of the KPIs included in your plan is changed, your Plan could become incompatible with the KPI. In this case, you will see the KPI in the compatibility banner turn from grey to red with a warning sign letting you know that you need to fix the compatibility issue before you can use your Plan accurately.
When a KPI is highlighted with this warning symbol, click into the KPI to enter the compatibility fixing page. The compatibility fixing page will display all the channels, spike groups and context variables that are causing the incompatibility. Click through each issue and select from the options to ‘fix’ them. Depending on the type of issue you will be provided with different options to solve the incompatibility.
Upper Funnel Channels:
Provide total amount for Recast to spend
Let Recast predict your total budget and spread
Upload spend
Lower Funnel Channels:
Uncapped - Recast predicts what you will spend
Capped - Set a maximum spend
Off - Set 0 spend in the channel during the plan time period
Provided - Set your total spend or upload a file with daily spend
Spikes:
Delete Spikes that are no longer in the model
Contextual Variables:
Carry forward the last value of the contextual variable
Upload daily contextual variable
Once you have resolved all the compatibility issues, your Plan will once again be compatible with your KPIs and ready to use.
Creating a Plan from an Optimization
Coming soon: Creating a Plan using Recast Predicted spend, Creating a Plan using an uploaded budget
Once you run an optimization and reach a budget you are ready to execute on, you can use the “Create Plan from Optimization” button and Recast will save your Optimization as a Plan with the recommended budget and contextual settings.
This allows you to:
Save your optimized budget and track the forecasted outcome of your optimization as you execute on your optimal budget.
Stay on top of incremental improvements to your optimized budget using the Recast automatic recommendations provided in your Plan which give updated suggested on how to further improve your optimized outcome.
Build an up-to-date home for your current optimized marketing budget in Recast.