A store with traffic and no orders is nearly always one of four things: the traffic is not buyers, the store fails quietly on a phone, the offer loses at the moment of decision, or the orders are happening and you are not seeing them. Check them in that order, because the first three are free and the fourth explains a surprising number of panics.
What follows is the sequence we run on an account before proposing any work, and the reason it is in this order is that each step is cheaper than the one after it.
First: is the traffic people who could buy?
Split your sessions by source and look at each one on its own. Traffic is not a single thing. A thousand sessions from an audience that clicked a curiosity post and a hundred sessions from people searching your product name are not comparable numbers, and blending them produces an average that describes neither.
The signal to read is not the conversion rate on each source. It is add to cart rate. If people from a source are not adding to cart at all, they never wanted the product, and no amount of work on the checkout will fix that. If they add and do not buy, the traffic is fine and the problem is further down.
No add to cart means a traffic problem. Adds without orders means a store problem. That single split saves most of the arguing.
Second: open your own store on your own phone
Not the theme preview, and not a desktop browser resized. Your actual phone, on mobile data rather than the office wifi, starting from a paid ad if you run them. Buy something. Use a real card and refund yourself afterwards.
Mobile converts at roughly half the desktop rate as a matter of course, which the benchmark data covers in more detail, so a gap is normal. What you are looking for is not a gap. It is the moment where you personally hesitate: a field that will not focus, a discount box that pushes the total off screen, a shipping cost that appears for the first time three steps in, a payment method missing that everybody in your market expects. Write down every hesitation. That list is usually the whole job.
Third: is the offer competitive at the moment of decision?
Not whether your product is good. Whether the specific page a stranger lands on answers the three questions they are holding: what is this exactly, what does it cost me in total, and what happens if it is wrong. Total delivered cost and the returns policy do more damage by being absent than by being unfavourable. A shopper who cannot find the returns terms assumes the worst version of them.
Check what a first time visitor sees against what a competitor's first time visitor sees, on a phone, with no prior knowledge of either brand. Most stores lose here on clarity rather than on price.
Fourth: are the orders happening and not being counted?
This one is worth checking before you spend money on the other three, because it is the only one where nothing is actually wrong. Compare orders in your Shopify admin against orders in your analytics and against the conversions your ad platforms report, for the same dates, and see whether the store agrees with itself.
Platform numbers and store numbers disagree by default, and the gap is usually much bigger than people expect. We measured it on an account we owned outright, and the write up is here. If your admin shows orders that your reporting does not, you have a measurement job and not a conversion job, and they cost very different amounts to fix.
How long should I give it before deciding something is broken?
Enough sessions for the answer to mean anything. At a two per cent conversion rate, a hundred sessions produce two orders on average, so a hundred sessions and zero orders is an entirely ordinary run of luck and not evidence of a fault. Somewhere around a thousand sessions to a page you start to be able to tell the difference between bad and unlucky.
This matters because most of the damage in this situation is self inflicted. A store gets a quiet fortnight, panics, changes the theme, the offer, the audience and the budget in one week, and then cannot tell which change did what. Change one thing, give it enough traffic to answer, and keep the others still.
What we do about it
We run the four checks above, in that order, and we report which one it was before proposing anything. Most of the time it ends at step two, and the fix is a list of specific moments in the phone journey rather than a project. What we look at, and in what order, is on the conversion page.