Tell a local business owner the three things to do this week, priced in jobs, and prove in 30 days that they worked.
For owners of local service businesses with one to five locations, and the agencies that run their profiles. The job is more calls from the map pack without learning SEO. The one idea: every finding from the profile, the website, the reviews, the photos and the listings lands in one queue sorted by average job value times the extra jobs the fix should bring. Nothing writes to the Business Profile; every change is copy-paste or a dated proposal the owner confirms. The example throughout is Brightwater Plumbing, a fictional Austin plumber.
How it works
- A free scan from the business name, no login. It shows the map pack top 3, a 9-point grid (top 3 at 4 of 9 points), the category the pack uses, the website field pointing at the homepage, 15 reviews this quarter against 48 for the leader, and a headline of about $4,550 a month of jobs going to the three businesses above (7 jobs at the plumbing default of $650 a job; estimate, default rates). Three fixes are shown in full; the rest sit behind the plan.
- The owner enters an average job value per service ($2,400 for a water heater, $9,500 for a repipe), and imports or types the jobs. The queue re-sorts: 14 open actions, 6 of them S tier, $26,200 per 28 days across 5 priced fixes (actions 4 and 9 share one job, so it is counted once).
- The profile audit compares against the pack, not a blank field. Primary category is the first card. 16 checks, 8 fail. The website field has to point at /locations/austin on the canonical host. A suspension-risk panel runs the website cross-checks. A change monitor shows Google's own edit to Saturday hours and never reverts it.
- One service, one page. The money pages map shows the blog that ranks 9 for the water heater query, the position-6 query that becomes a "cost" section instead of a new page, and the internal link to add with both URLs and the anchor. A city page publishes only with two real facts; Pflugerville passes, Leander fails.
- After the job: reviews and photos. Velocity against the pack with a 21-day warning and a 30-day alert. A reply queue with a 48-hour timer and an AI-phrase lint. Asks go out once, with one follow-up, with no incentive. A tech uploads real job photos from a phone; the profile is nagged at 14 days; generated images stay off.
- Listings are a short list and a diff. 8 listings checked, 3 fail, 2 missing. Link gaps are classified and drafted; a person sends them. Competitors sit side by side on the fields that matter.
- Measure and prove. A weekly 9-point grid, one diagnosis per keyword (the page is the gap, the profile is, or both are weak), a monthly AI-visibility log per engine, profile outcomes that stay empty until the real API is connected, and a 30-day proof report the owner can show a partner or an agency can send a client.
What it is and isn't
What's in it
- Free scan and paywall on the work
- Money-ordered weekly queue with account-manager approvals
- Profile audit vs the pack, suspension risk, change monitor, appeal pack
- Service-to-page map, striking distance, cannibalisation, internal links, city-page gate, builder work order
- Review velocity, reply timer with lint, review mining to FAQ, deleted-review watch, ask clock
- Photo inbox with a tech phone upload, posts behind an approve gate
- NAP diff across 8 listings, link gaps, competitor table
- Geo grid, pack vs organic, AI visibility log, 30-day report
- Connections: profile read, performance API, Search Console, GA4, jobs CSV, job values
Not in this version
- Writing to the Business Profile (proposals only until the API is approved)
- Sending asks, replies or outreach from the product (copy and WhatsApp hand-off)
- Keyword in the name, blanket "show your address", bought links, press-release syndication
- City-page templates, 40-site directory blasts, listings shown as "linked" to Siri or Uber
- Autopilot posting and generated images on by default (switched off, O8 and O9)
- A single SEO score to maximise, credits for AI actions, call tracking on the profile
- 169-point grids by default (paid add-on)
Release order
| Phase | What ships |
|---|---|
| A | The queue (AQ01, AQ02), job values and jobs import (DC03, DC04), the S-tier audit cards (PA01, PA02, PA06, PA05), the money pages map with cannibalisation and internal links (MP01, MP03, MP04), striking distance (MP02). Free scan with public data only (CV01 to CV04). |
| B | Reviews: velocity, reply queue with lint, ask clock, mining (RV01 to RV08). Photo inbox and tech upload (PH01, PH02), post approve gate (PH03). Listings diff and competitor table (NL01, NL02, NL04). Change monitor and suspension risk (PA03, PA04). |
| C | Geo grid pilot and pack vs organic (ME01, ME02), 30-day report and agency share link (ME04, CV05), AI visibility log (ME05), link gaps and mentions (NL03), city-page gate and work order (MP05, MP06), performance API connection and outcomes (DC02, ME03), approvals view (AQ03), appeal pack (PA07). |
Decided (1 Nov 2026, recommendations accepted)
- Is the money-ordered queue the Dashboard's first tab? Recommended: yes, replacing the Overview audit card; feature pages keep their detail (O1, O20).
- Job value before signup: industry default or ask on the scan? Recommended: a default per trade ($650 for plumbing) with an inline edit and the formula visible; the owner's values replace it after signup (O11).
- Does the pack-vs-organic diagnosis wait for the grid? Recommended: ship the organic half on Search Console now, with the pack column marked "not tracked yet" until the 9-point pilot runs (O3).
- Who owns the service page contract, Local SEO or the website builder? Recommended: Local SEO owns the audit and the work order; the builder generates (O4).
- Tech photo upload: our app or WhatsApp? Recommended: both into one inbox; the app screen ships first, a forwarding WhatsApp number in week two (O5).
- Autopilot posting and generated images are on today. Recommended: review mode and inbox photos become the defaults; existing autopilot users get a banner and a one-click switch (O8, O9).
- Pricing shape? Recommended: one price per location covering S and A plus the 9-point grid, no credits; larger grids and white label as priced add-ons (O14).
- AI visibility: our prompt runner or a vendor feed? Recommended: our fixed 10 prompts, three engines, monthly, per engine only; cited domains feed the listings list (O15).
- Review target: fixed number or pack median? Recommended: the gap to the pack median per month, capped by the ramp guard so asks never spike (O17).
All decisions O1 to O20 are recorded in REQUIREMENTS.md; every recommendation was accepted on 1 Nov 2026. Sources: the practitioner notes from X (Oct 2026), four Grok research runs (about 290 posts), and a public and logged-in teardown of RightChoice.ai (48 screens).