I’ve been reading a lot of store feedback threads lately, and one thing I keep noticing is that the advice often goes in many directions at once: improve product photos, add reviews, clarify shipping, change pricing, speed up the site, simplify checkout, and so on.
All of those can matter, but I’m curious how store owners decide what the real blocker is before changing things.
If visitors are browsing but not buying, do you usually rely more on funnel data, session recordings, customer emails, live chat questions, feedback from buyers, or just testing changes one by one?
What signal has been the most useful for figuring out what to fix first?
My suggestion would be that that don’t decide by picking a fix, decide by finding where people leave, then why. Step one is funnel data: look at what percentage go from product page to add to cart to checkout to purchase, and the biggest drop-off stage tells you what to work on. If they don’t add to cart, it’s your product page, offer, price or trust; if they add but don’t reach checkout, it’s cart or shipping cost; if they reach checkout but bail, it’s shipping shock, payment options or a forced sign-up. That one number stops you guessing.
One reframe worth making first though: check it’s even a store problem and not a traffic one, because if you’re sending wrong-fit visitors (like cheap broad ad traffic), they’ll browse and never buy no matter how good the store is, and no page tweak fixes bad traffic.
Hope that helps! If it did, a Like and Marking it as Solution goes a long way and helps others find the fix faster too.
For me funnel data is always the first place to look. If people are viewing product pages but not adding items to the cart its usually an issue with the product page, pricing or trust. If they are adding to cart but abandoning checkout then I look at shipping costs checkout friction or payment options.
After that session recordings are incredibly helpful because they show where visitors hesitate or get confused. I also pay attention to customer questions they often highlight missing information on the site.
Rather than changing everything at once I try to fix the biggest bottleneck first measure the results and then move on to the next improvement. That approach has worked much better than making multiple changes simultaneously.
I’d separate the decision into three passes: where the drop happens, who it happens to, and what evidence explains it.
Funnel stage gives the first cut, but I’d segment it before changing the store. A paid-search visitor landing on one product, a TikTok visitor landing on the homepage, and a repeat visitor coming back to cart can all look like “browsed but didn’t buy” while needing different fixes. Compare drop-off by source, landing page, device, and product or collection.
Then use recordings, on-site search, support questions, and chat transcripts to explain one drop-off point. If product views are high but carts are low, look for repeated hesitations around images, variant choice, delivery timing, returns, or price. If carts are high but checkout starts are low, test shipping and discount clarity in the cart. If checkout starts are high but purchases are low, check the checkout step when accessible, payment options, and surprise costs.
The best signal is usually the same problem showing up in two places: numbers show the drop, and recordings or customer questions explain why. That turns the next change into a small test instead of a full redesign.
If you would share your story, that would be easier, but we need to look at the entire funnel. The funnel doesn’t start from the story. Start from the ads you’re running.
It might be that your ads are not good enough, not convincing enough. Maybe you need to add steps in the funnel to make it more convincing. It depends on the product you’re selling and a bunch of other stuff. If you can share more info, I would be able to give you a better answer.
This is really helpful. The pattern I’m hearing is: find the biggest leak first, segment it by source/device/landing page, then look for evidence before changing anything.
One thing I’m still curious about is the “why” part.
If recordings show hesitation around things like shipping, price, variants, or product details, do you usually treat that as enough evidence, or do you try to get direct feedback from those visitors somehow?
I feel like recordings can show the behavior, but sometimes the actual reason still needs a bit more context.
A lot of store owners seem to start with the questions customers keep asking. If the same things keep coming up shipping, delivery times, sizing, returns it usually means the store isn’t making those answers obvious enough.
After that, checking the funnel helps confirm whether the numbers match what customers are asking about. Session recordings help too, but there’s no need to watch hundreds of them. Once the same behavior keeps showing up across recordings, analytics, and customer questions, that’s usually a pretty solid clue about what needs fixing first.
The biggest thing is looking for patterns instead of reacting to one-off comments. It’s really easy to end up fixing the wrong thing if every random piece of feedback gets treated like the main problem.
Hi there @zhong_yan
It is always better to find out where your customers are dropping off in the buying proccess before you go making changes and shifting things around. Look at the analytics for your product views, add to carts, checkout starts, and purchases completed, then add that to what you’re hearing from customers and seeing in sessions. Minor tests such as modifying product details, trust elements, or clarity of the checkout process can help validate what will move the needle on conversions. Don’t feel like you have to change everything at once, prioritize your changes by the biggest areas of friction.
One thing I’d add is to compare converters vs non-converters. If both groups hesitate at the same point, it may not be the real blocker. The most valuable insights come from behaviors or questions that are disproportionately common among visitors who don’t end up buying.