--- name: launch-standard-shopping-campaign description: >- Launches a Standard Shopping campaign, the maximum-control variant with manual or smart bidding, campaign priorities, full search-term visibility, and hand-built product-group trees. Use it for a first Shopping launch, for accounts under 30 conversions/month where Manual CPC is the only sane option, or whenever the operator needs bid-level control and query transparency that PMax hides. It builds campaign, ad group, product groups with per-group bids, bidding strategy chosen by conversion volume, budget, brand/universal negative lists, and runs the day-3 search-term review. Do not use it when hands-off automation is wanted at 30+ conversions/month (launch-pmax-feed-only-campaign), for cross-channel creative reach (launch-pmax-full-assets-ecommerce-campaign), before the feed is ready (set-up-and-optimize-product-feed), or for tiered bucket design (set-up-performance-based-shopping-segmentation). --- # Launch a Standard Shopping Campaign ## Purpose Create a Standard Shopping campaign with the control levers PMax removes: per-product-group bids, campaign priority for query sculpting, full search-term reporting, and transparent auction behavior. This is the right chassis for low-volume accounts, for operators who want to learn product-level truth before automating, and for brand/non-brand architectures that need negative keywords to steer queries. ## When to run - First Shopping campaign for a store with an approved feed. - Conversion volume under ~30/month, too thin for PMax or tROAS, fine for Manual CPC. - The account needs query sculpting (priority tiers + negatives) or brand/non-brand separation. - A segmentation project needs Standard Shopping bucket campaigns per tier (called from set-up-performance-based-shopping-segmentation). ## When NOT to run - 30+ conversions/month and the owner wants automation over control -> launch-pmax-feed-only-campaign. - Creative-driven cross-channel reach is the goal -> launch-pmax-full-assets-ecommerce-campaign. - Feed unapproved or unoptimized -> set-up-and-optimize-product-feed first. - Conversion tracking unverified -> set-up-google-ads-conversion-tracking first. - The campaign exists and needs maintenance -> run-shopping-campaign-optimization-cycle. ## Prerequisites - Merchant Center linked to the Google Ads account (this skill can create the link), feed approved with no critical errors. - Verified conversion tracking; value tracking if tROAS is on the roadmap. - Daily budget number, and a ROAS/CPA goal to pick the bid strategy against. - Brand terms listed if brand/non-brand separation is planned. - Structure decision made: single campaign to start (default), two for brand split, three-tier priority stack only for deliberate query sculpting. ## Procedure 1. **Session guardrail.** Before the first Google Ads write of this session, call `gads_policy_guardrail` and honor its output. 2. **Verify the Merchant link.** If Merchant Center is not yet linked, call `gads_link_merchant_center`, preview first (validate_only=true default), confirm the merchant ID, apply after explicit user approval. **HUMAN STEP:** approve the link from the Merchant Center side if it requires confirmation. 3. **Plan the build.** Call `gads_plan_shopping_campaign` to assemble the plan: naming (encode country/language/segment), country of sale (MUST match the feed's target country), campaign priority (low unless sculpting), location targeting restricted to shippable regions with presence-based matching, and the bid strategy from the Decision rules volume table. Present the plan to the user. 4. **Create the budget (write).** `gads_create_budget`, daily amount derived from revenue goal / target ROAS where history exists, otherwise an agreed test budget. Preview first, apply after approval. 5. **Create the campaign (write).** `gads_create_shopping_campaign` per the plan. Preview first, apply after approval. Keep priority LOW for a single-campaign setup; High/Medium only exist to route queries across multiple campaigns. 6. **Create the ad group (write).** `gads_create_ad_group` named for the segment (Shopping needs no ad creative, listings render from the feed). Preview first, apply after approval. 7. **Build the product-group tree (write).** With `gads_create_product_group`: start from all products, then subdivide to match the catalog's logic, typically brand then product type (or the reverse), down to item ID for products that will need individual bids. If this campaign is a segmentation bucket, subdivide by the custom label instead and exclude foreign tier values. Preview first, apply after approval. 8. **Set per-group bids (write, Manual CPC only).