What methods do you use to forecast demand for new products or features?

Forecasting demand for new products or features requires a blend of quantitative data and qualitative insights. Methods like market research, customer surveys, pilots, historical data, and predictive analytics help outbound PMs reduce uncertainty and make informed forecasts.

1. Market Research and Trend Analysis

   • Example: Apple’s Launch of AirPods
     Apple identified the growing trend of wireless technology and shifting consumer habits toward convenience, using market research to predict strong demand for AirPods.

   • Approach: Analyze industry reports, competitor trends, and emerging technologies to gauge demand for new products.

   • Action: Use tools like Gartner or IDC reports to benchmark trends and identify market demand signals. Track search trends and social media mentions to validate emerging opportunities.

2. Customer Surveys and Feedback

   • Example: Dropbox Business Expansion
     Before expanding its product from personal use to business, Dropbox conducted customer surveys to understand if existing users would pay for team collaboration tools.

   • Approach: Use surveys, focus groups, and interviews to collect quantitative and qualitative feedback from target customers.

   • Action: Run email surveys or in-product polls asking customers if they would use or pay for a new feature. Use Net Promoter Score (NPS) data to identify enthusiastic customers who might demand new offerings.

3. Pilot Programs and Beta Testing

   • Example: Tesla’s Full Self-Driving Beta
     Tesla released its autonomous driving software to a select group of drivers to assess demand and collect feedback.

   • Approach: Launch a pilot or beta program to a small group of customers. Measure usage rates, adoption, and feedback to determine market appetite.

   • Action: Track KPIs like feature engagement and satisfaction scores during the pilot to forecast broader demand upon full release.

4. Sales Forecasting Based on Historical Data

   • Example: Microsoft Office 365 Subscription Model
     When Microsoft transitioned to a subscription model, it used historical data from earlier product sales to predict customer migration and demand for the new model.

   • Approach: Use sales trends, adoption patterns, and growth curves from similar products or features to forecast demand for the new offering.

   • Action: Analyze existing customer behavior to identify segments likely to adopt the new product and use historical sales data to estimate future demand.

5. Conjoint Analysis and Willingness-to-Pay Studies

   • Example: Uber’s Dynamic Pricing Model
     Uber used conjoint analysis to understand how much users value features like shorter wait times and willingness to pay surge pricing during peak hours.

   • Approach: Use conjoint analysis to determine how customers trade off features, benefits, and price points, helping forecast demand at different pricing tiers.

   • Action: Conduct conjoint surveys where customers rank preferences for various product configurations, then model demand based on the results.

6. Pre-Orders and Early Sign-Ups

   • Example: Tesla Cybertruck Pre-Orders
     Tesla used pre-order campaigns to gauge market interest in its Cybertruck, collecting deposits to forecast production needs.

   • Approach: Offer pre-orders, waitlists, or early access programs to measure interest before launch.

   • Action: Monitor the number of pre-orders and early sign-ups as leading indicators of demand.

7. Sales Team Feedback and Channel Insights

   • Example: Salesforce Lightning Adoption Forecast
     Salesforce gathered insights from its sales teams and partners to forecast customer interest in migrating to the new Lightning platform.

   • Approach: Collect input from sales teams, partners, and channel data to forecast demand. Frontline salespeople often have valuable insights into customer interest.

   • Action: Hold regular feedback sessions with sales and partner teams to capture signals of market readiness and customer demand.

8. Predictive Analytics and Machine Learning Models

   • Example: Amazon’s Inventory Forecasting for New Products
     Amazon uses machine learning models that consider historical sales, seasonality, and external factors to forecast demand for new products.

   • Approach: Use predictive analytics tools to identify correlations and patterns in customer behavior, pricing, and market trends.

   • Action: Build predictive models based on internal data (past sales, web traffic) and external signals (market trends, economic factors) to estimate future demand.