In the Reporter tab, the reports available for you to run are similar to what you have seen in the Insights tab (such as the Overview, Intercept Summary etc) however, you can make your own customizations to the input parameters.

When you are creating a new report, you can select to create a one-off report or automatically update the report every time the model updates.

Step 1: Navigate to the Reporter

Step 2: Select "Recurring Reports"

Step 3: Select your report type

Step 4: Fill in the information required for the report type.

To see your recurring reports, click the check box next to "Show only auto-run" in the report library. Recurring reports will be marked with the ♻️ symbol for easy identification.

What happens to the selected date range?

Depending on your selection of date range, the recurring report will automatically generate either at every model refresh or when the model has complete date for the time period. For date range selections that are relative to the last date of data on the model, each time the model refreshes, the dates will shift to be relative to the new last date of data.

For the selections "Last month, quarter or year", the report will automatically generate when the model has data for the entire period specified.

For the custom selection, if the new model date is n days in the future compared to the date of the last model refresh, then all report dates will be shifted by n days. The next time the model is refreshed, all the report dates will be shifted by the number of days between the last model refresh and the latest model refresh.

For the “Fixed to Rolling” selection, the start date will remain fixed while the end date will shift by the number of days between the new old model refresh and the new model refresh.

You can use this feature to set up reports that will automatically generate for a given time period so every time the model updates, you can log into your dashboard and view the latest numbers in your report.

This report shows the same information as the Overview page in the Insights tab. It contains four graphs.

In the Outcome & ROI box you will see high level metrics as well as the change in each of the reported metrics comparing this period to last period, and this period to the same time last year. The metrics are:

• Baseline Outcome: This is the estimated outcome gained in the period independent of your marketing spend from organic conversions and spikes (promotions and holidays).

• Paid Outcome: This is the estimated outcome gained in the period as a result of your paid media spend.

• Total Outcome: This is the total of your KPI observed in the period (not estimated)

• Blended ROI/CPA: This is the total return on investment calculated as Total Outcome/Spend (not estimated)

• Paid ROI/CPA: This is the estimated return on investment of your in-period paid media spend (regardless of when the outcome is realized).

The grey bars show the spend in each of your channels throughout the selected time period while the colored bars show as the ROI, MROI and Uncertainty associated with each channel.

Open

Notes:

• The ROIs/CPAs shown are "total" ROIs/CPAs, meaning that we calculate both the predicted lower funnel spend caused by the upper funnel spend and the predicted lower funnel effect of the lower funnel spend and use these to adjust the estimate to more accurately reflect the total impact of the spend. For lower funnel channels, no additional adjustment is necessary.

• The CI Width is the width of the interquartile range of the Bayesian confidence interval on the ROI for ROI models. For CPA models, the CI width is the interquartile range for the number of acquisitions per $1000 spent in the channel. We report confidence interval widths on the ROI scale because it helps us avoid problems with CPA uncertainty expanding as CPAs increase. For example, if your overall CPA was $100, and one particularly bad channel had a CPA estimate of $2000-$2500, the CI width ($500), doesn't properly convey the model's high confidence that the channel is not performant. By reporting widths on the ROI scale, we properly convey that we have high confidence in the poor performance.

• These are estimates based on the total return earned (but not necessarily realized) in time period. The model expects that additional conversions will come in the future.

• The marginal ROI will always be less than the average ROI due to diminishing returns. You can see each channel’s diminishing returns curve on the Channel-Level Deep Dive tab.

• The "▼" symbol represents channels that are configured as Lower Funnel Channels, which are configured differently within the Recast platform. More details are available on Lower Funnel Channels.

• Downloads for all the confidence intervals are available at the bottom of the page.

This graph shows the channels that a have relatively high or low share of effect compared to the share of spend. Share of spend is calculated by dividing the channel's spend by the total spend in the time period. Share of effect is calculated by dividing the estimated impact of the spend by the estimated impact of all other channels. The percentage shown on the graph is share of effect - share of spend.

Channels that represent a large portion of the total spend compared with the amount of KPI they are driving are considered ‘underperforming’ channels

Channel that represent a smaller portion of the total spend compared with the amount of KPI they are driving present outsized returns and are considered “over-performing’ channels.

Use this graph alongside the Marketing effectiveness graph to identify channels to further explore as potential places where there is an opportunity to increase or decrease spend to improve your overall performance.

The graph shows a breakdown of your KPI during the timeframe and how much each marketing channel contributed to that KPI based on the Recast model.

Notes:

• The Baseline is all of the KPI that’s attributed to non-marketing, organic sales.

• Spikes are things like promotional events, store closures, or new product launches. You can see more detail on the “Spike Summary” tab.

• “Unexplained Variation” is due to the random variation where sometimes the model estimates miss high or low. This should average out to zero over long time frames, but within a given week you may randomly have more (or less) unexplained variation just due to randomness.

• Channels contributing a small percentage of the total will be grouped into an "Other" channel.

To learn more about each of these graphs visit Overview in the Insights section

The baseline summary shows the non-marketing (organic or “base sales”) portion of your business in a specified time period. That is, how many sales would you have in the absence of any marketing activity (or at least the marketing activity included in your Recast model). You can choose whether to include spike effects (holidays, promotions, etc) or not. This report can be run into the future to see the changes in the forecasted intercept.

The graphs in this report are line graphs showing the monthly and weekly changes in the baseline

throughout your reporting period. Hover over the line to view the mean, median and range of the intercept during a particular week or month.

Open

This report also contains a summary plot showing the daily intercept value every 7 days in the reporting period.

The shift curve report shows the estimated time it takes between when marketing activity happens in a channel (generally when money is spent) and when the effect of that activity is realized (e.g., in terms of conversions or revenue), for any channels you specify. You may specify which channels you want to run the report for.

The day when marketing activity takes place is day 0

Open

The spend response curve report lets you see the expected return at various levels of spend in the form of impact, ROI/CPA, and MROI/MCPA. You can specify a previous or future date to understand (for example) how a major shopping holiday like Black Friday may affect the total impact of your spend. To use, select which date you’d like to calculate this for and the channel you would like to include. Recast forecasts these curves 730 days into the future from the last modeled date.

Use your cursor to hover over the graph to see how much direct impact each of your channels contributed at the given level of spend as well the aggregate impact at that spend level at the top of the hover over box.

Using the Multi-Day Spend Response Report, you can find the expected return over a range of spend spread across your selected number of days.