Most theme decisions are made in the wrong order. You fall for a demo, then spend six months trying to make your products look like the demo's products. Reverse it. Start with your store, then find the theme that serves it.
Bring your ugliest product
Every demo has one perfect photo per product and a six-word name. You have a bottle with a forty-word name, a size chart, an allergen list and three photos of varying quality. Load that product into the trial first. If the page still looks composed, the theme is doing real work. If it falls apart, no amount of settings will save it.
A theme is only as good as it looks with your worst content.
Count the sections you will actually use
A theme with ninety sections is not richer than one with thirty. It is harder to keep consistent. Write down the eight to twelve sections your homepage, collection and product pages genuinely need. Check that those are excellent. Ignore the rest.
Read the type before you look at the pictures
Photography is yours to change. Typography is the theme's opinion, and it shows on every page. Look at a product title, a price and a paragraph of description together. If the hierarchy is clear at a glance, the theme has been designed. If you have to hunt for the price, it has been decorated.
Judge it on a phone, in daylight
Most of your customers will never see the desktop layout. Test the collection page, the product page and the cart on a phone, outside, with one hand. Buttons should be reachable, prices legible, and the add-to-cart button never more than a thumb away.
Find out what happens when you need help
Read the documentation before you buy. Look for a changelog with dates on it. Send the support team one real question and time the answer. A theme is software. You are choosing who maintains it for the next two years.
Buy the fit, not the feature list
Features are cheap to list and expensive to live with. Choose the theme that makes your particular store clearer to shop, then add the one or two features you miss with an app. That trade is far better than the reverse.
Choose for the store you run, not the store in the demo.
