Open your profile's performance chart and look for a date. If the count holds, breaks between two consecutive days, and then sits flat at the lower level, a switch was thrown, and the break says when. Failing that, it is a slope over six or eight weeks with no single day in it, and something accumulated or the field moved underneath you. Finding that date, or failing to, is the first fact.
Neither shape names the cause. A call chart varies for reasons tied to different systems: how your calls are counted, how many listings they are counted across, what the profile says, and who else is in the results. That is the order to check them in, cheapest first, because each step only means something once the one above it is known good.
The Call Button Is One Number Among Several
"Calls dropped" is a statement about your phone. Your profile's performance data is a statement about one event, a tap on the call button attached to that listing. Two measurements, moving independently.
A customer who finds you on Maps, clicks to the site, and dials the number in the header made a call with no button tap. One arriving through a directory, or a number that forwards first, lands on the same handset and is counted elsewhere or nowhere.
So find out which series moved before you look for a cause. If views and searches held steady and only call taps fell, your exposure is intact, and something about the button or the number changed. If views and searches fell alongside the calls, exposure moved. If every series holds flat and only your handset is quieter, the drop sits where the profile never recorded it.
Two measurement faults are cheap to rule out. A call-tracking number swapped onto your profile starts a fresh counted series from zero, and the old series continues nowhere. If it replaced the old number rather than joining it, that number may still be live in your website header or a directory, splitting the calls across two lines with only one read. Both show as a clean step dated to the day of the change.
A Second Pin Can Take Calls the Dashboard Never Sees
Your dashboard reports on the listing that the account manages. Nothing in it reports a second listing exists.
Two things produce a second record, each found by a different search. A claim nobody remembers -- a marketing vendor or a former employee verified a listing years ago on an account you cannot open -- turns up when you search the old phone number it still carries, in quotes. An address that moved and left a record standing at the old street turns up when you pan the map around that street.
The effect on the call chart is arithmetic. A call placed from the pin nobody manages never reaches your dashboard at all, and if that record carries a number no one answers, the caller reaches a dead line and lands nowhere either. This is the version that reads as a clean step with no edits on your side, because the edits were on the other listing.
Run both searches, then confirm the reviews you remember sit on the listing you manage. A listing's own "Suggest an edit" menu in Maps carries a close-or-remove option with a duplicate reason, and that is how one is reported. Write down what each record holds first, since what survives on a removed listing is not yours to decide.
The Profile Has Other Writers, and a Record of Them
You edit the profile. So does anyone you ever granted manager access, any listing service on the account, and the public, through suggested edits. A field can move without anyone here touching it.
Which field moved matters less than when, because you already have a break date. A phone number replaced with one that does not ring sends every call nowhere from the hour it was saved. Hours set to temporarily closed, or a special-hours entry left standing past its date, change how the listing is presented from the day it takes effect, a smaller step on the same scale. Both leave a date, and it either matches the break in the chart or it does not.
So the audit is a comparison. Read the public listing on a phone, in a private browsing window, against the dashboard field by field, because the state you see logged in is not always the state a searcher sees. Then read the profile's edit record for entries you did not write, and set their dates beside the break date.
What triggers a review or a suspension is not visible from outside it. The observable part is the state the dashboard reports and which fields differ from the ones you set. This audit sits third because a duplicate makes it worthless: audit the wrong PIN and every field looks correct.
The Field Moved While Your Profile Stood Still
Local results are a ranked set with a limited number of visible spots, so your position is relative to whoever else is in it. Nothing on the profile has to change for your position to.
The events that change the set leave no mark on the dashboard, which reports only on you. A new pin nearby is the one to look for first, because it is the only kind visible today without a record of yesterday: pan the map and a branch that was not there before simply is. The rest -- a competitor who rewrote a primary category, or who was absent a while and came back -- changes what a listing says rather than whether it is there, and is invisible unless you wrote the old set down.
So run your two or three highest-value searches and write down who is in the set. With a record from before the drop, the comparison names a new arrival or rules one out. Without one, the diagnosis stops here, and for a reason worth knowing: the previous state of a search result is not stored anywhere retrievable, and there is no history to request.
Run them from the same place every time. Where you search from changes what comes back, which is a separate question from why it changed.
It is the one finding here that does nothing to repair.
Last Month Is the Wrong Baseline for a Seasonal Trade
Every check so far assumes the number of people searching held constant. For weather-linked work, it is wrong in both directions: heating and burst-pipe searches climb in cold months and fall away in warm ones, while cooling, roofing, and exterior work run the other way. Your calls are downstream of demand, and demand is not flat.
Month-over-month comparison builds that swing into the baseline, so the month after a peak, measured against the peak itself, shows a drop for the seasonal reason and nothing on the listing. What separates a demand change from a visibility change is the same weeks a year earlier: the same eight-week window against the same one last year.
A demand change is proportional: impressions, searches, and calls fall together, your position is unchanged, and the year-earlier window shows a similar dip on similar dates. A visibility problem is disproportionate: calls fall further than impressions, or impressions fall where the year-earlier record shows none.
Query mix moves underneath the total: where the performance data splits people who searched your business name from those who searched what you do, a fall in the name half points toward awareness and a fall in the service half toward visibility.
A year-over-year comparison requires a record that goes back a year, and the date picker in the performance view stops where it stops. Whatever was not exported before that edge is gone.
Frequently Asked Questions
Give it two consecutive comparison windows. A small operation's daily call count is noisy: one holiday week, one week the crew was tied up on a single large job, and the chart moves. Compare rolling four-week totals with the four weeks before them, and treat a single flat week as noise until the next window confirms it.
Yes. A duplicate does not have to outrank you to cost you calls; it only has to be the listing a share of searchers happens to tap. If it carries a number you no longer answer, those callers reach a dead line and never appear in your data at all, which is why the loss looks bigger than any ranking movement explains.
Start with what your own website publishes, since a stale hours block on your contact page can end up on the listing. The other route is access: manager permissions granted to a vendor or an employee stay attached until somebody removes them, so open your profile's user list and remove anyone no longer working on your business.
It shows up in your phone log while the profile's own numbers hold flat. Calls generated by a search campaign's call asset or by Local Services Ads are dialed from the ad rather than from your Business Profile, so they never register as a call button tap. Pause the campaign and the handset goes quiet with nothing moving in the profile data.
Yes, and the recovery itself is evidence. A drop that ends as sharply as it began points to a state that expired rather than a change somebody reversed: a special-hours entry running past its end date and then clearing, a pending edit resolving, a competitor's temporary closure ending. Write down both dates. The gap between them is the life of whatever expired.
Not necessarily a lost one. The profile counts calls, messages, website clicks and direction requests as separate series, so a chat button switched on, or a booking link added by an integration, hands a searcher somewhere else to tap. Add the four together across both windows. A total that holds while the split moves is a change of shape; a total that falls is a loss.
Some drops resolve in an afternoon, and some have been running for months before anyone opened the chart — a review of your listings, your edit history, and who else entered your results tells you which one you are holding. Green Thumb Local helps home-service businesses get found, get chosen, and get more calls with local SEO, Google Business Profile optimization, and content. Call (480) 360-0101.
