Key Factors to consider:

1. Value to the Customer: Understanding the perceived value of the ML product to the customer to different customer segments helps in aligning the price with what customers are willing to pay. Ex: An ML product that increases revenue or significantly reduces cost to the customer, and has a high barrier to entry can command a higher price.

2. Cost Structure: Consider direct costs (Data Acquisition, Model Training and Infrastructure), and also indirect costs (R&D, Support). Make sure these costs are clearly understood to achieve desired profit margins.

3. Usage patterns: Different customers may use the product in various ways and volumes. Analyze the usage patterns to suit both light and heavy users, while maximizing the market reach and revenue. Ex: Subscription, Pay per use, tiered pricing models.

4. Scalable nature of ML products: As usage increases, the marginal costs of ML products decrease due to economies of scale. So, considering this makes usage-based or tiered pricing strategies more favored.

How to Manage Pricing for ML models:

1. Understanding Customer Value
2. Market Analysis
3. Experimentation (ex: A/B Testing)
4. Being flexible while product evolves
5. Being transparent about costs, avoiding hidden costs can boost customer trust.

Common Pricing models for ML Products:

1. Pay-per-usage (Usage based pricing)
2. Subscription
3. Freemium
4. Pay-per-model