The short answer
DuckDive reads each location's Local Falcon results through one tool, get_local_rankings. Ask without a keyword and it gives an average rank and a top-3 share per tracked term; name one and the latest point-by-point grid comes too. Your AI agent sets those beside the same website's GA4, Search Console and Ads data. Scan history, trend charts and new scans stay in Local Falcon.
- Top-3 share is the nearest thing to Local Falcon's SoLV; check which of ARP or ATRP the average rank should match before you compare reports.
- Read a grid by comparing points that sit the same distance from the pin. Distance cannot explain why one corner loses and its mirror wins.
- Same keyword spelling, grid size and radius at every location, or a cross-location table compares nothing.
Local Falcon's metrics and what DuckDive returns
A Local Falcon scan report is dense: per-point ranks, two averages, a share score, a benchmark against the strongest rival in the scan, and an AI analysis. DuckDive hands your agent a smaller slice. When a regional manager's Local Falcon PDF and your agent's answer differ by a decimal, this table explains why.
| Local Falcon term | What it means there | Through DuckDive |
|---|---|---|
| ARP | Average position among the top 20 visible results | Partly, one average rank per keyword |
| ATRP | Average position across all grid points, beyond the top 20 too | Partly, same single average rank |
| SoLV | How often you rank in the top three, scored out of 100 | Yes, as top-3 share per keyword |
| Grid point ranks | Your position at each point; 20+ means not found there | Yes, latest grid when you name a keyword |
| Top SoLV, Competition SoLV, OSoLV | The top competitor's SoLV, how many competitors are active, and your gap to the leader (OSoLV) | No, your own listing only |
| Trend Report | Charts across two or more scans with identical settings | No, Local Falcon is the better tool here |
Sources: How To Use ARP & ATRP, What Is SoLV?, How To Read a Scan Report and Trend Report, Local Falcon, checked November 2026.
- ARP (Local Falcon)
- Average Rank Position, counted over only the top 20 visible results in the scan.
- ATRP (Local Falcon)
- Average Total Rank Position, across every grid point, including positions outside the top 20. ATRP well above ARP means you do well near the center and fade with distance.
- SoLV (Local Falcon)
- Share of Local Voice: how often the business lands in the top three, out of 100 possible. A score of 0 means it never reached the 3-pack in that scan.
- Average rank (DuckDive)
- One average position per tracked keyword, from the latest scan, as
get_local_rankingsreturns it. - Top-3 share (DuckDive)
- The share of the latest scan's points where the listing ranked first, second or third, per tracked keyword.
- Latest grid (DuckDive)
- Point-by-point ranks from the most recent scan, returned only when you name the keyword.
Before you compare
A gap of a few tenths between your agent's average and a Local Falcon report is usually ARP versus ATRP, not a sync fault. Ask how points marked 20+ were counted.
Latest scan only: what that means for trend questions
Most DuckDive data tools take a fixed window, up to the last 12 months; get_local_rankings takes none. It returns where each keyword stood on its most recent scan, and nothing earlier. That decides which questions your agent can settle alone.
Questions a single scan answers well
- Which locations are weakest on a keyword right now, and where on the map the weakness sits.
- Which keywords sit below a top-3 share you set, at any location.
- Whether today's weak area matches a page or query that also gets few clicks in Search Console.
Questions that need Local Falcon's history
- Did Wexley climb since the spring? Local Falcon builds a Trend Report once two or more scans share the same keyword, location, grid size, radius and platform.
- Did a competitor overtake us last quarter? Competitor trends live in the same report.
The bridge is a file. The Trend Report page offers a CSV download of the trend data. Attach it to the chat and the agent can line the history up against today's figures, after checking DuckDive's latest numbers match the CSV's last row.
Ask your agentI've attached Local Falcon's Trend Report CSV for Wexley on "brake repair". Check that DuckDive's current top-3 share matches the CSV's last scan, then tell me whether the last four scans trend up, sit flat or slide.