** `gads_update_product_group_bid` to price each group by its expected value, higher on margin-rich or proven groups, conservative on unknowns. Preview the full bid table, apply after approval. 9. **Set the bidding strategy (write).** `gads_set_campaign_bidding_strategy` per the volume table in Decision rules. For tROAS entry, start at trailing actual ROAS minus 10-20%. Preview first, apply after approval. 10. **Attach negatives (write).** `gads_create_shared_negative_set` (or reuse existing lists), then `gads_add_keywords_to_shared_set`: a brand list (brand terms as broad match, catching misspellings and brand+product combos, applied to non-brand campaigns only) and a universal waste list (wrong-intent and not-carried-competitor terms). Link with `gads_link_shared_set_to_campaign`. Preview first, apply after approval. 11. **Pre-launch audit and enable.** Read the build back: country of sale matches feed, priority correct, locations right, product-group tree complete with no orphaned "everything else" surprises, bids set, negatives linked, budget attached. Enable only after user sign-off. 12. **24-48h verification.** Via `gads_run_gaql_query`: campaign eligible, products approved and serving, impressions starting, budget pacing sanely. Missing products trace back to feed diagnostics (HUMAN STEP in Merchant Center), not campaign settings. 13. **Day-3 search-term review.** Call `gads_get_search_terms_report`; negate clearly irrelevant queries via `gads_add_negatives_from_search_terms`, preview the term list, apply after approval. This first pass catches the worst mismatches before they eat the first week's budget. 14. **Handoff.** Day 7: first performance read. Days 14-30: watch bid-strategy learning if on Smart Bidding. From there, run-shopping-campaign-optimization-cycle owns the routine; when volume reaches 50+/month on Manual CPC, revisit step 9 for a tROAS migration. ## Decision rules - **Bid strategy by monthly conversions:** <30 -> Manual CPC (works at any volume, teaches product-level truth); 30-50 -> Maximize Clicks to build history (or Maximize Conversion Value if values are reliable); 50+ -> Target ROAS at trailing actual minus 10-20%. - **Budget formula:** daily budget = daily revenue goal / target ROAS (e.g., 500/day revenue at 400% -> 125/day budget). - **Country of sale is a hard constraint:** it must equal the feed's target country or products simply will not serve. - **Priority discipline:** single campaign -> low priority, always. High/Medium exist solely to decide which of several campaigns enters the auction for a query (sculpting with shared negatives), not to signal importance. - **Location matching:** presence-based, not presence-or-interest, Shopping intent from outside the shipping area is waste. - **Search Partners:** vertical-dependent; leave on initially, judge from partner-segmented data after a few weeks rather than by doctrine. - **Structure ceiling:** start with one or two campaigns. Three-tier priority stacks are earned once the simple build's search-term data proves the need. - **Negative match style:** brand negatives broad match (max variation catch); waste patterns phrase match; exact only for surgical single-query blocks. ## Common failure modes - **Country-of-sale mismatch.** Campaign looks perfect, zero impressions, hours lost, the feed targets a different country. Check it in step 3, not after launch. - **One giant "all products" group forever.** Every product gets the same bid, winners subsidize losers, and nothing is learnable. Subdivide at least one level at build time. - **Priority misunderstood as importance.** Setting High on the only campaign changes nothing except confusing future sculpting. Low priority for singles. - **Brand negatives forgotten on the non-brand campaign.** Brand queries flood in, ROAS looks heroic, and the brand campaign starves. Attach the brand list at build time (step 10). - **Skipping the day-3 review.** The first search terms are the cheapest lesson the campaign will ever offer; missing them means paying for the same lesson for weeks. - **Premature tROAS.** A target on 20 conversions/month oscillates and chokes volume. Respect the volume table; Manual CPC is not a failure state. ## Related skills - Before: set-up-and-optimize-product-feed, set-up-google-ads-conversion-tracking. - Called by: set-up-performance-based-shopping-segmentation / set-up-scoring-model-segmentation (per-bucket builds). - Alternatives: launch-pmax-feed-only-campaign (automation at 30+ conv/month), launch-pmax-full-assets-ecommerce-campaign (cross-channel creative). - After: run-shopping-campaign-optimization-cycle (routine care), analyze-search-term-reports (deeper query work), select-a-bidding-strategy / migrate-from-manual-to-smart-bidding (when volume outgrows Manual CPC).