--- name: calculate-bid-targets description: >- Derives defensible CPA, ROAS, or profit-per-spend targets from the account's real unit economics: compute the breakeven point where ad spend consumes all gross profit, pick how much of that profit the business is willing to reinvest in acquisition, and back the operational target out of those two numbers. Use it at onboarding, before any strategy switch that needs a target, whenever margins/pricing/close rates change, or when current targets were inherited or guessed. It ends with a validated number and (optionally) applies it to the campaign or portfolio strategy. Not for choosing the strategy type (select-a-bidding-strategy), configuring strategies end to end (set-up-conversion-based-bidding / set-up-value-based-bidding), or iterating targets on a live stable campaign (scale-bids-and-budgets). --- # Calculate CPA / ROAS / Profit-Based Bid Targets ## Purpose Smart Bidding converges on whatever target you state, so the target must encode the business's economics, a looser-than-necessary target donates margin to the auction, a tighter-than-viable one starves volume. This skill produces the number: verified economic inputs, breakeven levels, a deliberate profit-reinvestment choice, and a feasibility sanity check before anything is applied. ## When to run - Onboarding a new account (after unit economics are validated). - Before switching to tCPA/tROAS or creating a portfolio strategy that carries a target. - Pricing, margins, conversion rates, or close rates changed. - Stakeholder shifted the growth-vs-efficiency balance. - Live results sit persistently outside the acceptable range and the target's provenance is unclear. ## When NOT to run - Deciding which strategy type fits -> select-a-bidding-strategy. - The number exists and needs applying with learning-period management -> set-up-conversion-based-bidding or set-up-value-based-bidding. - Nudging a proven target toward a profit optimum -> scale-bids-and-budgets. - Unit economics themselves are unverified -> calculate-and-validate-unit-economics first; this skill assumes trustworthy inputs. ## Prerequisites - Verified economics (backend/CRM-sourced, not client guesses): deal values or order values, margins, order-level expenses, funnel conversion rates, churn/lifetime for subscriptions. - Documented growth and efficiency goals. - Conversion tracking live at the stage the bidding will optimize toward. ## Procedure 1. **Classify the vertical and pick the metric.** Sales-team lead gen -> CPA. Online store -> ROAS, or profit-per-spend if profit tracking exists (set-up-cart-data-and-profit-tracking). Subscriptions -> CPA bounded by lifetime value. 2. **Gather inputs (mostly HUMAN STEP, outside VigilDog).** Pull from backend/CRM: - Lead gen: average deal value, margin %, lead-to-close rate (and per-stage rates for multi-step funnels), lifetime value if repeat business matters. - Ecommerce: average order value, product margin %, order expenses % (shipping + payment + returns), effective margin = margin % - order expenses %. - Subscriptions: average revenue per user, lifetime = 1 / monthly churn, gross margin. Cross-check anything the client states; wrong inputs cascade into wrong targets. 3. **Compute breakeven** (the point of zero retained profit): - CPA breakeven = gross profit per closed deal x lead-to-close rate (chain multiply each funnel stage's rate for multi-step funnels). Compute it at the stage Google Ads optimizes toward. - ROAS breakeven = 1 / effective margin (35% effective margin -> 286%). - Profit-per-spend breakeven = 100% always, because margin is already inside the conversion value. 4. **HUMAN STEP (outside VigilDog), choose the reinvestment share.** Decide what fraction of gross profit goes to acquisition vs is retained: ~0.75 for aggressive growth, ~0.50 balanced, ~0.25 efficiency-first. This is a business decision the stakeholder must own, not a formula output. 5. **Compute the operational target:** - CPA target = breakeven CPA x reinvestment share (600 breakeven x 0.5 = 300). - ROAS target = breakeven ROAS / reinvestment share (286% / 0.5 = 572%), division, because a higher ROAS is stricter. - Profit-per-spend target = 100% / reinvestment share (0.5 -> 200%). Record breakeven, share, and target together; the trio is the audit trail. 6. **Read current state.** Call `gads_list_bidding_strategies` and `gads_get_campaign` for the affected campaigns: current strategy type, current targets, trailing actuals. The gap between the computed target and current actuals determines the rollout plan (large gaps close in 10-15% steps, they are never jumped). 7. **Feasibility check.** Approximate achievable volume at the computed target with `gads_keyword_forecast` (expected clicks/cost at relevant CPCs -> projected conversions at current conversion rates), and optionally pull `campaign_simulation` rows via `gads_run_gaql_query` where available. HUMAN STEP (outside VigilDog): for a richer curve, review Performance Planner / bid simulator in the Google Ads UI. Evaluate per Decision rules. 8. **HUMAN STEP (outside VigilDog).** Present the stakeholder the trio (breakeven / share / target) plus projected volume, and get sign-off on the tradeoff, especially if the validated target implies less volume than the growth goal wants. 9. **Apply (optional, write).** If asked to implement now: `gads_set_campaign_bidding_strategy` for a single campaign, or `gads_create_portfolio_bidding_strategy` for a pooled target. Consult `gads_policy_guardrail` before the session's first write. Preview (validate_only=true default), show old vs new target and the step plan if closing a gap, apply only after explicit approval. For full setup with learning-period management, hand off to set-up-conversion-based-bidding or set-up-value-based-bidding instead. ## Decision rules - **Never at or past breakeven.** A target equal to breakeven runs the account at zero retained profit; only a deliberate, time-boxed land-grab justifies it, and never silently. - **Reinvestment share bands:** 0.75 growth / 0.50 balanced / 0.25 efficiency. Default a new account to 0.50 and revisit after 2-4 weeks of data. - **Formula direction check:** CPA multiplies by the share, ROAS and profit-per-spend divide by it. If a "more efficient" target came out looser, the direction was flipped. - **Mixed-margin catalogs:** compute per category, or prefer profit-based values (which absorb margin variance automatically); for a single blended ROAS target use the sales-weighted effective margin. - **Optimization-stage consistency:** breakeven must be computed at the same event the campaign optimizes on (lead vs qualified lead vs closed deal). Mismatched stages are the most common silent target error. - **Feasibility outcomes:** volume sufficient at target -> done. Volume short -> raise the reinvestment share (accept thinner margin), expand targeting, or lower the growth goal, never quietly push the target past breakeven. Budget insufficient -> allocate-budget-across-campaigns. No combination works -> the problem is upstream (economics, market size, goals), not bidding. - **Refresh triggers:** any input change (pricing/margin/close-rate), goal change, sustained divergence from target, and a quarterly review regardless. ## Common failure modes - **Client-stated margins taken at face value.** Margins are systematically overestimated; verify against backend financials before they become bid targets. - **Revenue-ROAS ignoring order expenses.** Shipping, payment fees, and returns quietly eat the margin the ROAS math assumed; always use effective margin. - **Stage mismatch.** Breakeven computed at closed-deal level while the campaign optimizes on raw leads (or vice versa) mis-scales the target by the funnel rate. - **Skipping feasibility.** A correct target can still be unachievable at needed volume; the forecast check turns that discovery from a month of bad performance into a planning conversation. - **Set-once targets.** Economics drift; a target older than a quarter is a rumor. Schedule the recalculation. ## Related skills - Before: calculate-and-validate-unit-economics (inputs), set-up-cart-data-and-profit-tracking (enables profit-based targets). - After: set-up-conversion-based-bidding / set-up-value-based-bidding (implementation), set-up-portfolio-bid-strategies (pooled targets), allocate-budget-across-campaigns (budgets use these targets), scale-bids-and-budgets (later target iteration).