Lukas developed a sophisticated multi-signal approach using Valley that combines website intent data, viral content engagement, and rigorous A/B testing methodology. His strategy focuses on two primary high-intent signals rather than broad demographic targeting.

Signal-Based Targeting Strategy:

1. High-intent website visitors: Prospects from Western countries in relevant B2B positions who have demonstrated product interest

2. Viral content engagement: Scraping competitor and industry influencer posts that generate hundreds of engaged prospects interested in specific product subsets

"I am never pitching blindly and my signals are based on the people general affinity to the idea," Lukas explains.

Campaign Configuration Methodology:

• Separate product voices: Distinct writing styles and product descriptions for each offering to appeal to different demographics

• A/B testing framework: Always launching two campaigns per product with different rules to optimize performance

• Research agent optimization: Selecting all five available agents for maximum information gathering, customized by product type

• Quality filtering: Deleting low-fit prospects, manually reviewing medium-fit prospects, keeping high-fit prospects

Message Approval Process: Lukas maintains a rigorous 80/20 approval ratio—80% of messages require no editing, while 20% are either regenerated due to hallucinations or scraped entirely for poor ICP fit. "I never used Valley in fully autonomous mode. I still approve all of the messages."

Channel Strategy: Connection requests significantly outperform InMails, delivering 70% reply rates compared to 38% for InMails. "Connection requests, absolutely. I never got it higher than 38% reply rate on InMails, which for me is actually a low reply rate."

"Valley only recently launched blank connection messages, which I actually requested from the start. I absolutely love written connection messages with at least a short pitch which would actually fit."