The restaurant's website has one job: turn a hungry search into a booked table. This screen shows whether it's doing that, in restaurant terms, with the bot traffic already stripped out.
Sessions · June
9,420
↑ 12.6% vs May · humans only
Reservation starts
861
9.1% of sessions tapped Book
Completed bookings
604
~1,690 covers attributed
Calls from site
278
tap-to-call, mostly mobile
Where sessions come from
Branded search ("blue juniper kitchen") is 58% of organic; that's the review engine paying rent.
Top pages · what people actually look at
/menu
6,180
/ (home)
4,390
/reservations
2,540
/patio
1,360
/private-events
920
The menu PDF died in March; the live Menu Builder page is why menu views doubled: it loads fast and it's always current.
The only funnel that matters · search → site → booked table
9,420
sessions
861
tapped Book (9.1%)
604
completed (70% of starts)
~$45,900
revenue attributed at $27.16/cover
Drop-off between start and complete is almost all the date picker on mobile Safari; the fix ships with the July site update and is worth roughly $1,900/month if it closes half the gap.
📈
Insight · scanners were flattering the numbers
All-traffic showed 14,210 June sessions; 4,790 were bots and AI scrapers. The humans-only view is what pricing and marketing decisions should read.
Toggle "All traffic" above to see the difference. The bot share tripled since January; any restaurant still reading raw GA4 is overestimating its site by a third.