Tools it calls search get_local_rankings
Know your campaign schedule too. If scans run every four weeks, the latest grid can be four weeks old next to a GA4 window ending today. Read it as the month's state of play, not yesterday's.
Worked example: reading one grid, quadrant by quadrant
A grid of colored dots invites a gut call. A slower read finds the cause faster. Google's Business Profile Help names relevance, distance and popularity as the main inputs to local results. A grid lets you hold distance still and look hard at the other two.
- Find the pin. The center point is the business. The points on its row and column (the center cross) are the closest to it.
- Split the rest into four corners. On a 5 x 5 grid, each corner holds four points, and all four corners sit at exactly the same distances from the pin.
- Compare corners with each other. If one corner fails and its mirror wins, distance is not the cause. Relevance or a stronger rival in that area is.
- Then click the bad points in Local Falcon. Clicking a point in its scan report lists all the competitors found there. DuckDive does not carry that list.
Ask your agentShow the latest grid for 'tire shop' at the Greyfriars location and describe which quadrant ranks worst.
Tools it calls list_websites get_local_rankings
Show the latest grid for 'tire shop' at the Greyfriars location and describe which quadrant ranks worst.
list_websitesfinds the Greyfriars website and confirms Local Falcon is connectedget_local_rankingsGreyfriars, keyword "tire shop": average rank, top-3 share and the latest 5 x 5 grid
Greyfriars reaches the top three at 11 of 25 points (44%), with an average rank of 5.4. North is at the top; the dark-ringed dot is the shop, and pale rings mark the center cross.
| Area | Avg rank | In top 3 |
|---|---|---|
| Center cross | 3.2 | 5 |
| Northwest | 2.0 | 4 |
| Southeast | 5.8 | 1 |
| Southwest | 6.3 | 1 |
| Northeast | 12.3 | 0 |
Northeast is the worst quadrant: none of its four points reach the top three, and its far corner shows 20+ (I counted it as 21, so the real average there is no better than 12.3). Northwest sits at the same distances from the shop and holds the top three at all four points, so distance does not explain the northeast. Someone stronger, or more relevant to "tire shop", owns that side. Open the 20+ point in Local Falcon to see who.
Four patterns cover most grids you will meet, and each points somewhere different:
- An even ring that fades outward. Distance doing its job. Fine if the outer ring is beyond where you serve.
- One wedge or one side. Like Greyfriars' northeast: a rival sits there. Open those points before touching the website.
- A grid stuck at 4s and 5s. Distance is fine and the listing is close but not chosen. Look at relevance (does the site cover the service?) and at prominence, which Google's help article ties to how many reviews the business has and how many sites link to it.
- A checkerboard. Top three beside 15th with no geography behind it. Rerun before acting.
If the question is whether the map pack and the regular results moved together, the walkthrough on telling a map pack drop from an organic one picks up where this grid stops.
Find the weak corner at every location, not just one
Connect each location's website and Local Falcon, then ask your agent which quadrant loses at each one, keyword by keyword.
Free for 2 sites · No credit card
Choosing keywords that make locations comparable
A table of locations ranked by top-3 share is only fair if every row measured the same thing. Most broken comparisons trace back to keyword lists that grew one location at a time, each manager adding terms in their own words.
- One core set, spelled once. Pick the service terms every location offers and type them identically everywhere. "tire shop" and "tire shops" are two keywords to Local Falcon and two rows to your agent; so are "tire" and "tyre".
- Local extras stay out of the comparison. A location with a quick-lube bay can track "oil change", but leave it out of chain-wide rankings.
- Mix one broad term with specific ones. Local Falcon's SoLV guide warns that broad, general keywords carry the highest competition, and that longer, more specific phrases are easier to win. A broad term shows the ceiling; the specific ones show progress.
