The Common Breaking Point
Almost every brand hits this moment. Shopping campaigns are delivering acceptable ROAS at current spend. The logical next step is to increase budget and capture more revenue. So you add 50% more budget.
Within two weeks, ROAS drops 30%. Conversion rate falls. CPCs spike. You pull budget back, but performance doesn't recover to previous levels. The algorithm seems "broken."
Scaling exposes problems that were always there. At lower spend, waste was hidden in the overall numbers. At higher spend, waste compounds and becomes visible.
This pattern is so consistent that we now assume any brand asking to "scale shopping" has underlying data problems that scaling will surface. Usually, we're right.
What Happens When You Increase Budget on a Dirty Feed
When your feed data is vague, you're matching to a mix of relevant and irrelevant searches. At lower spend, the algorithm might be selecting the better ones. At higher spend, it expands to include more of the worse ones.
Stage 1: Expansion to Marginal Queries
More budget means the algorithm looks for more opportunities. It expands to queries that are technically matches but have worse intent.
Stage 2: Conversion Rate Drops
Marginal queries have worse conversion rates. Overall conversion rate declines because you're mixing good traffic with increasingly bad traffic.
Stage 3: Algorithm Learns Wrong Signals
Smart Bidding learns from the traffic you generate. If that traffic doesn't convert, it starts optimising toward user profiles that don't buy.
Stage 4: Performance Volatility
The system oscillates trying to find what works. Performance becomes unpredictable. Days vary wildly. No bid adjustment stabilises results.
This is why pulling budget back doesn't restore previous performance. The algorithm has now learned from the bad traffic. It takes time (and often fresh campaigns) to reset those learnings.
How Google Learns the Wrong Signals Faster
Smart Bidding is a learning system. It observes what converts and what doesn't, then optimises toward more conversions. This is powerful when the inputs are right. It's destructive when inputs are wrong.
The Negative Feedback Loop
Bad feed data → wrong queries → wrong traffic → poor conversions → algorithm learns "this audience type doesn't convert" → targets similar (still wrong) audiences → performance degrades further.
At higher spend, this loop runs faster. More data means faster learning. If what you're teaching the algorithm is wrong, faster learning is worse, not better.
This is why performance becomes unexplainable. You're looking at bids and budgets when the problem is upstream in the feed. No bidding change fixes a matching problem.
What to Fix Before You Scale Spend
The answer isn't "don't scale." It's "fix the feed first, then scale." Scaling clean data scales revenue. Scaling dirty data scales waste.
Audit Your Search Terms Report
Before scaling, export the last 30 days of search terms. Are you appearing for queries that match your products and price points? If more than 20% of impressions come from irrelevant queries, fix the feed first.
Qualify Your Titles
Titles should repel wrong traffic as much as attract right traffic. Add material, quality indicators, and specific product details that filter out bargain hunters if you sell premium.
Complete Critical Attributes
Gender, age group, size, colour, material, and pattern help Google narrow matching. Missing attributes force inference, which expands matching in unpredictable directions.
Review Images for Price Alignment
Product images are the first filter. Do your images look premium if you charge premium? Do they communicate quality before the click happens?
Pre-Scale Checklist
Before increasing Shopping budget by more than 20%, verify each of these:
- 1Search terms are relevant: 80%+ of impressions come from queries where your product makes sense
- 2Conversion rate is stable: Not declining over the past 4 weeks
- 3No price-modifier queries: Not appearing for "cheap," "sale," or "under £X" unless that's your positioning
- 4CTR is reasonable: Above 1% for Shopping (below indicates matching or image problems)
- 5Titles qualify intent: Include material, quality, and product-specific terms that filter traffic
- 6Critical attributes complete: Gender, size, colour, material filled for all products
If Any Check Fails
Fix the feed first. Scaling with unchecked items multiplies the problem. Every week you scale dirty data teaches the algorithm wrong patterns that take longer to correct later.
Frequently Asked Questions
Why does performance get worse when I increase Shopping budget?
When you scale budget on a feed with data problems, you're amplifying waste. Google learns from the traffic you generate. If that traffic doesn't convert because the feed attracted wrong-intent searches, Google optimises toward more of the same bad traffic. Scaling accelerates this negative loop.
How do I know if I'm ready to scale Google Shopping spend?
Check three things: Is your search terms report showing relevant queries? Is your conversion rate stable (not declining with more spend)? Are you appearing for queries that match your price point and product type? If all three are true, you're likely ready to scale.
Why does my Google Shopping performance become volatile at higher spend?
Volatility at higher spend often indicates the algorithm is trying to find converting traffic but can't because the feed data is attracting wrong audiences. The system oscillates between placements trying to hit targets, but the underlying matching problem prevents stability.
Related Reading
Planning to Scale? Let's Check the Feed First
We'll audit your feed and search terms to identify what needs fixing before you increase spend. Better to know now than learn the expensive way.
Request a Feed AuditGet our insights in your inbox
Plain-English thinking about Google Ads. No spam, unsubscribe anytime.