Resources

Marketing Anomalies

What are Marketing Anomalies?

Morphio’s automatic anomaly detection helps track your data across different channels, metrics, and sources on a daily basis. Morphio’s machine learning is based on a predictive model, normalized historical data, trends and a few other parameters. When a significant difference is detected from the predicted normal value, Morphio will notify you via Slack, email and in-platform.

How is an Anomaly defined?

Morphio’s algorithm crunches metrics from multiple integrated platforms. The data is analyzed alongside ‘the same day’ from the previous three months and one year prior. The baseline for each channel and metric is calculated using a deep neural network model that considers many factors to predict benchmarks that are specific to your website.

In a normal distribution, it is anticipated that 95% of data fits inside two standard deviations of the average. Data variations outside of 2 standard deviations are labeled as a Level 2 Anomaly or Level 3 Anomaly. Check out the diagram below.

Anomaly Levels

What does your Anomaly card mean?

  1. Metric (Avg Time On Site, Sessions, Goal Completions, etc.)
  2. Channel (Paid Social, Organic, etc.)
  3. Anomaly level:
    a. Level 1 is a slight change
    b. Level 2 is a more severe change
    c. Level 3 is an extreme change
  4. Actual values
  5. The expected range of the values for this date, channel, and metric
  6. Related Goals, Prediction of the goal most likely to be affected.
  7. Colours:
    a. Dark and light green anomaly cards represent positive changes in your data
    b. Red and orange colours represent negative changes in your data

What channels and metrics are being tracked?

The below shows the metrics, dimensions, and channels that Morphio tracks for anomalies:

  • Multichannel anomalies:
    Source: Google Analytics
    Dimensions: All Users, Organic, Paid Search, Paid Social, Display
    Metrics: Sessions, Goal Completions, Transactions, Conversion Rate, Avg Session Duration and Bounce Rate
  • Device/Country anomalies:
    Source: Google Analytics
    Dimensions: Device/Mobile, Canada/US
    Metrics: Sessions, Goal Completions
  • Spend anomalies:
    Source: Google Ads, Facebook Ads, Microsoft Advertising, LinkedIn Ads
    Dimensions: All Users
    Metrics: Spend
  • Campaign anomalies:
    Source: Google Ads, Facebook Ads, Microsoft Advertising, LinkedIn Ads
    Dimensions: Campaigns
    Metrics: CTR, Clicks, Impressions
  • Search keywords anomalies:
    Source: Google Ads
    Dimensions: Search Keywords
    Metrics: Quality Score
  • Search ad anomalies:
    Source: Google Ads
    Dimensions: Search Ads
    Metrics: CTR
Back to Top