- Extras skew Local Falcon's own rollup too. Its Location Report appears once a location has been scanned for at least two keywords and averages ARP, ATRP and SoLV across all of them. A location with one extra keyword gets a different average, so per-location rollups only line up when the scanned sets match.
The rules hold outside auto repair. A roofer whose demand swings after storms needs a core set that stays put all season; the roofing analytics page covers that rhythm.
Run a keyword-coverage audit
Before any cross-location table, audit the lists themselves. Start by asking which sites have no Local Falcon data and whether the rest share one spelling:
Ask your agentWhich of our location websites have no Local Falcon data at all? For the rest, do they all track the same core keywords, spelled exactly the same way? List any near-duplicates like plurals or UK spellings.
Tools it calls list_websites get_local_rankings
At the fictional tire chain, that question returned this:
| Location | Keywords tracked | Flag |
|---|---|---|
| Greyfriars | tire shop, brake repair, wheel alignment, oil change | "oil change" tracked here only |
| Wexley | tyre shop, brake repair, wheel alignment | "tyre" breaks the match; found under the old wexleytyres.example site |
| Brindle Park | tire shop, brake repair, wheel alignment | None |
One rename and one exclusion. Change Wexley's keyword to "tire shop" in Local Falcon; until then, "tire shop" compares two locations, not three. Keep "oil change" out of chain tables, and expect Greyfriars' Location Report to average four keywords against the others' three.
The same audit works sideways across an agency's book, one service at a time. This version is for a set of plumbing clients:
Ask your agentFor every location tracking 'water heater repair', list average rank and top-3 share, and tell me which keywords are tracked at only one location.
Tools it calls list_websites get_local_rankings
A keyword that only one client tracks is either a genuine specialty or a term nobody else thought to add. Either way it cannot appear in a cross-client ranking until a second location scans it.
Scans, trends and credits stay in Local Falcon
DuckDive reads results. It does not run scans, schedule campaigns, change a radius, or show who holds the 3-pack at a point. Scan settings, competitor reports, AI visibility scans and Trend Reports stay in Local Falcon. So does billing: Local Falcon's radius and grid guide says each data point uses one credit (as of November 2026), spent from your Local Falcon account when the scan runs. A 7 x 7 grid costs more than a 5 x 5 every time it runs, whoever reads the results.
Local Falcon has its own connector, listed in Claude's connector directory, made by Local Falcon and marked Anthropic verified (as of November 2026). Its setup guide says it can read scan reports, run new grid scans on demand and open trend and campaign reports.
- Use Local Falcon's connector when the whole question is about rankings: a fresh scan, competitors at a point, a keyword over months. DuckDive does none of those.
- Use DuckDive when the question crosses from the map into the website, such as whether a weak keyword lines up with few Search Console clicks, at every location at once.
- Use both if your agent client lets you add more than one connector. Ask Local Falcon's connector for the scan detail and DuckDive for the traffic side, in the same chat.
DuckDive is also the wrong pick if you need Google Business Profile insights (not a DuckDive source) or a white-label PDF for clients. For listing searches, views and actions, the Yext data source page covers what DuckDive reads for locations on Yext.
Worked example: one keyword, three locations, three stories
Average rank and top-3 share look like one fact said twice. Read together, they describe a grid's shape before you open it. Here is one keyword at all three fictional locations, on the same 5 x 5 grid and radius.
Compare "brake repair" at Greyfriars, Wexley and Brindle Park. What kind of problem does each one have?
list_websitespicks out the three Calder sites that carry Local Falcon resultsget_local_rankings"brake repair" at each location, with its latest grid, one call per siteget_search_console_pagesBrindle Park's top pages by clicks, last 90 days
Three different problems, and only Brindle Park's starts on the website.
| Location | Avg rank | Top-3 share |
|---|---|---|
| Greyfriars | 2.9 | 64% |
| Wexley | 7.6 | 48% |
| Brindle Park | 4.2 | 20% |
Greyfriars wins 16 of 25 points; the misses are 4s and 5s on the outer ring, all sides alike. That is distance, and it needs nothing urgent.
