Two days after moving a clinic to a brand-new website, the owner sent me two screenshots. Visitors: near zero. Search clicks: near zero. Both graphs fell off a cliff on the day of the move.
That is the worst feeling in small business. You change something, the number dies, and the obvious story writes itself: I broke it.
The story was wrong. The website traffic dropped to zero on the chart because the thing that counts visitors had not been installed on the new site. Real people were arriving, booking, and calling. Nobody was counting them.
What actually happened
The old website had a small piece of tracking code on every page. It had been there for years, quietly recording every visit.
The new site was built from scratch. The tracking code was never added, because nobody thought to add it — it had never been a thing anyone touched. The moment the address pointed at the new site, the counter stopped counting.
So the graph was honest about one thing only: it had no data. It was not measuring a loss of customers. It was measuring its own blindness.
The fix took a few minutes. We put the same tracking tag back on every page, checked five pages by hand to confirm it was really there, and the numbers started coming in again that afternoon.
Why did my search traffic drop right after a change?
This is the part worth remembering, because it caught us too.
The search report showed clicks falling to zero two days before the website even changed. That is impossible. The old site was still live on those days.
The reason is lag. Search reporting is always a few days behind, and the most recent days on the chart are always incomplete. If you look at the graph today, the last stretch will always slope down toward zero — whether anything happened or not.
That means:
- The newest days on any analytics chart are not evidence. They are a partial count.
- A drop that starts before your change cannot have been caused by your change.
- Wait for the reporting to catch up before you draw a conclusion, then look again.
We waited. It filled back in.
The three things to check before you believe a collapse
When a number falls off a cliff, work through this in order. It takes ten minutes and it will save you a week of panic.
Is the measuring device still connected?
Did you change your website, your booking system, your email tool, or your phone number? If so, check that whatever was counting is still attached to the new thing. A tracking tag, a call-tracking number, a form, a spreadsheet somebody stopped filling in — any of them can quietly get left behind during a move.
Is the drop real, or is it just the last few days?
Look at where the decline starts. If it starts in the most recent two or three days, suspect reporting lag first. If it starts on a specific earlier date and stays flat, that is more likely real.
Does the real world agree with the graph?
This is the check almost nobody runs. Go look at the actual business. Did the phone ring? Did anyone book? Did anyone walk in?
In the clinic's case the answer was yes — there was a real booking, and three real phone calls, on the exact days the chart said nobody existed. That single check settles the argument faster than any dashboard.
The harder question underneath
Here is where it got interesting. Once we established the traffic story was a measurement problem, the owner asked a second question: but new bookings really did feel slower.
That one we could not answer, and that is the honest finding. There was no record anywhere of how many bookings the old system used to take in a normal week. So "slower" had nothing to be slower than.
We could see real friction we had added on changeover day — a deposit at booking where there used to be none, narrower appointment windows, a longer intake form. Each is a plausible cause. None of them is proven, because there is no baseline.
So we set a date a week out, agreed what we would count, and decided nothing until then. No deposit changes, no hours changes, no reacting to a feeling.
The takeaway
Before you change your business because a number moved, make sure the number is still measuring the thing you think it is — and write down today's figures, so the next time something shifts you have something to compare it to.