What do you test first when a product page gets traffic but no sales?

When a product page gets steady traffic but barely any purchases, what is your first test: price, images, copy, reviews, shipping details, or checkout friction? Do you change one element at a time or begin with a full page review?

I would start by checking where the drop-off happens before changing the whole page. If people scroll and interact but do not add to cart, I would test the main image, value proposition, and price framing. If they add to cart but do not buy, shipping costs, delivery timing, or checkout friction are stronger suspects. I prefer one meaningful test at a time so the result is easier to trust.

The first thing I check before touching the page is where the traffic actually comes from. A page that converts fine for branded or high-intent visits can look broken when it's suddenly getting cold traffic. I'd split sessions by source, compare view-to-cart against cart-to-checkout, and then test the element closest to the drop-off point - one change at a time, and only after the data says where the leak is.

Checking whether the page answers the basic questions a customer might have before buying can help. Sometimes people are interested but still hesitate because they’re unsure about the product, how it works, or what they’ll actually get.

Sometimes the problem can be the offer itself rather than the page. If the product doesn’t feel different enough or there isn’t a good reason to buy it now, people can browse and leave even if everything on the page looks fine.

NathanielBr said:

I would start by checking where the drop-off happens before changing the whole page. If people scroll and interact but do not add to cart, I would test the main image, value proposition, and price framing. If they add to cart but do not buy, shipping costs, delivery timing, or checkout friction are stronger suspects. I prefer one meaningful test at a time so the result is easier to trust.

Click to expand...

Checking where people drop off first makes a lot of sense. There’s no need to change the whole page when you can find the part that’s actually causing the problem.