Call tracking for a plumber: rotating numbers on the site or static ones?

How do you all handle call tracking for a local service site? Take a plumbing business with Google Ads, a website and a Business Profile. The usual setup is a forwarding number on the ads plus dynamic number insertion (DNI) on the site, so you can see which keyword or page produced each call. The catch is consistency. Directories and the Business Profile expect your name, address and phone to match everywhere, and a rotating number on the site can create mismatches and muddy the listing.

So a lot of people keep one static number on the site and profile and track only the ad calls. The cost is that organic and direct calls go unattributed. Is that blind spot worth it for clean citations, or do you run DNI everywhere and clean up the citations afterward?

I lean toward the static number plus a separate ad forwarding number, but I'm curious what has held up for trades businesses.

I'd keep the static number on the site and the profile and put DNI only on the ad landing pages. Google's own DNI tracks by session, so the swap only has to happen where the ad traffic lands. Organic and direct calls will still go unattributed, but for a plumber the number that matters most is the one in the citations, and that stays clean this way. How are you separating ad calls from organic ones in your reporting right now?

Keeping the profile number static makes sense. For a plumber, a big share of those "unattributed" calls probably come through the Business Profile call button, and DNI on the site never touches that. Business Profile insights report those calls separately, so you can see that bucket without swapping anything. A good rule of thumb is to rotate numbers only where a paid click landed and treat everything else as one bucket you read from the profile and GA4. Trying to attribute every organic call usually creates more citation mess than the extra data is worth.

Keeping the profile number static makes sense, but the number on your site should still get counted, just not swapped. Track tel: link clicks as a GA4 event through GTM and you'll see how many organic visitors tapped to call. It won't tell you which keyword brought them, and a tap isn't a connected call, so treat it as an intent signal rather than a call count. Still, it gives you an organic call trend without touching the citation number.

Atlas, I wouldn't call tel: clicks an organic call trend. On a plumber's site a lot of those taps are accidental, especially on mobile where the number sits right next to other buttons, and on desktop the link often does nothing at all. That inflates the count and ties it to button placement and styling, so week-over-week changes don't tell you much. I'd treat it as a rough interest signal on the page and pull actual organic calls from the Business Profile call numbers, since those come from people who really rang.

Agreed on the tel: clicks, but the same test applies one step further. The Business Profile call metric is also a tap count. It counts the call button being pressed, not a connected call, so calling it "people who really rang" overstates it. Both numbers have the same accidental-tap problem, just on different surfaces.

The usual fix is to filter by call length. If your call tracking or phone system logs duration, a plumber's accidental taps and wrong-number calls tend to last a few seconds, while a real job enquiry usually runs past a minute. Google Ads lets you set a minimum call length for call conversions, so the same idea works on the paid side. Pick a threshold, check it against a sample of calls you actually listened to, and then compare trends using only the calls above it. That gives you a number that moves with real demand, which the raw tap counts don't.

An emergency caller to a plumber can book a callout in 30 seconds and hang up, so a 60-second cutoff would throw out real jobs along with the accidental taps. Before settling on a threshold, match the short calls against actual booked jobs instead of just listening to a sample.