Getting found on Google
How Many Website Visits Before You Trust the Data
Find out how many days of data to wait for before touching price, photos, or text, and which problems you fix right away.
You check your dashboard on a Tuesday and see three visits. Monday had one. Did business just triple? For a second you think about changing something: the price, the photo, the text on your homepage.
The following Thursday, nobody books an appointment, and you wonder if the site is broken or if the ad you ran stopped working. Three visits, or zero bookings in a single day, don't tell you anything yet. What you do know is that a small business works with very little data each day, and with little data, everything looks like a signal.
The question that actually matters isn't whether today's number went up or down. It's how many days you need before that up or down actually means something, instead of just being the normal noise of an ordinary Tuesday.
Why a single day doesn't tell you anything
A local website getting between five and twenty visits a day doesn't have a fixed number waiting for it every day. It moves around within a range for reasons you don't control: whether it's a Monday or a Saturday, whether it rained, whether someone shared your site in a group text without you ever knowing.
Going from one visit to three is, in percentage terms, a two-hundred-percent jump. Put that way, it either scares you or thrills you, depending on the day you're having. But in plain numbers, it's the difference of two extra people passing by. Below twenty total visits, forget the percentage: just look at whether the total went up or down, and wait for more days before deciding that movement means anything.
There's exactly one exception where you don't need to wait at all: a visible technical problem, like a booking form that doesn't submit or a site that takes so long to load that nobody sticks around to see it. Fix that the same day you find it. Everything else, prices, photos, text, whether an ad is working, gets judged over several weeks, never a single day.
The chart above is just an ordinary week, no promotion, no changes, nothing special. Saturday is almost four times Sunday. If you only looked at Sunday, you'd think business had stopped. If you only looked at Saturday, you'd think things were booming. Neither one is true, that's just the normal shape of that week.
That's why comparing a Tuesday to the Saturday before it, or this week's Monday to last week's Saturday, doesn't work. Compare each day to its own past: Tuesday to Tuesday, Saturday to Saturday.
The rule for telling noise from a trend
- 1Rule out a technical problem
Check that your booking form actually submits and that the site loads properly on a phone. If that's the issue, fix it now, there's nothing to wait for. - 2Compare the same day of the week
A Tuesday gets compared to the Tuesday before, not to Saturday. A business that's busier on weekends will always have fewer visits on weekdays, and that's not a drop. - 3Gather at least three weeks of that day
If Tuesday has been below the week before for three weeks running, that's no longer a coincidence. With just one week, it still could be. - 4If you ran a promotion, compare the week after it ends
A discount bumps up visits while it lasts. What matters is whether, once it ends, some new people stuck around, or everything went right back to the old number.
- With fewer than twenty total visits, don't read percentages, just look at whether the total went up or down.
- Always compare the same day of the week, and only draw conclusions after three or four weeks in a row.
- A technical problem gets fixed the same day; a change to price, photo, or text gets judged over several weeks.
- Every change stays untouched for at least two or three weeks before you check if it worked.
Tell us what you do and your site is live this afternoon
Say it in one line, the way you would say it to a customer, and it goes live. Free, no card needed.
Build my site free
