---
name: set-campaign-goals-and-kpis
description: >-
  Turns a stakeholder's business ambition into a written, numeric Google Ads goal with a three-tier
  KPI framework (primary targets, hard guardrails, diagnostic metrics) and then reality-checks that
  goal against search volume, impression-share headroom, and historical trend before anyone spends a
  cent chasing it. Reach for this at engagement kickoff after unit economics are validated, at the
  start of a quarter, whenever the stakeholder changes objectives mid-flight, or before expanding
  into a new campaign type or market. Do not use it to compute margins or break-even thresholds
  (that is calculate-and-validate-unit-economics), to decide which bottleneck to attack next (that
  is scaling-os), or to run an ongoing review against already-agreed targets
  (run-a-weekly-performance-review / run-a-monthly-performance-review). Read-only: this skill makes
  no Google Ads changes.
---
# Set Campaign Goals and KPIs

## Purpose

Most Google Ads accounts fail quietly because nobody ever wrote down what the account is supposed
to achieve in numbers. This skill converts a business objective into a platform-level goal
statement with explicit assumptions, picks the metrics that define success and the guardrails that
define failure, and pressure-tests the whole thing for feasibility using VigilDog read tools. The
output is a document the stakeholder has agreed to, every downstream optimization skill assumes
it exists.

## When to run

- Kicking off a new account or client engagement, immediately after unit economics are validated.
- A new quarter starts and business objectives have been revised.
- The stakeholder changes the goal mid-cycle (new revenue target, new market, budget change).
- You are about to add a new campaign type or geography and need its expected contribution defined.
- An existing goal keeps being missed and you suspect it was never feasible in the first place.

## When NOT to run

- You need break-even CPA/ROAS, margins, or LTV computed, run calculate-and-validate-unit-economics first; this skill consumes those numbers.
- You already have agreed targets and just need to check progress, run run-a-weekly-performance-review or run-a-monthly-performance-review.
- You are deciding which problem in the account to fix next, run scaling-os.
- You want to pick a bid strategy for a specific campaign, run select-a-bidding-strategy.
- You need to split budget across existing campaigns, run allocate-budget-across-campaigns.

## Prerequisites

- Validated unit economics: break-even ROAS or maximum viable CPA, margin, and (for lead gen) lead-to-sale rates. Without these, guardrails are guesses.
- Access to historical Google Ads performance (ideally 12+ months for seasonality; 90 days minimum).
- GA4 property linked and reporting, so channel contribution can be measured rather than assumed.
- A stakeholder available for a goal conversation, this skill cannot be completed unilaterally.

## Procedure

All steps here are reads or documentation, this skill performs no Google Ads writes. If the
session later moves into a skill that does write, consult gads_policy_guardrail before the first
write, and remember every write goes preview (validate_only=true) -> explicit user approval -> commit.

1. **HUMAN STEP (outside VigilDog): run the goal conversation.** Ask the stakeholder: what outcome
   (revenue, customers, qualified leads, MRR)? What target number and by when? What budget is
   available? What efficiency floor must hold (minimum ROAS / maximum CPA / margin)? Are there
   capacity limits (inventory, sales team bandwidth, fulfillment)? Record answers verbatim.

2. **Classify the goal.** If the emphasis is more volume/revenue/new markets, the primary type is
   growth and efficiency metrics become guardrails. If the emphasis is cheaper acquisition or
   better returns, the primary type is efficiency and volume metrics become guardrails. Never
   accept both as co-primary, one must yield (see Decision rules).

3. **Write a testable goal statement.** One outcome, a quantified target, a deadline, and at least
   one guardrail. Example shape: "Grow monthly Google Ads conversion value from $120k to $150k by
   end of Q3 while holding non-brand ROAS at or above 350%." Reject anything like "more leads" or
   "better ROAS", if it has no number, volume floor, and date, it cannot be diagnosed later.

4. **Pull the Google Ads baseline.** Call gads_get_account_performance for the trailing 90 days and
   the same period last year: conversions, conversion value, cost, CPA/ROAS, and the trend slope.
   This is the "current" side of the gap the goal must close.

5. **Measure channel contribution.** Call ga4_run_report with sessions and conversions broken out by
   default channel grouping over the last 90 days. Compute Google Ads' current share of total
   conversions/revenue. HUMAN STEP: decide with the stakeholder what share of the business goal
   Google Ads should carry, that allocation is a judgment call, not arithmetic.

6. **Translate the business goal into platform metrics.** Revenue maps to conversion value;
   customers or deals map to conversions (deals require offline conversion import to be visible, see set-up-offline-conversion-tracking); CAC maps to CPA; blended ROAS or MER maps to a stricter
   in-platform ROAS scaled by the channel-contribution share from step 5.

7. **Document every assumption.** For each of: channel contribution %, expected conversion rate,
   expected CPC range, and typical conversion lag, record the value and where it came from
   (baseline pull, auction data, GA4). These assumptions are what you audit when the goal is missed.

8. **Select campaign types.** Core intent-capture types (Search, Shopping, Performance Max) serve
   both goal types. Add reach types (Display, Video, Demand Gen) only when a growth goal exceeds
   what existing search demand can supply. For an efficiency-primary goal, add nothing beyond
   remarketing. Record each chosen type with its expected contribution.

