Search “plumber near me” on a phone in Austin, click an ad, and land on a page with Denver’s skyline and a 303 number. The keywords may be sorted and the bids automated, but the paid click is going to the wrong place. Most franchise Google Ads setups fail at the landing page, not the campaign. Before I let an AI tool touch a multi-location account, I want to know whether it can generate, route, and keep in sync a separate page per location per intent without a person rebuilding the set whenever an offer changes. “Does it do local?” tells me nothing. The cheapest tool that cannot handle the page work costs more than the expensive one that can. Run these 15 pass/fail checks in a live trial, not a polished demo.

A. Can you control each location without disturbing the others?

  1. Give each location a campaign and budget, or its own account under an MCC. Pass: pause one location’s spend without pausing the other nine, then pull a budget and performance view for that location alone. Fail: one national campaign held together by city settings and geographic bid adjustments, with no independent budget to defend. When franchisees fund their own marketing, separate accounts under an MCC are the cleaner cut; when corporate funds everything, a campaign per location in one account works. Settle who owns the account before building either structure. Within each location, split ad groups by service rather than multiplying every service into its own thin campaign; that was the practical conclusion for a 12-location salon account. The fuller structure decision is in the multi-location account guide.

  2. Set location targeting to Presence, not Presence or interest. Pass: open Settings → Locations → Location options in every campaign and confirm the setting. Google’s default includes people interested in a location, not only people in it. That can put a franchisee’s budget in front of searchers outside the territory they serve. Presence is not a perfect fence either; Google says targeting a city can include nearby suburbs. Its reported conversion gain for advertisers moving the other way in Travel, Real Estate, and Education is not a reason to let a local HVAC campaign spend across a territory line. Check the setting yourself. A vendor saying it “handles geo” is not the pass test.

  3. Check matched locations, then exclude other franchisees’ towns. Pass: open the Locations report, switch from Targeted locations to Matched locations, and compare where ads appeared with the territory map. If another franchisee’s town appears, the leak is visible; a bid strategy will not close it for you. Build location negatives for neighboring territories, including the suburbs and colloquial region names people use in searches, then check the report again. Do not stop after reading the list of places you intended to target. The matched view is the one that shows whether the account obeyed you.

B. Does the page match both the place and the job?

  1. Generate distinct pages for distinct locations. Pass: ask the tool to build pages for three locations during the trial and inspect three URLs populated from the relevant location records. Fail: a setup-hours estimate, a per-page build quote presented as automation, or one URL that swaps its headline according to the searcher’s city. One 17-location service franchise manager described managing hundreds of location-and-service pages by hand, with labels and filters to cope. That is work a page workflow must account for, not evidence that a single national page is enough. If the tool succeeds on the first location but needs a rebuild for the second, mark it fail.

  2. Test the location and intent axes. Pass: compare an Austin “emergency AC repair” page with an Austin “AC tune-up” page, then compare each with its equivalent in another city. The offer and service need to fit the query as well as the place. Franify describes a page per location, with a unique form and phone number; GenPage describes pages per ad group or keyword cluster. Each description addresses an axis I would test, not the whole franchise matrix. Fail any tool that leaves you rebuilding the other axis yourself. A changed city string alone is not real local substance; inspect the actual page before accepting a grid of URLs as a grid of useful pages.

  3. Pull NAP, hours, and offers from one location record. Pass: change one branch’s Saturday hours in the data source, then inspect its page and the other branches’ pages. The changed branch should update without making Denver inherit Austin’s schedule. Each record needs the details that differ locally: its contact information, priority service, offer, and page destination, rather than generic inputs copied across every city. A breakdown of clean location data for AI ads makes that distinction concrete with different offers and priorities across markets. Fail when a correct-looking page is maintained separately from the data everyone else uses. The expensive error is not ugly design; it is a customer calling when the page says the branch is open and it is not.

  4. Change one offer and time the sync. Pass: edit Austin’s offer in the data the tool reads, verify its page reflects the change inside a day, and confirm Denver stays untouched. Fail: a support ticket, re-export, or person editing ten templates. Try a service change or holiday hours too if those are the updates your team makes most often. Franchise offers drift: a coupon runs in three markets while another location drops a service. The location record should drive the page, not sit beside a document somebody must remember to fix. Ask who notices a failed update and where you can see that failure; a sync process you cannot inspect is another manual check waiting to be missed.

C. Can you trace a lead back to the right branch?

  1. Attribute conversions to the location that served the click. Pass: pull a week of conversions and confirm the serving location is identifiable for each one. Fail: a central lead form with no location field, or account-level reporting that cannot tell which branch earned a lead. Verify this before automation starts learning from the numbers; otherwise the next bid decision uses the wrong lesson. Multi-location tracking gets skipped precisely because a total conversion count can look reassuring. Check diversion paths as well. An operator with more than ten remodeling locations reported roughly 30% of clicks going through location-detail or directions actions rather than the landing page, without credit for the resulting conversions. Keep location assets, but measure what happens when they take the click.