Wexley has the worst average but not the worst share. The western half is top three; the eastern half runs from 10th to 20+. A bad average with a middling share means a cliff, not a slump. Check who owns the east in Local Falcon.
Brindle Park rarely wins and rarely loses badly: 20 of 25 points sit at 4th or 5th, so distance is not the issue. Its top 100 pages by Search Console clicks include no brake page. Start with relevance: a real brake repair page.
Brindle Park's pattern is the one a website fix can move, so screen the chain for it:
Ask your agentWhich tracked keywords at Calder Tire & Brake have a top-3 share under 20%, and for each one, does that location's site have a page about it among its top Search Console pages?
Tools it calls list_websites get_local_rankings get_search_console_pages
Then follow strong keywords into the website. A keyword can own the map and still land visitors on a page that does not convert:
Ask your agentFor Greyfriars, put each tracked keyword's top-3 share next to GA4 sessions, conversions and conversion rate for the landing page that matches it, over the last 30 days. Flag any keyword with no matching page.
Tools it calls search get_local_rankings get_ga4_landing_pages
The Search Console data source page and the GA4 data source page list what those tools return.
Matching scans to websites
Local Falcon's unit is a saved location, meaning one business listing, plus the keywords scanned at it. DuckDive's unit is a website, holding the site's GA4 property, its Search Console site and any Ads accounts. A grid is only useful beside traffic when each scanned listing sits with the site its customers actually land on. Calder lines up one to one: three listings, three location sites.
Brands that run every location from one domain have a single DuckDive website, so check with us before pairing several scanned listings with it.
| Scanned listing | DuckDive website | Grid comes back? |
|---|---|---|
| Greyfriars | greyfriars.calder.example | Yes |
| Wexley | wexleytyres.example (retired) | Yes, beside an empty site |
| Brindle Park | brindle.calder.example | Yes |
Wexley is the trap. Its scans arrived all along, which is why the keyword audit found them, but under a website kept from before the rebrand. Every question that set Wexley's grid against GA4 came back with zero sessions, and the cause was the pairing, not the market. Once the listing sits under wexley.calder.example, the same questions start working.
When a location comes back with no grid
- No keyword was named. Without one,
get_local_rankingsreturns averages and shares only. Ask for the grid by keyword. - The wording differs. The grid exists only for the exact term scanned at that listing. Ask the agent to list the location's keywords first, then copy one.
- The listing was never scanned for that term. A keyword scanned at Greyfriars says nothing about Brindle Park. Add it to that location's campaign in Local Falcon.
- The listing sits under another website. Like Wexley above: the grid exists, but on a site you were not asking about.
- The import has not finished. Ask the agent for the Local Falcon source's sync state on that website; a source still importing or marked failing explains an empty answer.
Ask your agentFor each Calder website, list the Local Falcon keywords it returns and its GA4 sessions for the last 30 days. Flag any site that has a grid but almost no sessions, and any site whose Local Falcon import is still running or failing.
Tools it calls list_websites get_local_rankings get_ga4_overview get_sync_status
Not on DuckDive yet? Continue with Google, or paste this line into your agent and it adds each location's site and links the Google sources for you.
Paste this into your agentSet up DuckDive for my website. Follow https://analytics.duckdive.ai/agents.md
The grids take a separate route once the sites exist: open Integrations inside DuckDive, bring in the Local Falcon account that already runs your scans, and make sure each scanned listing lands beside its own location's site.
The multi-location overview shows rankings beside the other sources for a whole chain, and the Local Services Ads data source page covers the lead side for locations that buy those ads.
Scan cadence and keyword mapping
Because your agent sees only the newest scan, the schedule decides what "latest" means. Local Falcon's campaign scheduling guide covers most of it, and following it pays off in its trend charts and in your agent's answers.
