Both rates are platform averages, not your store's. The split between the two losses is the number worth calculating for yourself.
There are two completely separate populations behind a low conversion rate, and they need opposite fixes. At platform-average rates, about 94 of every 100 visitors leave before they ever reach a cart: they did not find it, did not believe it, or were never the right person. About 5 more add something and then stop, which is a friction and price problem. Work out which of those two numbers is unusual for your store before you change anything, because a redesign aimed at the wrong one is a month spent for nothing.
Two populations, one number
"My store does not convert" is one sentence describing two unrelated failures. Separating them is the entire job of this chapter, because almost every wasted month in ecommerce comes from fixing the wrong one.
Around 94 in 100, at average rates. They never signalled intent, so you have almost no record of them. Three sub-causes, and they are genuinely different:
- They never found it. The product exists, and their path to it did not.
- They did not believe you. Nothing on the page made a stranger comfortable handing over a card.
- They were never the buyer. Traffic quality, not store quality.
Around 5 in 100. These people told you they wanted it. Something after that changed their mind, and it is usually the total, the delivery date, or the number of steps. Chapter 3.3 is entirely about this group.
They are a fifth the size of group one and about ten times easier to fix, which is why they are worth doing first even though they are the smaller prize.
A store with a discovery problem and a store with a checkout problem report the same conversion rate. They need opposite work. The Dropshipping Playbook
They never found it
The biggest of the three on most dropshipping stores, and the one that gets the least attention.
Between 15% and 30% of visitors use a store's search box rather than its navigation, and those visitors are worth far more than the rest: industry compilations put them at 40 to 45% of revenue on the sites that measure it. Econsultancy's frequently cited figure has searchers converting at 4.63% against 2.77% for everyone else.
Now the other half of that. Industry estimates put the share of on-site searches returning nothing at somewhere between 10% and 20%, with best practice under 5%. A visitor who searches and gets an empty page is a visitor who arrived with the highest intent you will see all day and was told the store does not have it.
The uncomfortable part: none of that appears in a Shopify report. You will see the session and the exit. You will not see the query. Chapter 3.4 explains why that is structural rather than an oversight.
They did not believe you
Trust is not a design opinion, and treating it as one is why so many redesigns change nothing. It resolves into a short list of specific, checkable things.
| What a stranger checks | What most new stores show | Cost to fix |
|---|---|---|
| Does anyone else buy this? | No reviews on the product page | Low |
| When does it arrive? | No delivery window stated | Free |
| What if it is wrong? | Refund policy unwritten or unlinked | Free |
| Who am I buying from? | No About page, default domain | Low |
| Is this the real product? | Supplier photos, low resolution | Medium |
Two of those are free and are the two most often missing. A stated delivery window and a written refund policy also reduce chargebacks, which is the thing that triggers the payment reserves in chapter 1.3. Rarely does one afternoon of work pay in three places.
Imported or fabricated reviews are one of the specific flags that appears in Shopify store-termination threads. The fix for having no social proof is getting some, slowly. It is not manufacturing it.
They were never the buyer
The third cause is the one nobody wants, because the fix is not in the store at all.
If your ad promises one thing and the landing page delivers another, the click was bought and the visit was wasted. This shows up as a high bounce on one specific source while the rest of the store behaves normally, which is why the diagnostic in chapter 3.1 splits by traffic source before anything else.
It also shows up as the pattern chapter 2.2 warned about: optimising a campaign for clicks or add-to-carts rather than purchases produces exactly this. The ad account reports success while the store reports strangers.
The stranger test, in fifteen minutes
Before any redesign, do this. It costs fifteen minutes and nothing else, and on a store under 500 sessions there is no better use of the time.
- Hand your phone to someone who has never seen the store and name a product, without pointing at it
- Say nothing while they look for it, and write down every place they hesitate
- Note whether they used your navigation or went straight to the search box
- