Using customer searches to determine product assortment

About the Project

Companies utilize customer search data to shape their product strategies. A firm is able able to discern what types of restaurants and brands they should include, while we could identify new products that were in demand on their platform.

• Complex Search Data: Both companies found working with search data challenging due to its complexity, with millions of rows of custom strings entered by customers across different geographical locations.
• Data Filtering Difficulty: Identifying interesting and actionable search trends was difficult. Top search results (like "Pizza" or "Chinese" for Food, "Water" and "Snacks" for CPG) were common and unenlightening.
• Stakeholder Management: It was unclear how the chain of stakeholder management worked to convert insights into new products or actions, meaning the analytics needed to be integrated across the company.

Actions Taken:

• Dashboard and Data Pipelines: The companies built a dashboard and data pipelines to analyze customer search behavior. Key filters included highly searched items by area, trending items, highly searched but low converting items, and seasonal changes.
• Prioritizing High Demand Items: Sellers & merchandisers received force-ranked lists of highly searched, low-converting items to prioritize sourcing without overloading existing workloads.
• Routing Search Terms: Search terms were directed to appropriate teams, with opportunities divided between the national and local sales teams and the relevant category teams within merchandising.

Results:

• Improved Voice of the Customer & Predictability: The companies gained better predictability of customer needs and preferences, resulting in higher search conversion rates.
• Efficient Item Sourcing: New restaurants and items debuted by the companies performed better compared to items that were not highly searched.
• Easier Discovery of New Verticals: Identifying new verticals became easier. For example, boba went from accounting for just 2% of searches to becoming its own category of restaurants on the app. Similarly, other companies are able to predict which seasonal assortments were worth sourcing, thereby avoiding unwanted seasonal inventory spoilage.
• Search is the Basis for Ads and Retail Media: having a strong sense of search behavior allows for more relevant and nuances customer profiles which can be used to create content for greater CRM conversion.