THE PROBLEM
Client supports e-commerce merchants and retailers scale their business by automating sales and support conversations. Client helps brands and retailers stand out, build and enhance their brand image through hybrid solutions leveraging Conversational AI, and virtual customer assistance with the objective to deliver more personalized and memorable CX.
Client wanted to help e-commerce merchants and brands, develop their Amazon Campaign Bidding Strategy to enhance Return on Advertising Spend (RoAS) through optimized bid price recommendations for their advertisement on Amazon, where merchants bid on keywords and the merchant with the highest bid and best-targeted keyword finally wins the auction.
Client was already following a rule-based mechanism to find the bid price, but the client’s merchants were experiencing, very poor to zero Impressions, Click-Through Rate (CTR) and Return on Advertising Spend (RoAS) for many of their products. Hence, they partnered with us with the objective of improving Return on Advertising Spend (RoAS).
INXITE OUT APPROACH
Campaign Performance Assessment
The client sells a myriad of brands and ASINs through its vast merchant network. The very first step of our solution involved understanding the products and keywords performance trends during the previous bids in the business context.
Bid Price Responsiveness-based Product Profiling
The products were broadly categorized into three different classes (Highly Elastic, Moderately Elastic, and Non-Elastic) based on their sales responsiveness to changes in advertising spend during previous advertisement campaigns.
Bid Price Forecast Model Development
For each product class, products, sales, and advertising spend-related features were prioritized. A few key features include:
• Advertisement group
• Keywords and Match type
• Impressions
• Clicks
• Spend
• Sales
• Actual Return on Advertising Spend (RoAS)
• Cost-Per-Click (CPC)
Leveraged a deep-learning-based algorithm on the identified features for forecasting bid price for a business-determined Targeted Return on Advertising Spend.
Bid Strategy Finalization
Finally, while incorporating the actionable insights from the above-forecasted bid price, the client applied specific business logic wherever necessary and assisted the merchants with the right bid strategy to achieve the Targeted Return on Advertising Spend. Model recommendations were validated by running advertisement campaigns on Amazon.com.
RESULT
Results achieved for the different segments are as follows:
• Non-Elastic Products: A $1 increase in advertising spending generated $30 in additional revenue. Basis overall 12% increase in advertising spending, impressions increased by 24%, and cost-per-click was optimized by 37%.
• Highly Elastic Products: A $1 increase in advertising spending generated $5 in additional revenue. Basis overall 7% increase in advertising spending, cost-per-click was optimized by 29%.
• Moderately Elastic Products: A $1 increase in advertising spending generated $4 in additional revenue. Basis overall 64% increase in advertising spending, impressions increased by 35%, and cost-per-click was optimized by 22%.
Improvements achieved from the first round of bid strategy optimization convinced the client to extend the automated bidding approach to an additional 1000 product groups.
