Google Ads for multi-location & franchise: structure, attribution, and location truth

Direct answer — how should multi-location and franchise Google Ads work?

Give every location visible attribution (geo, calls, landers, CRM tags), separate brand and national demand from local non-brand where economics differ, stop one national Target CPA from hiding weak markets, and coordinate Local SEO / GBP so paid is not the only “local” system. Multi-location Google Ads fails when HQ runs one soup campaign, franchisees get unreadable lead dumps, calls route to the wrong store, and Performance Max claims every city while nobody knows which yard booked the job.

This insight owns Google Ads multi-location / franchise structure and attribution. Vertical siblings: home services, automotive multi-branch. Measurement: multi-path, call tracking, OCI. Architecture: intent account structure. Local sibling: US Local SEO, Local SEO service. Map readiness: Map Pack Readiness.

Working rule: If a location cannot see its own booked jobs and cost, you do not have a multi-location account — you have a national blender.

What “good” looks like after 60–120 days: each priority market has readable CPA/ROAS or cost per booked job; call routing matches the ad’s location promise; landers are location-true; franchisee reports stop being political fiction; PMax (if used) has market or region asset groups; Health Score recovers because soft HQ micro-conversions are gone.

If you only remember three moves: attribute by location, split national brand from local money themes, fix geo + call routing before “scaling AI” — or the AI will scale the wrong store.

Proof patterns: US home services multi-metro Ads (four metros, location-aware rebuild), Malaysia four-branch automotive, Singapore two-location automotive. Local SEO twin for franchise GBP networks: US franchise Local SEO.

What multi-location Google Ads actually buys

Operators in this class include:

• Corporate multi-branch (owned locations) — tint studios, clinics, home services metros.
• Franchise networks — HQ media + franchisee co-op, territory rules, politics.
• Service-area multi-van — fewer storefronts, still location economics.

They buy the same Google inventory as single-location SMBs — Search, LSA (where eligible), PMax, Demand Gen — but the failure mode is attribution bleed: strong locations subsidize weak ones; franchisees revolt; HQ “proves” national ROAS while local P&Ls scream.

Out of scope as full essays: full franchise Local SEO OS (US franchise case + Local SEO guides); Meta territory fences (Advantage+ compare); SaaS multi-region (different product — SaaS Ads).

The location truth stack (non-negotiable)

1. Geo settings that match dispatch

Presence vs interest, radius vs cities, exclude markets you do not serve. Loose “interest” targeting on multi-location accounts buys tourism and wrong-city leads. Tighten before debating bid strategies.

2. Call routing that matches the ad

Dynamic numbers per location (or per market campaign). Ads for Location B must not silently ring Location A’s front desk without a rule and a CRM tag (call tracking). Franchise politics invents fiction when national brand campaigns convert on a random store number.

GBP NAP: Keep stable local numbers on GBP/citations; use site DNI under ads. Do not train franchisees to put rotating tracking lines into GBP.

3. Conversion labels / actions by location or market

At minimum: CRM source fields for location ID. Prefer conversion actions or value rules that let reporting separate markets. Soft HQ events (“ebook,” “franchisee portal login”) must not be primary in the customer acquisition account (primary vs secondary, bad conversion data).

4. Offline imports with standardized stages

Franchise CRMs fragment stage names across locations. Standardize “Qualified / Job booked / Won” at HQ or map carefully into one Ads action set (OCI). Fragmented labels destroy match reporting and QBR discipline.

5. Deduplication across LSA, Search, PMax, and organic

One booked job must not become three primaries. Tag channel and location. Score Lead Path Coverage.

Self-check: Tracking Trust, Google Ads Health Score.

Structure models that survive QBRs

There is no single correct tree — but there are wrong ones.

| Model | When it fits | Failure mode | | --- | --- | --- | | Market campaigns (city/metro non-brand) | Different CPCs and margins by market | Too many micro-campaigns if you oversplit | | Shared brand + market non-brand | Strong national brand; local money themes | Brand starved while PMax “looks good” | | Region portfolios | 20+ locations; need manageability | Hides store-level losers if reporting stops at region | | Location asset groups in PMax | After Search foundations | One national PMax soup |

Working default for SMB multi-location: Fund brand (national or by brand+city where needed). Build non-brand by market or priority cluster. Shared negative lists at HQ. Location landers mandatory for local themes (intent structure).

