The blunt answer: doubling your Google Ads budget overnight rarely doubles your results - and on Smart Bidding it usually halves your ROAS for 2-3 weeks before things settle. The auction expands, CPCs rise, the bidding algorithm re-enters a learning phase and conversion quality drops while Smart Bidding tests new audiences and queries. Below, Ben breaks down exactly what happens day by day, why, and the disciplined way to actually scale.
The short answer
Doubling daily budget triggers four predictable things on a Smart Bidding account: (1) a 7-21 day learning phase where bid stability collapses, (2) CPA rises 30-60% temporarily as the algorithm tests broader queries, (3) conversion volume rises - but typically only ~1.4x, not 2x, (4) ROAS often halves before recovering. The exceptions: brand campaigns, manual CPC, and accounts already constrained by >50% lost IS to budget. For everyone else, scale at 20% increments instead.
The 30-day timeline of doubling budget
| Period | What you see | What's happening |
|---|---|---|
| Day 1-3 | Spend rarely hits new budget; CPCs creep up | Smart Bidding cautiously testing higher bids |
| Day 4-10 | Spend ramps; CPA jumps 30-60%; conversions wobble | Algorithm exploring broader queries and audiences |
| Day 11-21 | Volume rises; CPA still elevated; ROAS down | Learning phase exiting on highest-volume actions |
| Day 22-30 | CPA partially recovers but rarely back to baseline | New equilibrium - efficiency permanently lower |
Why Smart Bidding panics
Smart Bidding (tCPA, tROAS, Maximise Conversions / Conversion Value) builds a model of which auctions convert. When daily budget doubles, three things change in its inputs at once:
- Auction footprint expands - the algorithm has to enter auctions it previously skipped because they were too expensive or too speculative.
- Pacing changes - it can spend more aggressively during high-value hours, but also needs to find enough volume to absorb the extra budget.
- Signal noise rises - new query patterns, new audiences, new placements all generate fresh data that the model has to reweight.
Result: a fresh learning phase. The campaign status may not literally show "Learning" but performance behaves as if it is.
What changes in the auction
Three concrete shifts UK advertisers see in account data after doubling:
- Average CPC rises 15-40%. Smart Bidding bids higher to win incremental volume the previous budget couldn't afford.
- Search query mix broadens. Queries that were previously ignored (longer-tail, weaker intent, partial matches) start triggering ads.
- Conversion rate drops 10-25%. Because the broader query mix has lower commercial intent on average.
What happens by campaign type
Search (Smart Bidding): the textbook bad case described above. 2-3 weeks of inflated CPA.
Search (Manual CPC): usually fine. With manual bids and stable Quality Scores, doubling budget mainly buys the auctions the campaign was already losing on impression share - no algorithm to retrain.
Performance Max: the worst case. PMax rebalances spend across Search, Shopping, Display, YouTube and Discover at once. Doubling budget can spike Display and YouTube spend - usually the lowest-quality channels - and CPA explodes for 3-4 weeks.
Shopping (standard): handles it better than PMax. CPA still rises ~20-30% for 10-14 days but recovers cleanly.
Brand campaigns: no real impact. Search demand for your brand is finite and intent is so high the algorithm doesn't need to re-explore.
Demand Gen / Video: CPM spikes, CPA volatility lasts 14+ days. Treat like PMax.
Real numbers from a UK account Ben rescued
An ecommerce client doubled spend on a Performance Max campaign from £200/day to £400/day to "go big" on Black Friday prep. The 30-day after-action read like this:
- Daily spend: £200 → £400 (delivered ~£385 average).
- Conversions: 8.2/day → 11.4/day (a 1.39x lift, not the 2x expected).
- CPA: £24.50 → £33.80 (+38%).
- ROAS: 4.8 → 2.9 (-40%).
- Day-30 CPA: £28.10 (still 14% above pre-double baseline).
The "extra" 3.2 conversions/day cost £185 in incremental spend - £58 each. The original conversions cost £24.50 each. Net: a campaign that was healthy at £200/day became unprofitable above that threshold for the entire month.
When doubling actually works
Three scenarios where doubling is safe:
- Heavily budget-constrained accounts. If lost impression share (budget) is >50%, you're missing so much demand that the auction expansion won't be material - you're just funding auctions you were already winning, more often.
- Brand campaigns. Demand is finite and intent is high - doubling rarely affects efficiency.
- Manual CPC with stable Quality Scores. No Smart Bidding to retrain; you're literally just paying for more clicks at the same price.
The right way to double your budget
Use 20% steps every 7-10 days. Doubling takes 4-5 weeks instead of one day, but you keep CPA in target the entire way:
- Week 1: +20% (e.g. £200 → £240).
- Week 2-3: hold, monitor CPA.
- Week 3: +20% (£240 → £288).
- Week 4-5: hold.
- Week 5: +20% (£288 → £346).
- Week 6-7: hold.
- Week 7: +20% (£346 → £415) - now slightly above the original "double".
Total: ~7 weeks from £200 to £400/day with stable performance, versus 4 weeks of broken performance from a single overnight double.
How to recover if you've already doubled
- Don't panic-revert. Pulling budget straight back triggers another learning phase in the opposite direction.
- Hold the new budget for 14 days. If CPA hasn't stabilised by day 14, then drop budget by 20% (not all the way back) and hold another 14 days.
- Don't change tCPA / tROAS targets at the same time. One variable at a time.
- Check the Insights tab for new search themes, audiences and placements - sometimes the post-double mix surfaces a new winning segment that justifies a campaign split.
- If the new mix is genuinely lower quality, add the worst-performing search terms / placements as negatives and let Smart Bidding rebuild around what's left.
