Methodology · As of 2026
What we test
Counter & Crew is an independent buying guide for AI at local businesses. We sit on real lines, read the cash price the vendor will print, and publish a verdict. Nobody pays for a nicer rank.
Method as of September 2026. We refresh this page when the test changes. Canonical site counterandcrew.com.
Who tests
Counter & Crew tests. Not a vendor marketing team. Not a paid ranker. We write for the owner who still answers the phone: plumbers, HVAC shops, electricians, dentists, law firms, and the rest of the trades on For Your Business.
We do not publish a staff bio page with invented names, titles, or lab credentials. The byline is the site. The inbox is support@counterandcrew.com. The crawler files are /llms.txt and /llms-full.txt.
The stance is on About. Independent buying guide. Nationwide installs when you would rather not become the software person. Scores before the sale.
How a phone-line test works
A feature page is not a test. We place calls and send the missed-call texts on live numbers. If it cannot pick up, book, or text back when a person is on a job, it does not get a kind writeup.
The homepage says the same thing in one line. We do not just read the sales page. We call it. We sit through the booking. We score what happens when a real person is on the other end.
Two published artifacts sit on the homepage so you can hear the kind of call we mean. A Ridge HVAC missed-call text thread on a business line. A Riley receptionist audio clip from an after-hours new-patient call at a chiropractic clinic. Those are the live-test objects already on the site. This page does not add a stopwatch, a pickup rate, or a lab n we never printed.
Hard calls are the ones the demo script skips. The caller talks over the bot. They change the job mid-sentence. They ask for a person. They want a 2 a.m. dispatch, not "are you open tomorrow." A plumber under a sink. An HVAC tech on a roof. An electrician with the cover off the panel. If the product only works on the happy path, the review says so.
- Real lines. Forward the shop number or take the vendor number. Place the call. Send the miss text. Listen to whether it greets, follows the caller, and leaves a note you can open later.
- Live prices. The number on the card is the public monthly cash on the day we sat through the page, not the annual teaser. Custom means they would not print a dollar. We do not invent one.
- A verdict you can shop with. Recommended or Solid pick maps to BUY, Situational maps to CONSIDER, and Skip for most maps to SKIP. Finder uses that map, and the review body keeps the original wording.
What we score
Crew score is reliability on a real line, plus value, plus how easy it is for a one- or two-truck business. Not a feature checklist. Skip is a feature. If we would not put it on your phone, the table says so.
When a review carries a scorecard, the dimensions are the ones already on the card. Sounds human. Books appointments. Emergency calls. Transfers. Integrations. Setup. Value. Each dimension needs evidence in the writeup. If the writeup does not support a dimension, we leave it out. We do not invent a number to fill a cell.
The six test questions on the homepage are the same shopping lens:
- Answer. Does it actually pick up, or dump the caller into a maze?
- Understand. Can it follow a real caller, or only the demo script?
- Interruptions. What happens when they talk over it, change the job, or get impatient?
- Book. Does the appointment land on the calendar, or die in a voicemail?
- Human. Can it hand off to a person when the call gets messy?
- Save money. Does it earn its keep, or just add another monthly bill?
Scores update when the product changes. Last checked on a review is the updated date in the frontmatter. The compare table is the dated snapshot of monthly price, verdict, and aisle for every published review.
Dated method, as of 2026
This method is current as of September 2026. Reviews carry their own updated dates. Rosie is August 27, 2026. Frontdesk is August 26, 2026. Other cards print the day we sat through that page. When a vendor changes the meter, the cancel, or the booking path, we update the card and the date.
We refresh this methodology page when the test itself changes. New score dimensions. A different verdict map. A change in how affiliate or install money works. Until then, treat the 2026 rules on this page as the ones that produced the live cards.
We will not back-fill a 2025 lab we did not run. We will not add a "tested N tools in N cities" line we cannot point at. Tool counts on the homepage are read from published reviews at build time. They are not typed in.
How a vendor gets on the list
We review the aisles owners actually shop. Missed-call text-back. AI receptionists. Review asks. Website chat. Local SEO. Shop OS. The aisle index is /categories/. Trade hubs remix those aisles so a plumber does not get a dental phone.
A vendor gets a card when we can sit through a public pricing page, a demo line, or both, and write a review that names the cash, the gotcha, and one reason not to buy. Nobody pays for a nicer rank. Nobody pays to move from SKIP to BUY. A salesperson cannot buy the score.
We do not invent agencies on the installer map either. Matches come from a declared specialty on the same aisle, or a named tool slug. Listed means the agency told us what they do, or we built the card from public information, and the checking is not done. Verified means we looked at the work and called businesses they installed for. Nobody pays to pass.
If we have not reviewed a stack, we say so. Barbers do not get a fake Mindbody price. Restaurants do not get a fake POS. Medical practices do not get an invented Athena overlay. Custom means no public dollar. We leave it Custom.
How Counter & Crew makes money
We make money two ways. Vendors may pay affiliate commissions. Businesses may pay for installs. Neither buys a better ranking. Scores publish before we sell an install.
Affiliate links sit on some reviews. The visit button may use that URL. The verdict does not. Some links pay us. They do not buy the score.
Done-for-you installs are the other door. Flat-rate remote work in all 50 states. Packages on /installs/ are Missed-Call Recovery at $1,500 plus $99 a month care, Review Engine at $1,250 plus $99 a month care, and Full Front Desk at $3,500 plus $199 a month care. The fit call is 30 minutes. If none of it pays for itself, we say so on the call.
The house listing on the map is Counter & Crew Installs. We list it plainly instead of pretending the map is neutral about it. Finder still points at published reviews first, then the map or the DFY page.
Referral fees and install fees do not move a score. That is the "skin in the game" line on the homepage. We rank the tools, and we install the ones we would actually buy. The ranking publishes first.
What we will not print
We will not invent a pickup-rate, an average handle time, or a "jobs recovered this month" number we did not measure. If a review or a guide already printed a vendor claim, we label it as theirs. "Designed for seconds after the missed-call event" is a wiring claim on the missed-call guides. It is not a Counter & Crew lab timing.
We will not invent customer quotes. We will not invent team bios. We will not invent a metro we serve as if the site were local. Framing is nationwide.
We will not sell you Weave for a plumber or Rosie for a dentist. Wrong aisle is a skip, even when the affiliate link would still pay.
Where to go next
Read a card on Reviews. Shop one trade on For Your Business. Or answer three questions on Find AI. The name, the inbox, and the two doors (map and DFY) are on About.