Do not use one Target CPA across Phoenix and a low-CPC secondary market if allowable CAC differs by 2×. The US home services rebuild split by metro for this reason — Phoenix and Dallas earned budget after truth was visible.

Franchise co-op: Document who pays for brand vs local non-brand. Co-op funds without location attribution create free-rider fights.

Landing pages and location assets

• Location (or metro) pages with real NAP, maps, reviews, hours, and service proof — not doorway templates that only swap city names.
• Message match: “[City] emergency plumber” lands on that city’s emergency CTA (message match).
• Location assets / feed where used — keep address accuracy or Ads will send people to closed stores.
• Franchise template sites: allow enough local proof that conversion rate is not national-average fiction.

Doorway-page risk is a Local SEO and Ads problem. Corporate templates that swap city names without unique proof hurt both map pack and paid CR — see US franchise Local SEO.

Performance Max for multi-location

PMax after:

• Location-aware conversion truth
• Brand Search funded
• Search foundations for high-intent local themes in priority markets
• Asset groups by market/region (not one “all locations” black box)
• Brand exclusions where brand cannibalization is real

PMax before that stage averages markets and claims brand. PMax playbook, attribution inflation, PMax Readiness.

Location assets help Maps inventory — they do not replace geo hygiene or call routing.

HQ vs franchisee reporting

Franchisees need:

• Leads and booked jobs tagged to their location
• Cost and CPL/CPA they can reconcile to co-op invoices
• Spam/dispute context (especially LSA)

HQ needs:

• Roll-up by region + brand vs non-brand
• Health of measurement (match rates, soft primary purge)
• Which markets deserve scale vs Local SEO investment

Kill vanity national ROAS as the only slide. Strong brand Search in one state will subsidize weak markets forever if you only show the blender.

Weekly ops (45–60 min at HQ): search terms on shared lists; top/bottom markets by cost per booked job; answer-rate by location; Ads vs CRM reconcile; franchisee ticket queue on misrouted calls.

Co-op invoice fights: When franchisees dispute lead quality, the only defense is location-tagged CRM truth and call recordings (where lawful). “Google said 40 conversions” without store IDs is how networks lose trust in paid media for a year.

Budget governance

• Cap or pause markets where answer-rate SLAs fail.
• Move incremental budget to markets with proven cost per booked job — not to the loudest franchisee.
• Keep brand funded even when local non-brand is throttled.
• New openings get a measurement checklist before spend (lander, DNI, GBP, CRM location ID).
• Use break-even / allowable CPA math per market when margins differ.

National “always on” budgets without market kill switches are how weak yards burn co-op cash.

Local SEO coordination (paid is not Maps)

Multi-location paid without GBP discipline overpays for demand map pack should own. Franchise networks especially need per-location GBP, citations, and reviews — not one corporate profile (Local SEO, Map Pack Readiness, US franchise case).

Coordinate geo modifiers, service categories, and offers so Ads and organic are not blind rivals for the same “near me” query. When a market’s CPCs exceed efficiency, Local SEO is often the cheaper long-term owner of that intent — Ads should not be asked to replace a neglected GBP forever.

Vertical notes

Home services / trades: Emergency vs scheduled still applies inside each market (home services insight). LSA eligibility may vary by location — tag carefully.

Automotive aftermarket: Package tiers + branch pacing (automotive insight). WhatsApp-first markets need chat attribution per branch.

Clinics / regulated: Location landing compliance; call recording disclosures by market.

UK: Consent Mode and LSA overlap — LSA vs Search. Do not average LSA and Search into one national CPA.

Sequencing playbook (90 days)

1. Days 1–14 — Audit. Health Score; location bleed map; soft HQ conversions; geo/call routing diagram.
2. Days 15–35 — Measurement. DNI per market; CRM location IDs; OCI stages standardized; demote soft primaries.
3. Days 35–60 — Structure. Brand isolation; priority market non-brand; shared negatives; location landers.
4. Days 60–75 — Bidding. Smart Bidding per market or honest portfolios after clean history (Smart Bidding, learning phase).
5. Days 75–90 — Controlled PMax / budget shifts. Market asset groups; move budget to markets that book; Local SEO for markets Ads cannot efficiently own.

