Using Augmented Analytics

This article provides an overview of the functionality and use cases of the Reveal element.

Introduction

The Reveal element enables you to analyze your marketing data to detect trends, anomalies, and gather campaign optimization recommendations. The data analysis capabilities of the Reveal element work out of the box. There is no setup necessary.

After you authorize an account from a supported advertising system, Adverity constantly analyzes its data to detect marketing insights and pro-actively displays them to you.

There are three major parts of the Reveal element:

  • Proactive Analytics

  • Forecasts

  • Channel Mix Optimization

To see the value proposition and the use cases for different aspects of the Reveal element, read Use Augmented Analytics on our website.

See our blog post Is Augmented Analytics the Ultimate Digital Marketing Tool for more information.

Proactive Analytics

With Proactive Analytics, you can visualize important information about:

  • Anomalies

  • Trends

  • Segment analysis

  • Budget suggestions

Proactive Analytics displays a timeline of findings based on your marketing data. Adverity updates the timeline once a day. Every finding is rated on an impact score on a scale:

  • Low

  • Medium

  • High

The impact score shows you how much attention you should pay to the displayed information.

Filtering findings of Proactive Analytics

Filtering the findings in Proactive Analytics helps you to see only the information of your interest. For instance, you can apply a filter to only show the high-impact findings. By default, you can see all of the findings. To filter and sort the findings, follow these steps:

  1. Click the Reveal element and select the Workspace you work with in Reveal.

  1. In the left navigation panel, click Proactive Analytics.

  2. In the top right corner of the page, click Show filters.

  3. Use the right panel to filter the results. You can filter for the following:

    • Finding types

    • Impact score

    • Show or hide acknowledged findings

    • Datasources

    • Funnel steps

    • Accounts

  4. Click Apply Filters.

Configuring notifications

Proactive Analytics allows you to receive notifications about anomalies, trends, and any other important information you configured. The configuration and filters made in Proactive Analytics user interface are applied to notifications automatically. To reconfigure the notifications you receive, change the filters and other configuration settings in Proactive Analytics.

Slack notifications

To configure Slack notifications in Proactive Analytics, follow these steps:

  1. Click the Reveal element and select the Workspace you work with in Reveal.

  1. In the top right corner, click Configure Notifications.

  2. Click Slack.

  3. Click Authorize Slack.

  4. Follow the authorization steps in the Slack web user interface. Select a channel in which you want to receive notifications.

  5. Click Create.

Microsoft Teams notifications

To configure Microsoft Teams notifications in Proactive Analytics, follow these steps:

  1. Click the Reveal element and select the Workspace you work with in Reveal.

  1. In the top right corner, click Configure Notifications.

  2. Click Teams.

  3. Provide a Microsoft Teams weebhook URL. For more information, see the Microsoft documentation.

  4. Click Create.

Understanding data completeness statuses

The data completeness statuses are displayed in a bubble next to the News Feed title. An example is shown in the figure below.

The data completeness status displayed in a bubble next to the News Feed title

The data completeness statuses for Proactive Analytics are as follows:

Complete

The Complete status is displayed once all scheduled data source collection tasks are complete.

Fetching

The Fetching status is displayed if any data collections associated with your dashboard have an active running task (for example, fetching, an ongoing enrichment, an ongoing data transfer).

Incomplete

The Incomplete status is displayed if there are any errors in the data collection from the data source connected to your dashboard. This error can happen in any of the previous steps, including fetching the data, enriching the data, transferring the data. You can only resolve authorization errors. For all other errors, contact Adverity support.

Computing

The Computing status is displayed while Adverity computes the findings to be displayed in the news feed. This status is typically displayed in the mornings while Adverity computes the findings during the night.

Forecasts

Forecasts help you to monitor your spending against your planned budget. This feature provides you with forecasting data for up to 30 days in the future.

To set up your budget, follow these steps:

  1. Click the Reveal element and select the Workspace you work with in Reveal.

  1. In the left navigation panel, click Forecasts.

  2. In the Forecast Range field at the top of the page, select the time range.

  3. In the Budget field at the top of the page, enter the planned budget for the selected period.

Filtering and sorting results of forecasts

To filter information in the budget forecast, follow these steps:

  1. Click the Reveal element and select the Workspace you work with in Reveal.

  1. In the left navigation panel, click Forecasts.

  2. Use the right panel to filter the results. You can filter the following:

    • Data sources

    • Accounts

    • Campaigns

  3. Click Apply.

Channel Mix Optimization

Channel Mix Optimization helps you find your optimal budget allocation to increase your Return on Investment (ROI). This is provided by forecasting the output of thousands of different budget allocations based on your historical data. Channel Mix Optimization can help you determine what type of marketing channels you need to use to achieve the maximum revenue for a defined budget.

