The reason this hurts is that traffic is the thing you were told to buy. You bought it, the chart went up, and nothing happened to the bank account. The natural conclusion is that the site is broken.
Sometimes it is. More often it is one of four other things, and it is worth eliminating those before commissioning a redesign.
Cause 1: the new traffic is different traffic
This is by far the commonest explanation and the easiest to check. If your extra visitors came from a broader keyword, a cheaper ad placement, or a piece of content that went further than usual, they are not the same people as your old traffic. They are earlier in their thinking, less sure they want the thing, and they convert worse. That is not a fault; it is arithmetic.
Cause 2: the traffic is real but wrong
Related and worse: the visitors are genuine people who were never going to buy, because the thing that attracted them is adjacent to what you sell rather than the same as it. An article that ranks for a question your customers do not ask brings traffic that will never convert, no matter what the page does.
This shows up as high volume, high engagement, and no revenue. It is a content and targeting problem wearing the costume of a conversion problem.
Cause 3: the extra traffic is not human
Worth ruling out early because it is quick. A jump in sessions with no corresponding change in anything else — no more scrolling, no more clicks, an implausible geographic spread, a suspiciously round pattern by hour — is often crawlers and bots. It costs nothing to check and it saves you from optimising for visitors who do not exist.
Cause 4: the site cannot carry the extra load
More traffic sometimes makes a site slower, and slower sites lose people before they see anything. If the increase coincided with a real change in how long the page takes to become readable, you have your answer and it is an infrastructure answer rather than a design one.
Cause 5: the page really is the problem
Only after the first four are eliminated is this the sensible conclusion — and then it is very sensible, because it is the one you can actually fix.
The usual suspects, in order of how often they turn out to matter: the page does not say what it is for within five seconds; there are too many decisions before the important one; nothing establishes that there is a real business behind it; the form asks for things nobody needs; and the phone experience is materially worse than the desktop one that everybody tested.
The order to check in
- Segment by source. Free, takes ten minutes, and resolves this outright more often than not.
- Check for bots. Also free, also quick.
- Check speed as a visitor experiences it, on a phone, not on your desk.
- Compare phone against desktop conversion. A large gap is a finding in itself.
- Then, and only then, audit the page.
The thing nobody says out loud
Sometimes traffic went up because somebody was selling traffic, and traffic was what they were able to deliver. It is a legitimate service and it is often bought in the belief that it is the same thing as growth. It is not: getting more people to the page and getting more of them to buy are different problems, and the second one is rarely included.
That is not an accusation of bad faith. It is a description of how the market is organised, and knowing it lets you buy the thing you actually needed.
What to do in the first week
None of this costs anything, all of it can be done in an afternoon, and it resolves the situation outright more often than not.
- Pull last month and the month before, split by source. Compare each channel against its own history rather than against the blended average.
- Split the same two months by device. A phone-heavy increase against a desktop-heavy baseline explains a great deal on its own.
- Check the new traffic for signs of life. Time on page, scrolling, any interaction at all. Volume without any of those is usually not people.
- Load your own page on a phone, on mobile data, away from the office. Time how long until you can read something. This is the test everybody skips.
- Run the fourteen checks on the page the new traffic lands on. Free, thirty seconds, and it will tell you whether the mechanics are sound.
If all five come back clean, you have learned something genuinely valuable: the problem is not the page, and the money is better spent on where the traffic comes from or on what is being offered when it arrives.
Questions people also ask
How long should I wait before concluding sales have not moved?
At least a full month, and longer if you have a long consideration cycle. Revenue lags traffic, sometimes by weeks, and reading a fortnight of data as a trend is how a healthy month gets treated as a crisis.
Could the extra traffic be hurting my rankings?
Not directly. More traffic does not lower rankings. What can hurt is if the extra traffic came from thin content that also weakened the site's overall quality — but that is a content problem, not a traffic one.
Should I stop the campaign that brought the traffic?
Not until you know whether it converts on its own terms. Judged against the blended average it will usually look bad; judged against its own cost per acquisition it may be perfectly fine.
Is a redesign the answer?
Rarely, and it is the most expensive thing to try first. A redesign that is not based on a diagnosis just changes which problems you have.