Liam Walsh
Liam is a Co-Director at Intelligency and heads up the agency's Digital Intelligence & Paid Social activity. Over the last decade, he has worked with brands from the world of sports such as Premier League clubs to entertainment such as Channel 4 and Disney.
Google Ads automation is powerful, but it’s not always right. Tools like Smart Bidding, Performance Max (PMax), and Responsive Search Ads (RSAs) can improve efficiency, but only if they’re learning from the right signals. Left unchecked, automation can drift – raising costs, wasting budget, or driving poor-quality leads. The key is knowing when automation is helping, and when it’s steering you off course.
When Automation Goes Wrong
Performance Max Cannibalisation: PMax often takes credit for branded searches you already would have captured with Search campaigns. Even with brand exclusions, it can still show against these queries, giving the illusion of strong results while draining budget from your Search activity. If your Search campaigns show rising “Lost Impression Share (rank)” while PMax spend climbs, that’s a red flag.
Auto-applied Recommendations (AARs): Google can automatically “optimise” campaigns by adding broad match keywords or changing targeting. While positioned as helpful, these changes often expand far, leading to irrelevant clicks and higher CPAs. Always review and opt out of auto-applied changes you don’t want.
Modelled Conversions: Sometimes Google reports more conversions than you see in your CRM. That’s because it fills tracking gaps with estimated data. If left unchecked, Smart Bidding may optimise toward inflated numbers rather than actual leads or sales. The fix is to import offline or qualified conversions, so Google optimises for what truly matters.
When It’s Just Limited
Not all warnings mean failure. Sometimes automation is simply restricted by your setup:
Limited Status: A red label often means your targets are too strict or budget too low. A yellow label usually just signals low volume or “learning mode.”
RSAs: Too few headlines limit learning, while too many spread thin data across combinations. Match the number of assets to your budget (around 8–10 headlines, 2–4 descriptions).
Conversion Lag: Google records conversions by click date, not completion date. This can cause reporting delays but it doesn’t mean automation is broken.
How to Keep Automation Honest
Use impression share metrics, bid strategy reports, and simple scripts to monitor automation. Scripts for anomaly detection, budget pacing, or poor query themes can flag issues before wasted spend builds up.
The Bottom Line
Automation won’t always fail dramatically – it usually drifts. Your role is to guide it: set clear goals, track real results, and step in when it veers off course. With the right guardrails, automation becomes an asset instead of a liability.
