How to Understand Your Users Better and Deliver a Top-Tier Experience with Custom Properties

In today’s digital landscape, understanding your users’ preferences and behaviors is crucial to delivering tailored services. Every interaction, click, and engagement offers valuable insights into what your user needs and desires, but harnessing this data effectively requires the right tools and approaches.

This is where Helicone’s Custom Properties come into play.

What are Custom Properties?

Custom Properties are metadata that you can attach to your LLM requests. In Helicone, you can tag your requests with specific information to create detailed segments of your data, such as session/conversation/app IDs, the prompt chain or metadata about the user who made the request.

In this blog, we’ll explore:

• the importance of custom properties, and how they enable a deeper understanding of user behavior, preferences, and patterns.
• how you can tailor your services for your customers to drive business growth.
• how companies like Journalist AI and Greptile use custom properties for user analysis.

Use Case 1: Environment Tracking

Tracking environments in Helicone allows you to monitor and differentiate between various operational stages such as development, staging, and production.

Implementation

client.chat.completions.create(
    # ...
    extra_headers={
        "Helicone-Property-Environment": "development", # simply add this header to your request
    }
)

Best for:

• Compare performance metrics across different stages on Helicone’s dashboard. Address any issues or performance bottlenecks before moving to production.
• Identify cost-saving opportunities by tracking costs incurred at each stage, removing unnecessary expenses and optimizing resource usage.
• Collect early user feedback from staging environments and compare across stages to track improvements and refine your product.

Use Case 2: User and Feature Segmentation

Understanding user behavior patterns makes it easier to find opportunities to personalize marketing campaigns, introduce new features, or expand into new markets based on trends or demand patterns.

1. Geographic Location

Monitoring how your application perform across different geographic regions.

    headers={
        "Helicone-Property-User-Type": "paid",
        "Helicone-Property-Country": "USA", # simply add this header to your request
    }

Best for:

• Optimizing resource allocation
• Refining pricing strategies
• Targeting marketing efforts towards high-value user segments

2. Feature Segments

Monitoring feature performance or emerging trends among users.

    headers={
        "Helicone-Property-User-Type": "paid",
        "Helicone-Property-Feature": "spellcheck", # simply add this header to your request
    }

Best for:

• Smartly optimizing development efforts
• Prioritizing features with proven demands
• Aligning development priorities with user needs and pain points

3. Plan Segments

Implementing precise usage-based billing models by user types (free vs. premium).

    headers={
        "Helicone-Property-User-Type": "paid",
        "Helicone-Property-Plan": "enterprise", # simply add this header to your request
    }

Best for:

• Tracking user-specific metadata such as subscription plans, usage tiers, or subscription renewal dates
• Identifying opportunities to upsell or cross-sell additional services
• Tailoring subscription offerings to better align with customer preferences

Companies Who Use Custom Properties In Production

Journalist AI

is a platform that automates content creation for writers seeking efficient and high-quality content production, by generating written articles, blog posts, and other materials based on user input.

Journalist AI segments their users by plans using Custom Properties to:

• Compare the behavior of subscribed vs. non-subscribed users to see which features they use more frequently.
• Gauge the value of that feature in driving subscriptions if certain features are more popular among subscribed users.
• Track the costs associated with providing services to subscribed vs. non-subscribed users to understand the cost-benefit ratio of subscription plans.
• Market certain features more aggressively within their premium plans, if certain features are used more frequently by free users.

Greptile

is a platform that helps users search and analyze text data from sources including websites, documents, and databases, using advanced natural language processing (NLP) techniques.

Greptile segments requests by repository (Helicone-Property-Repository) to measure:

• Number of Requests made to each repository to gauge which repository is the most popular.
• Latency and Costs associated with requests to each repository.
• User patterns among different repositories.

Resources you might find helpful:

→ Doc: Setting up custom properties.

→ Step-by-step guide: Using Custom Properties to Segment Data.