  2. Split calls and forms by location. Pass: submit a form from each trial page and place a real test call from each one; confirm both reach the intended branch and carry its location into reporting. Fail: Dallas and Houston sharing a tracking number and collapsing into one bucket. Give each location dedicated call tracking across calls from ads and calls from the website, plus a location-specific form path. Set the call-duration threshold before launch: 60 to 90 seconds is a common bar for counting a call as a qualified lead. If you leave it unset, wrong numbers and hang-ups can teach the bidding system that the branch generated useful demand. A dashboard total is not a substitute for two test calls.

D. Can budgets move without letting one branch take everything?

  1. Make budget shifts explainable by location-level returns. Pass: ask for 30 days of spend and cost per qualified conversion, side by side for every branch, then ask what budget the tool moved and why. Fail: an even split that stays fixed despite different costs and demand. I do not want “optimization” as the explanation; I want the location, the metric, and the change. Ten equal slices can be a deliberate franchise policy, but they are not evidence that a tool allocates money by return. If one branch books jobs at a third of the neighboring branch’s cost, ask what prevents the next available dollar from going there. The answer may be a business constraint. It should not be that nobody checked.

  2. Cap each location and name the human who owns the cap. Pass: set a daily ceiling on one branch and watch whether it holds through a full week, including high-demand days. Fail: a shared budget that lets one metro consume a third of national spend after demand spikes. Check overlapping territories too; the 28-location franchise example shows how branches can spend against one another across their boundaries. A machine can execute within limits, but somebody has to choose those limits and answer when they need changing. Get that person’s name during the trial, not after a branch owner disputes the bill.

Miniature main street with one shop drawing from a large, kinked hose while nine others get thin straws

E. What does the full page matrix cost, and who keeps it?

  1. Price pages at your location and intent count. Pass: get a written price for the pages you actually need, including the variations tested in item 5. Fail: a low entry price that excludes programmatic generation, caps active pages below your count, or charges separately for each new intent. Landingi’s listed tiers put 10 active pages at $24 a month, 100 at $119, and programmatic generation with uncapped page volume in a Scale tier starting at $229. Instapage lists dynamic text replacement and ad-to-page personalization at $99 and $199 a month, rather than on its entry plan. Those are pricing questions to put to any vendor, not proof that a text swap solves item 5. A 50-location franchise with multiple intents should compare the whole page bill with a flat fee that includes the work.

  2. Ask what happens to the fee when location eleven opens. Pass: the contract makes the pricing rule clear at your current count and the next one. Fail: “flat rate” meaning a rate that rises with every branch. A franchise SEO pricing example puts engagements at $2,000 to $5,000 a month per location; Third Marble’s franchise packages start around $499 to $549 a month and discount when more than three locations run. The models differ, so compare the full invoice, not the word flat. groas charges one flat monthly fee per service, regardless of ad spend or location count, with no setup fee, percentage of spend, or long-term lock-in. Ask every vendor for the same location-count scenario before calling one cheaper.

Paper-craft franchise bill fanning out into many small per-location bills

  1. Get account, page, and data ownership in writing. Pass: the contract says they stay yours, and the vendor can show you an export now. Fail: pages trapped on its domain, conversion history visible only in its dashboard, or a transition fee to hand over what you paid to build. Vendor-vetting advice rightly puts actual cost and asset ownership at the center of the decision. Do not save this question for a renewal conversation; that is when your alternatives are weakest. groas states that accounts, pages, and data stay yours, month to month. If another vendor will not put its answer as plainly in the agreement, treat that as an answer.

F. Run the phone test the demo cannot run for you

  1. Search three markets, click the ads, and compare the pages yourself. Pass: search a buying keyword plus a city in three of your markets, click through, and find distinct URLs with the right local service, offer, and phone number. Call each number and confirm it reaches the intended location. Then repeat from a browser set to a different location and compare what you see with the account’s Matched locations view. Fail: one URL with a swapped city headline, a wrong-city page, or a number that routes to the wrong branch. This is the check operators skip most. A demo renders from one office in one city; it cannot show what your customers see across three markets unless someone actually tests them. Say you spend $20k a month across ten locations. Wrong-city landings keep consuming that budget while somebody discovers the page matrix needs rebuilding, and the rebuild may be another charge for pages you thought you had bought. The phone test costs two minutes. Skipping it costs paid clicks and, potentially, the pages all over again.