In the right panel of the Channel Mix Optimization tab, you can:

  • Change the historical time range.

  • Change the media channels.

Calculating optimal ROI with Channel Mix Optimization

There are two essential pieces of data necessary for optimal functionality of Channel Mix Optimization ROI calculation:

  • Revenue: the revenue data of Google Analytics.

  • Costs: the costs data from any advertising data source which includes them.

To start your Channel Mix Optimization, follow these steps:

  1. Click the Connect element and select the Workspace you work with in Connect, Enrich & Transfer.

  1. In the left navigation panel, click Connections.

  2. In the upper right corner click + Add.

  3. Search for Google Analytics.

  4. Click Google Analytics. Add and authorize it as a Connect.

  5. Repeat this process with any advertising data source which includes costs. For example, Google Ads.

The following historical data is needed for a data model:

  • At least 2 years of historical data for an accurate data model.

  • Ideally 5 years of historical data for a very accurate data model.

ROI Advisor

Use the ROI Advisor to discover relationships between your investments and returns.

The ROI Advisor uses a Bayesian statistical model. The model estimates the revenue generated by your investment in a marketing channel. Unlike widely used attribution models, the unique statistical model of the ROI Advisor does not rely on unstable tracking techniques which can cause data quality issues.

The ROI Advisor gives you information about the following:

  • The amount of revenue each marketing channel contributed to total returns.

  • The amount spent on each paid channel over time.

  • The weekly returns for different weekly investment levels.

  • The delay between the investment and its impact on returns.

To start using the ROI Advisor, contact your Adverity Account Manager.

Appendix - Capabilities and limitations

Supported data sources

Proactive Analytics

Proactive Analytics supports a limited set of data sources:

  • Amazon Ads

  • Apple Search Ads

  • Criteo Marketing

  • Display & Video 360

  • Facebook Ads

  • Google Ads

  • Google Search Ads 360

  • Instagram Ads

  • Linkedin Ads

  • Microsoft Advertising

  • Twitter Ads

Forecasts

The Forecasts feature supports every data source available in the Connect element.

Channel Mix Optimization

To use the Channel Mix Optimization feature, Adverity needs to verify if your data sources are compatible on a case-by-case basis. Contact your Account Manager for details.

Getting valuable insights

For Proactive Analytics and Forecasts, there is no limit for metrics to generate widgets.

To get valuable insights, use the Proactive Analytics and Forecasts features for the following:

  • High-volume accounts (high amount of impressions, clicks, costs).

  • Campaigns that track conversions.

Anomalies: conversion lag

In digital marketing, conversion lag refers to the amount of time between ad engagement and conversion. As most users do not immediately convert after initially engaging with an advertisement, conversion lags can regularly appear. For more information, read our blog post Here's How Conversion Lag Corrupts Data Quality (and How to Fix It).

Adverity uses a regression model to forecast the effect of the conversion lag. This smooths the time series which is then subjected to outlier detection.

Trends: time range

The purpose of trend detection is to detect long-term trends. The time range is determined dynamically and can differ between trend widgets. A trend always ranges from the last turning point until yesterday.

Historical data necessary for accurate data model

For an accurate data model, some historical data is needed. The data you can provide to optimize the data model includes the following:

  • Advertising costs

  • Revenue

The time range of historical data which you can provide is the following:

  • At least 2 years of historical data for an accurate data model.

  • Ideally 5 years of historical data for a very accurate data model.

Data Models

Adverity uses one of two different models to forecast expenditure for the data sources, accounts, and campaigns that you select. Adverity chooses the model automatically based on the data you provide.

Autoregressive forecasting

Autoregressive forecasting is a complex and more accurate model. It takes into account both recent history and seasonal trends to make its predictions. However, you can only use this forecasting model if the data set satisfies specific conditions.

Adverity uses autoregressive forecasting if the selected data set fulfills the following requirements:

  • The data set has more than 30 days of data in the recent past.

  • The average daily cost for the last 30 days is more than 1.

Linear forecasting

Linear forecasting is a more simplistic model. Adverity uses linear forecasting when a data set does not satisfy the criteria for autoregressive forecasting. Linear forecasting relies on a mean average calculation from the available historical data, and does not take seasonal variations in data into account. Linear forecasting predicts the same number for every future day based on the average of the previous spending divided by the number of days.