Opening a new location: tracking + lander + GBP live before first dollar of local non-brand spend. Do not “add a pin to the national campaign” and hope. The same rule applies when a franchisee sells or relocates — pause the old geo, retarget numbers, and retire obsolete location assets so Ads does not send traffic to a closed yard.

Proof patterns (published)

US four-metro home services (Ads case)

National-ish soup + incomplete calls → metro split, call + OCI primaries, LSA coordinated → 2.9× → 6.4× ROAS; Phoenix/Dallas earned budget once visible.

Malaysia / Singapore multi-branch automotive

KL tint and SG automotive: branch bleed fixed with DNI, package tiers, branch pacing — same location-truth principle.

US franchise Local SEO (case)

Twelve locations, broken GBP templates → per-location profiles; map pack top 3 in 8 of 12 priority metros in 90 days. Paid cannot paper over that forever.

Common failure patterns

• One national non-brand campaign for 12 locations.
• One Target CPA across unequal markets.
• Calls routing to HQ or wrong store.
• Rotating tracking numbers in GBP.
• Soft HQ micro-conversions as primary.
• PMax “United States” asset group for everything.
• Franchisee reports without location tags.
• Doorway city pages under Ads.
• Scaling spend when answer rates collapse at specific locations.
• Ignoring Local SEO while overpaying Search for map-pack queries.
• Enabling AI Max / broad account-wide before negatives and geo hygiene (AI Max).

How this fits the cluster

• This insight (CP-039): multi-location & franchise Google Ads OS.
• Home services / automotive: vertical depth inside location structure.
• SaaS (CP-038): different vertical — not a franchise substitute.
• Call / OCI / structure guides: horizontal system.
• Local SEO + franchise GBP case: organic twin.
• Section 1F Google verticals complete with this ship.

FAQ

How should franchises structure Google Ads?

Shared brand framework, market-level non-brand themes, centralized negatives, location landers, and location-tagged conversions. Avoid national soup.

Should each location have its own Google Ads account?

Sometimes for franchisee-owned spend and billing — often HQ MCC with location structure is cleaner for measurement. Choose based on billing and control, not fashion.

How do I stop location attribution bleed?

Geo hygiene, per-location call routing, CRM location IDs, and reporting that cannot hide losers in a national average.

Is Performance Max good for multi-location?

After Search foundations and location-aware goals — with market/region asset groups. Not as day-one national automation.

Do franchisees need their own landing pages?

Yes for local non-brand. Template + unique proof beats doorway swaps.

How does Local SEO relate to multi-location Ads?

GBP and map pack own much of near-me. Paid should coordinate, not replace. See the US franchise Local SEO case.

What Target CPA should multi-location use?

Per market (or honest portfolio scopes). One national tCPA hides 2× economics differences.

How do I report to franchisees?

Location-tagged leads, booked jobs, cost, and spam context — not a national ROAS screenshot.

Should brand be national or local?

Usually national brand (or brand+city where needed) plus local non-brand. Do not starve brand to inflate non-brand dashboards.

What about Location Extensions / assets?

Keep them accurate. Wrong addresses destroy trust and waste clicks.

How long to rebuild a multi-location account?

Expect 2–4 weeks for measurement/structure foundations, then 60–120 days for compounding — longer if franchise politics block stage definitions.

Can AI Max help multi-location Search?

Optionally per mature market after geo and negatives — not as an account-wide switch that ignores territory rules.

What to do next

1. Draw the current call + geo + conversion diagram for every location.
2. Score Health Score and Lead Path Coverage.
3. Standardize CRM location IDs and money stages; demote soft HQ events.
4. Split brand vs priority-market non-brand; fix location landers.
5. Add PMax only after readiness.
6. Parallel Local SEO for GBP networks (Map Pack Readiness).

Get a free Google Ads audit — we will map location bleed, franchise reporting gaps, and whether your national ROAS is a blender.

Multi-location Google Ads is a governance and attribution system that happens to buy auctions. When each location’s booked jobs are visible, Smart Bidding and budgets become rational. When they are not, every AI feature scales the wrong yard faster — and franchisees will tell you before the dashboard does. Fix location truth first; then scale what the scoreboard proves.