4. The Next Frontier of Insights and Optimization
Opportunities Exists in Outside Data Sources
• Weather
• Prices
• Competitors
• Store Locations
• Audience Profiles
5. The Old Way – 3-9 months, $30K-$90K cost
• Prerequisite: data scientist or analyst with some free time
• Thesis (8-12 weeks)
• Identify potential data sources (2 weeks)
• Evaluate and decide on data source (2-4 weeks)
• License data (4-8 weeks)
• Cleanse and normalize data to fit your internal data hierarchy (2 weeks)
• Run statistical models to find correlations and insights (2-6 weeks)
• Now, how do we act on it?
6. The cClearly Way
• Prerequisite: data scientist or analyst with some free time not required
• Thesis (8-12 weeks) not required
• Identify potential data sources (2 weeks) – thousands already built-in
• Evaluate and decide on data source (2-4 weeks)
• License data (4-8 weeks)
• Cleanse and normalize data to fit your internal data hierarchy (2 weeks)
• Run statistical models to automatically find correlations and insights
• Now, how do we act on it? Actioned on automatically through integrations
7. Example: Store Location
cClearly Algorithms
cClearly identified that when the consumer is located less than
5 miles away from competitor store, conversion rate drops by 40%
cClearly compiles list of all zip codes that
have a competitor store within 5 mile radius
cClearly lowers bid for all relevant zip codes across all marketing channels,
resulting in a decrease in cost per conversion and an increase in conversion volume
Paid Search Ads Display Ads Video Ads Facebook Ads
cClearly Database Customer Data
8. Example: Zip Code based Audience Profiles
• Zip codes with higher percentage of college educated, non-smokers convert
significantly better
People who have a Bachelor’s Degree
have a higher ROAS for this Brand. In
other words, more educated people
provide a better return on ad spend.
Graph shows that Smokers have a
less favorable ROAS and this
Brand should avoid them