Bid Strategy Selector

Bid Strategy Selector

A skill for matching the right Google Ads bid strategy to each campaign's data and goals.

What this skill does

Given campaign data, this skill recommends one of seven bid strategies per campaign with:

1. The reasoning (data-backed)
2. The recommended target value (tCPA dollars or tROAS percentage) where applicable
3. A transition plan (don't switch cold)
4. Expected performance change
5. Risk factors and what to monitor

The seven strategies (when to use each)

Read references/strategy-comparison.md for the full breakdown. Quick guide:

Required inputs

1. Campaign data with at least 90 days of history
2. Per-campaign monthly conversion volume - required for Smart Bidding eligibility checks
3. Conversion tracking quality - does the user track conversion VALUE, not just count? (Required for Target ROAS / Max Conv Value.)
4. Business goal per campaign - what are they optimizing for? (Volume? CPA target? Specific ROAS?)
5. Budget headroom - are campaigns hitting daily budget caps?

If conversion value tracking isn't set up but the user wants ROAS-based bidding, flag this as a prerequisite.

Workflow

Step 1: Per-campaign assessment

For each campaign, gather:

• Current bid strategy
• Last 90 days monthly conversion volume (avg)
• Conversion value variation (CV - coefficient of variation = std dev / mean)
• Budget utilization (is daily budget capped?)
• CPA stability (90-day CPA standard deviation as % of mean)
• Performance trend on current strategy (improving, stable, declining)

Step 2: Eligibility filter

Apply hard rules first:

• Conv volume < 15/month → Manual CPC only. Smart Bidding needs data.
• No conversion value tracking → no Target ROAS or Max Conv Value possible. Recommend setting up value tracking as a prerequisite.
• Conv volume 15–30/month → Maximize Conversions OK; Target CPA risky (will work, but learning is unstable).
• Conv volume 30–50/month → Target CPA OK; Target ROAS still risky.
• Conv volume 50+/month with value tracking → all strategies on the table.

Step 3: Match strategy to goal

Within the eligible set, match to the user's stated goal:

Step 4: Recommend the target value

If recommending Target CPA:

• Start at the campaign's recent (last 30 days) CPA, NOT the user's wishful target. tCPA shouldn't be set 30% below current CPA - Google will choke volume to hit it.
• Tighten over weeks if performance allows.

If recommending Target ROAS:

• Start at recent achieved ROAS, then tighten 5–10% per month.
• Same principle: aspirational targets kill volume.

State the recommended starting target explicitly with the formula: "Set tCPA at $X, which is your last-30-day average. After 4 weeks of stable performance, consider tightening to $Y."

Step 5: Transition plan

Switching bid strategies cold is the #1 cause of post-switch CPA spikes. For each switch, recommend:

1. Test on one campaign first. Don't bulk-switch.
2. Use portfolio bid strategies if multiple similar campaigns exist - let them learn together.
3. Expect a 7–14 day learning period. Do NOT optimize during this window.
4. Set conservative initial targets (looser than current performance).
5. Don't pause keywords or change budgets during learning. It resets the model.
6. Monitor for 4 weeks before declaring success or failure.

Step 6: Risk factors

For each recommendation, surface 1–3 risks:

• "If conversion volume drops below 30/month, tCPA will become unstable."
• "If your competitor escalates, Maximize Conversions may bid up significantly."
• "Target ROAS at 400% may starve volume; willing to lower target if scale matters more."

Output format

Per-campaign recommendation table:

After the table, an overall summary:

• How many campaigns to switch this month
• Order of operations (which to switch first - usually start with the highest-volume, most-stable campaign)
• Prerequisites (e.g., "set up value tracking before recommendation X")

What this skill must NOT do

• Don't recommend Smart Bidding for sub-15-conv campaigns. It will perform worse than Manual.
• Don't recommend Target ROAS without confirming the user has reliable conversion value data.
• Don't recommend tightening targets aggressively. Tighten 5–10% at a time over weeks.
• Don't say "Smart Bidding is always better." Manual CPC is the right answer for low-volume campaigns and certain branded/specialty contexts.
• Don't recommend switching a campaign that's been on its current strategy for less than 8 weeks. Let learning complete first.
• Don't recommend Maximize Clicks unless the user explicitly has an awareness goal with no conversion tracking; it almost always wastes spend otherwise.

Reference files

• references/strategy-comparison.md - full strategy-by-strategy breakdown
• references/transition-playbook.md - step-by-step transition checklist with timing