How to read this

• Pacing is straight line: month-to-date divided by days elapsed, multiplied by days in the month. It assumes the rest of the month runs at the current daily rate.
• Charged leads are the leads Google billed for. Total leads includes leads that were credited or not charged, so charged leads is the figure that ties to spend.
• Previous year comparisons shift back 364 days rather than a calendar year, so each day lands on the same weekday. Local Services Ads volume swings hard by weekday, and a calendar-year shift would compare a Monday against a Saturday.
• Call connection rate is connected calls divided by total calls, summed across locations rather than averaged per location.
• Booking rate is not shown. Booked status has been recorded on under 1% of leads since June, down from 9.4% in May. That reflects how the status is being used in the Local Services Ads dashboard, not booked work.
• Weekdays are the working week here. Across the portfolio, an average Sunday draws about 7% of a Monday's spend and an average Saturday about a third, because these are home services businesses that run Local Services Ads on weekdays. A location that shows no activity on a Sunday is closed, not missing from the data.
• Weekdays served counts the weekdays a location actually ran in the selected range. A location well below the available weekdays is one whose campaign was paused or ran out of budget, which is a real gap in coverage rather than a reporting artifact, and is flagged in the location table.
• Lead types. Phone, message and booking are how the customer made contact. Google does not publish whether a lead came from a direct search for the business or from browsing a category, so that split is not available at any level.
• Service category is each location's own primary Local Services category, taken from the categories its leads actually came in under.
• Rates on small samples are withheld. A connection rate drawn from fewer than ten calls is marked with an asterisk and left uncoloured, because one or two calls would otherwise show as a perfect or a failing score.
• Locations with no spend in the selected range are grouped separately. They did not perform badly; their campaigns did not run.
• Map placement. Markers sit on the location's city, projected through an Albers equal-area conic. Circle area is proportional to cost, so a marker with twice the area is twice the spend, not twice the radius. A small number of markers are placed from the market name in the account rather than a city field.
• Comparing ranges. Because volume follows the working week, two ranges of equal length can hold a different number of weekdays and show a volume change that is purely calendar. Where that happens the affected tiles are flagged. The previous-year comparison shifts by 364 days and is weekday matched, so it is the cleanest basis for a volume read.