Stop benchmarking your conversion rate
Our site converts at 6.94% during evergreen stretches with everything at full price. I put in a screenshot below to prove it.
One note on the funnel: there's almost no drop between cart and checkout because ours is coded as basically one step.
Everything you hear about is top of funnel oriented. It's the new platform or tool... and your most valuable asset is sitting right in front of you and you probably invest the least in it.
We all set artificial ceilings for what's possible on our own site. Once we hit an industry range, we basically say good enough and leave it there.
The ceiling is a feel
What is the realistic ceiling on conversion rate? That's a real question... no one really knows. We just set averages based on good, bad, and ugly brands, and then that becomes the standard we all measure against.
This is the highest leverage lever you can pull... I can almost guarantee it is easier to get double digit improvement in your conversion rate than it is in your AOV or reduction in cost of traffic.
Our site conversion rate has materially risen every year, and this year it's up 61%, apples to apples. The number reflects what's possible when you treat converting people at the point of sale as a lever. For us that meant rewriting the pages where we sell, and testing headlines on top of them.
Where I'd start
Pull your funnel and get the three rates. The biggest leak points at the problem, the same way I'd read it auditing anyone else's business. If people aren't adding to cart, it's a product page problem, and if they're adding but not buying, it's pricing, offer, or checkout.
For most people, the best place you can spend your time is on your product pages. Rewrite them from customer language and fix the imagery, and ship the whole thing rather than testing your way there one button at a time. Run headline tests on top once the pages are sound. Then, keep iterating.
The real question
Why couldn't your conversion rate be higher? I don't think there's a good answer. Why do we limit ourselves to somebody else's average?
Whatever your funnel converts at today, pull it and find the biggest leak, then start there.