**TL;DR —** A single-location shop can survive with one person checking Google Business Profile (GBP). A clinic group, salon chain, or multi-city trades brand cannot. Volume multiplies by location count; brand risk multiplies when the wrong voice replies to a 1-star review. The durable pattern is **per-location GBP credentials**, a shared **n8n router**, **AI drafts for 4–5 star reviews**, and **human gates for 1–3 star reviews**—not hiring a full-time agency per city. --- ## Why multi-location review ops break | Failure mode | What happens | Business cost | | :--- | :--- | :--- | | One shared Google login | Token chaos; wrong location gets the reply | Brand confusion, policy risk | | One person, five cities | SLA collapses after ~50 reviews/week | Late replies; ranking decay | | One generic AI prompt | Same reply tone in every market | Sounds fake; misses local services | | No negative queue | Auto-publish angry 1-star replies | Public fights, legal exposure | | Burst campaigns per franchisee | Spam filters + uneven velocity | Shadow-deleted reviews | Single-location foundations: [GMB automation for service businesses](/blog/google-my-business-automation-service-businesses) and [review velocity algorithm](/blog/google-business-profile-algorithm-review-velocity). --- ## Target architecture (multi-token n8n routing) ``` ┌─────────────────────────────────────────────────────────────┐ │ Location A GBP API token Location B token Location C │ └───────────────┬──────────────────────┬──────────────────────┘ │ │ └──────────┬───────────┘ ▼ ┌──────────────────────┐ │ n8n Review Ingress │ Poll / webhook per location │ + location_id tag │ └──────────┬───────────┘ │ ┌────────────┼────────────┐ ▼ ▼ ▼ rating >= 4 rating 1–3 blacklist hit │ │ │ ▼ ▼ ▼ Brand AI draft Hold queue Hold + legal flag auto-publish Slack/TG owner-only │ │ ▼ ▼ Social graphic Human approve / edit (optional) then publish via correct token ``` ### Core components 1. **Per-location OAuth / service credentials** Never reuse one refresh token across unrelated GBP listings if Google scopes or account structure separate them. Tag every event with `location_id`, `brand_id`, and `market`. 2. **Ingress workflow in n8n** Schedule or push-based fetch of new reviews. Deduplicate by `reviewId`. Write to a store (Supabase/Sheets) for audit. 3. **Brand handbook pack per brand (not always per location)** Tone, refund policy, signature, languages. Locations inherit brand pack; override only when markets truly differ. 4. **Conditional publish path** Same rules as [automate vs approve](/blog/gmb-review-reply-automation-what-to-automate): auto for 4–5★; hold 1–3★. 5. **Escalation channel** Slack or Telegram topic per region: `location_id`, star rating, review text, draft reply, Approve / Edit buttons (or deep link to ops UI). 6. **Request loop still required** High reply volume without [WhatsApp post-job requests](/blog/whatsapp-vs-email-review-requests) only defends existing reviews—it does not grow velocity per location. --- ## Volume math for clinics and salons Example planning table (illustrative): | Locations | Jobs or visits / mo / site | Request conversion 10% | New reviews / mo | Replies needed | | ---: | ---: | ---: | ---: | ---: | | 1 | 200 | 20 | ~20 | ~20 | | 5 | 200 | 20 | ~100 | ~100 | | 15 | 200 | 20 | ~300 | ~300 | At 300 reviews/month, manual copy-paste is a full-time job. Automation is not a luxury—it is the only way to keep **24–48 hour reply SLAs** and steady velocity without a large agency retainer. --- ## Multi-location playbook (implementation steps) 1. **Inventory every GBP listing** Place IDs, categories, primary phone, and who owns the Google account. 2. **Map location → credential → brand pack** Store in config, not hard-coded nodes. 3. **Stand up n8n review ingress** One workflow family; parameters by location. 4. **Train AI on brand handbook** Include forbidden claims (medical outcomes, guaranteed results, competitor bashing). 5. **Wire Slack/Telegram for 1–3★** On-call rota per region if you operate across time zones. 6. **Add request automation per location stack** SimPRO/Jobber/Square/clinic PMS → delay → WhatsApp template with that location’s review URL. 7. **Dashboard weekly velocity per location** Flag locations with <2 reviews/week for ops intervention ([drip vs burst](/blog/google-business-profile-algorithm-review-velocity)). --- ## What to automate vs escalate at scale | Event | Automation | Human | | :--- | :--- | :--- | | 5★ with text | Draft + publish | Spot-check weekly | | 4★ with mild critique | Draft + publish | Optional QA sample | | 1–3★ | Draft only | Always approve | | Staff named in complaint | Draft only | Manager + HR path | | Legal / safety keywords | Freeze publish | Owner / counsel | | Review removal candidates | Never auto-appeal | Policy checklist first | Deep negative-review protocol: [Negative Google review defense](/blog/gmb-review-reply-automation-what-to-automate#negative-google-review-defense). --- > **⚠️ Calculate Your Review Leakage:** > *Five locations × 40 completed visits each × 2 reviews total = you are not running a reputation system—you are running a lottery.* > **[Request a free multi-location reputation audit →](https://vijayatechlabs.com/reputation-management?utm_source=blog&utm_medium=cta&utm_campaign=blog-reputation&utm_content=multi-location-google-review-automation#free-audit)** > *(1-page PDF in about 48 hours.)* --- ## Commercial box: agency vs productized multi-location | Model | Cost shape | Fit | | :--- | :--- | :--- | | In-house community manager | Salary + tools | High brand control; slow to scale | | Traditional agency | ~$1,500+/mo per brand | Heavy meetings; variable tooling | | **VijayaTech productized engine** | **$180/mo** single location · **from $500/mo** multi-location | Fixed scope: requests, AI drafts, holds, reporting | Product: [Google Review & Reputation Management](https://vijayatechlabs.com/reputation-management). --- ## Frequently Asked Questions ### Can one n8n instance serve all locations? Yes. One orchestration layer with per-location credentials and config is the standard pattern. Separate n8n instances only when compliance or client isolation requires it. ### Do multi-location brands need different WhatsApp numbers per site? Often yes for local trust and template quality. Some groups use a brand number with location variables. Choose based on Meta portfolio structure and customer expectation. ### How many reviews per location per week should we target? Plan **2–4 new reviews per location per week** for active markets, adjusted for visit volume. Chase steady drip, not end-of-month bursts. ### Does this replace local SEO content work? No. Reviews are a prominence and trust signal. You still need correct categories, NAP consistency, and service pages. Reviews compound those foundations. --- ## Sources - Google Business Profile Help, *How to improve your local ranking on Google* — https://support.google.com/business/answer/7091 - Google Business Profile API documentation — https://developers.google.com/my-business - BrightLocal, *Local Consumer Review Survey* — https://www.brightlocal.com/research/local-consumer-review-survey/ - VijayaTech Labs Reputation Management — https://vijayatechlabs.com/reputation-management