9. **Build the three-tier KPI table.**
   - Tier 1 (primary, reviewed weekly): conversions / conversion value / revenue for growth goals;
     CPA / ROAS / cost per qualified lead for efficiency goals.
   - Tier 2 (guardrails, reviewed weekly, hard thresholds from unit economics): a growth goal gets
     a minimum ROAS or maximum CPA and a minimum margin; an efficiency goal gets a minimum monthly
     conversion count, minimum conversion value, and a minimum impression share.
   - Tier 3 (diagnostic, reviewed bi-weekly): impressions, CTR, CPC, conversion rate, AOV,
     impression share plus IS lost to budget and lost to rank, Quality Score.

10. **Feasibility check, volume.** Call gads_keyword_forecast on the core keyword themes. Compare
    forecast clicks x assumed conversion rate against the incremental conversions the goal
    requires. If forecast supply covers less than the required increment, the goal needs new
    inventory (new campaign types, markets, or keywords) or a smaller number.

11. **Feasibility check, auction headroom.** Call gads_get_auction_insights for the core campaigns.
    If your impression share is already above ~80%, growth cannot come from buying a bigger share
    of the same auctions; the plan must name where the extra volume comes from. Also note which
    competitors sit above you, rising competition invalidates optimistic CPC assumptions.

12. **Feasibility check, trajectory.** From the step 4 baseline, compare the trailing trend slope
    to the slope the goal requires. A goal needing +25% in one quarter from an account that has
    been flat for three quarters requires a named mechanism (new budget, new inventory, new offer),
    not hope. HUMAN STEP (outside VigilDog): cross-check with Performance Planner, Bid Simulator, and
    Google Trends seasonality in the Google Ads UI, no VigilDog tool exposes these.

13. **Red-flag screen.** Reject or revise the goal if any of these hold: it assumes conversions
    scale linearly with budget; it demands maximum growth and maximum efficiency simultaneously;
    required spend (incremental conversions x expected CPA, plus a 10-20% buffer) exceeds the
    available budget; the timeline ignores conversion lag or learning periods; or a KPI in the
    framework cannot actually be measured with current tracking (route to
    set-up-google-ads-conversion-tracking before proceeding).

14. **Stand up the tracking sheet.** Call sheets_export_report (or report_builder) to create a
    KPI-vs-target tracker: one row per KPI tier entry with target, threshold, and review cadence.
    HUMAN STEP (outside VigilDog): build a Looker Studio dashboard instead if the stakeholder
    prefers one.

15. **HUMAN STEP (outside VigilDog): agree the review cadence with the stakeholder.** Weekly for
    primary and guardrail KPIs, bi-weekly for diagnostics, monthly stakeholder review, quarterly
    full goal revision. Record the agreement in the goal document. Done means: goal statement,
    assumptions, campaign-type selection, KPI table, feasibility results, and cadence are all
    written down and acknowledged.

## Decision rules

- **One primary, one guardrail set.** Growth and efficiency cannot both be maximized; whichever is
  not primary supplies hard thresholds. If the stakeholder refuses to choose, the goal is not set.
- **Never target break-even.** Operating targets need a 10-20% cushion above break-even ROAS (or
  below maximum viable CPA) so normal variance doesn't tip the account into losses.
- **Volume floor for efficiency goals.** Keep at least ~30 conversions per 30 days per campaign for
  target-CPA bidding and ~50 for target-ROAS; an efficiency goal that starves the account below
  these levels degrades Smart Bidding and eventually efficiency itself.
- **Budget sufficiency test.** Required monthly spend = incremental conversions x expected CPA
  x 1.1-1.2 buffer. If that exceeds available budget, cut the target or raise the budget, do not
  proceed on the assumption CPA will conveniently fall.
- **IS ceiling.** Above ~80% impression share on core campaigns, treat existing auctions as tapped;
  incremental volume must be sourced from new keywords, campaign types, or geographies.
- **Lag-adjust the timeline.** If days-to-conversion typically exceeds 7, evaluation windows and
  the goal deadline must both account for the lag, or the final weeks of the period will look
  artificially bad.
- **Re-run triggers.** Stakeholder changes the goal: redo steps 2-15. Targets missed for two
  consecutive review periods: redo steps 10-13 and either adjust the goal or escalate. Channel mix
  shifts materially in GA4: redo step 5.

## Common failure modes

- **Platform targets set in a vacuum.** A CPA target chosen without the business goal behind it
  optimizes toward a number nobody actually needs. Always start at step 1, even when it feels
  redundant.
- **No guardrails.** Pure growth goals quietly destroy margin; pure efficiency goals shrink the
  account until the "great ROAS" applies to almost nothing. Tier 2 is mandatory, with numbers.
- **Break-even targets.** Leaves zero room for variance, testing, or learning periods. Apply the
  10-20% cushion rule.
- **Single-source validation.** A forecast alone ignores competition; auction insights alone ignore
  demand. Steps 10-12 are a set, run all three.
- **Set-and-forget goals.** Business conditions drift; a goal untouched for six months is usually
  fiction. The quarterly revision in step 15 is part of the skill, not optional hygiene.
- **Untrackable KPIs.** Agreeing to report on qualified leads when no offline import exists
  guarantees a credibility problem at the first monthly review. Step 13 catches this, act on it.

## Related skills

- Run before: calculate-and-validate-unit-economics (supplies every guardrail number).
- Run after: scaling-os (decide the first constraint to attack), select-a-bidding-strategy,
  allocate-budget-across-campaigns, then the relevant launch skills.
- Ongoing: run-a-weekly-performance-review, run-a-monthly-performance-review,
  run-a-quarterly-business-review (each reviews against the framework this skill produces).
- If tracking gaps surface: set-up-google-ads-conversion-tracking, set-up-offline-conversion-tracking.