Scrub for industrial controls

Control engineers researching real solutions often appear in the same search-term report as students looking for homework help, hobbyists setting up home automation labs, and IT consultants reading about technology for general knowledge without buying intent. A search-term report that includes 'Arduino Modbus library', 'Raspberry Pi industrial gateway tutorial', 'PLC programming for beginners course' is contaminated with non-buyer intent. Scrub runs NLP heuristics to detect consumer patterns — tutorial vocabulary, hobby language, educational intent — and bulk-suggests negatives. For controls specifically, contamination patterns include: hobby automation language (smart home, DIY, hobby), educational context (beginner, course, tutorial, university, student), maker platforms (Arduino, Raspberry Pi) unless you sell industrial-grade variants, and generic IoT queries (IoT platform, IoT gateway, connected devices). Scrub flags these and suggests bulk negatives. The AI-assisted classification helps with ambiguous terms: 'edge gateway' could be a professional controls engineer or a hobbyist, but 'edge gateway Rockwell Studio 5000' is almost certainly professional. Recommendation: run Scrub monthly on your search-term reports. Negative out any terms that are flagged as consumer contamination or educational intent. Create a permanent negative list for hobby and DIY vocabulary and apply it at the account level — this prevents new contamination patterns from being discovered by Smart Bidding.