This guide covers why new campaigns need different targets, the three-phase approach, how to calculate your floor, target by campaign type, common mistakes, and when to tighten vs. loosen.
1. Why New Campaigns Need Different Targets
Google's Smart Bidding algorithms need data to work. Specifically, they need conversion data. A new campaign has zero conversion history, which means the algorithm is making educated guesses about who will convert and how much to bid.
If you set an aggressive tROAS target from day one (say, 5x), you're telling the algorithm "only bid on clicks that are very likely to return 5x." But the algorithm has no idea what a high-probability click looks like for your specific product and audience. So it becomes extremely conservative, bids very low, and your campaign barely spends anything.
The result: low spend, low volume, and inconsistent data. The algorithm never gets enough conversions to learn, so it stays conservative indefinitely. This is the classic "my new campaign won't spend" problem, and it's almost always caused by targets that are too tight too early.
2. The Three-Phase Approach
Here's the framework we use when launching new campaigns through our PPC management service:
Phase 1: Learning (Weeks 1-3)
Use Maximize Conversion Value with no ROAS target. Let Google spend freely within your daily budget and find conversions wherever it can. The goal in this phase is data collection, not profitability. You're going to overpay for some clicks. That's fine. You're buying learning data.
Set a daily budget you're comfortable with as a "learning investment." For most ecommerce brands, this is $50-$150/day depending on AOV and category.
Phase 2: Calibration (Weeks 4-6)
After you have 30-50 conversions, look at your actual trailing ROAS. Set your tROAS at 80% of that number. So if you're averaging 3.5x, set your target at 2.8x. This gives the algorithm a target it can actually hit while still pushing it toward better performance.
Monitor volume closely during this phase. If impressions drop by more than 30%, your target is too tight. Lower it by 10% and give it another week.
Phase 3: Optimization (Week 7+)
With 50+ conversions and stable performance, increase your tROAS by 5-10% every 1-2 weeks. The algorithm has learned enough to handle gradual tightening. Keep pushing until you find the sweet spot where ROAS is above your break-even and volume is still acceptable.
This entire process takes 6-8 weeks. Rushing it usually makes things worse, not better.
3. Calculate Your ROAS Floor
Before setting any target, you need to know your break-even ROAS. This is the minimum ROAS at which your ads don't lose money.
The formula: Break-even ROAS = 1 / gross margin (as a decimal). If your gross margin is 50%, your break-even ROAS is 2.0x. If it's 35%, your break-even is 2.86x. See our true ROAS guide for a detailed walkthrough.
Your ROAS floor is this break-even number. Never set a tROAS target below it (unless you're intentionally buying customer acquisition at a loss for LTV reasons). And during Phase 1, understand that you might dip below this floor temporarily. That's the cost of learning data.
4. Targets by Campaign Type
Not all campaigns should have the same target. Here's a realistic framework:
- Branded Search: tROAS of 8x-15x. These clicks are cheap and high-intent. Set an aggressive target from day one because branded search doesn't need a learning phase.
- Non-Branded Search: Start with no target, graduate to 2x-4x depending on margins. This is where the three-phase approach matters most.
- Shopping / Performance Max: Start with no target. Graduate to 3x-5x. Shopping has more data signals (product feed, pricing, images), so it tends to calibrate faster than Search.
- Remarketing: tROAS of 5x-8x. These are warm audiences who already visited your site. Higher targets are appropriate because conversion probability is higher.
- Prospecting (Meta/Display): Start with no target, graduate to 1.5x-3x. These are cold audiences and will always have lower ROAS than Search. That's fine. They feed the top of the funnel. For benchmarks, see our ROAS benchmarks by category.
5. Common Target-Setting Mistakes
Mistake 1: Setting the same target across all campaigns. A single 4x target means your branded campaigns are under-optimized (they should be at 10x+) and your prospecting campaigns are over-restricted (they'll never hit 4x).
Mistake 2: Using aspirational targets instead of data-driven ones. "We want 6x ROAS" is a wish, not a strategy. Base your targets on your actual break-even math and trailing performance, not what you hope to achieve.
Mistake 3: Changing targets too frequently. Every time you change a tROAS target, the algorithm enters a mini-learning phase. Changing targets every few days means the algorithm never stabilizes. Set it, wait 2 weeks, then reassess.
Mistake 4: Not accounting for seasonality. Your Q4 ROAS of 6x is not your baseline. Setting a 6x target in January will restrict your campaigns during a naturally slower period. Use trailing 30-day data for the current season, not peak-season numbers.
6. When to Tighten vs. Loosen Targets
Tighten (increase tROAS) when:
- Your actual ROAS is consistently 20%+ above your target for 2+ weeks
- You have conversion volume to support it (50+ conversions in the last 30 days)
- You want to improve profitability and are willing to accept less volume
Loosen (decrease tROAS) when:
- Campaign spend dropped significantly after setting or raising a target
- You're in a seasonal dip and need to maintain share of voice
- You want more volume and your current ROAS is well above break-even
- You're launching new products that need traffic to build conversion data
The key principle: ROAS targets control volume, not just profitability. A higher target means less volume. A lower target means more volume but potentially lower efficiency. Finding the right balance is an ongoing process, not a one-time decision.
Frequently Asked Questions
For new campaigns, start without a ROAS target. Use Maximize Conversion Value for the first 2-4 weeks to let the algorithm learn. Once you have 50+ conversions, set your tROAS at 80% of your actual trailing ROAS. Tighten gradually from there.
Most new campaigns take 4-8 weeks to reach stable ROAS. The first 2 weeks are learning phase with volatile performance. Weeks 3-4 show more consistency. By week 6-8, you should see performance close to your target if the campaign structure is sound.
Yes, always. Branded campaigns should target 8x-15x ROAS. Non-branded prospecting should target your break-even ROAS plus a small margin, usually 2x-4x depending on category. Using one target for both types will either starve non-branded or waste money on branded.
Find Out What's Costing You
COREPPC's free audit checks your Google Ads and Meta accounts in 60 seconds. Get your performance score and see where to improve.
Start Free Audit