Before you trust a cross-location answer
- Same weekday, every run. The campaign guide recommends scanning every week or every two weeks, and picking "every 4 weeks" over "monthly" so scans always land on the same weekday.
- Scan while every location is open. The same guide says to schedule for a time when all locations in the campaign are open.
- Baseline scans first. It suggests a few one-off "baseline" map scans to settle grid size and radius before you lock them into a campaign.
- Group by density. Separate campaigns for urban and rural locations let each use a fitting radius and spend credits where they tell you something.
- Drop points no customer stands on. Local Falcon lets you disable points over large bodies of water or other irrelevant ground, so dead points stop dragging averages.
- Log every settings change. A new radius or grid size breaks the Trend Report's same-settings rule, invisibly to your agent.
- Refresh after a big scan. An admin's agent can queue a fresh Local Falcon import for a site, though only one re-import per website fits in any 15-minute window.
Agencies juggling dozens of local clients will find the client-side version of this, from renewal calls to screening a whole book, on the local SEO agency page.
Local Falcon and DuckDive, question by question
Does DuckDive spend my Local Falcon credits?
Credits go when a scan or campaign runs, and DuckDive never starts one. Those scans run in your own Local Falcon subscription, which DuckDive does not provide; DuckDive picks up the finished results for each location's website.
Is top-3 share the same number as SoLV?
Both answer one question: at what fraction of grid points did the business make the 3-pack? Local Falcon scores SoLV out of 100, and DuckDive reports a share per tracked keyword from the latest scan. If the two disagree, check that you are looking at the same scan.
Why does my agent show no grid until I name a keyword?
By default get_local_rankings returns two numbers per tracked keyword, average rank and top-3 share. The point-by-point grid comes back only when the request names one keyword, so ask for it by the exact wording Local Falcon stores.
What does 20+ mean on a grid point?
In Local Falcon's scan report, 20+ at a point means the listing did not show up there at all. ARP leaves those points out and ATRP counts them, which is why the two averages drift apart on a grid with weak edges. Read a 20+ point as not found, never as a precise twenty-first place.
Do disabled grid points count toward top-3 share?
Local Falcon lets you switch off points over water or ground you do not serve. Before you rely on the agent's share for a location with disabled points, compare it once with SoLV in that scan's report; if they match, the switched-off points were left out of both.
Local Falcon has its own MCP connector. Why add DuckDive?
For ranking-only work, Local Falcon's connector is the stronger choice: it can run scans and read competitor and trend reports. DuckDive earns its place when the question reaches the website, putting each location's grid next to its GA4, Search Console and Ads data in the same chat.
Sources
- How To Use ARP & ATRP, Local Falcon Knowledge Base article KB25, checked November 16, 2026.
- What Is SoLV? How To Interpret and Use the Metric, Local Falcon Knowledge Base article KB32, checked November 16, 2026.
- How To Read a Scan Report, Local Falcon Knowledge Base article KB72, checked November 16, 2026.
- Local SEO Trend Report: Track Rankings Over Time, Local Falcon report types page, checked November 16, 2026.
- Tips to improve your local ranking on Google, Google Business Profile Help article 7091, checked November 16, 2026.
- How To Read a Location Report, Local Falcon Knowledge Base article KB75, checked November 16, 2026.
- How To Set the Map Scan Radius and Grid Size, Local Falcon Knowledge Base article KB48, checked November 16, 2026.
- Local Falcon connector, Claude connector directory listing, checked November 16, 2026.
- How To Use the Local Falcon Claude Connector, Local Falcon Knowledge Base article KB94, checked November 16, 2026.
- How To Schedule Scans Using Campaigns in Local Falcon, Local Falcon Knowledge Base article KB53, checked November 16, 2026.
Last reviewed . Numbers in EXAMPLE DATA blocks are fictional. Local Falcon is a trademark of its owner, named here only to describe compatibility.
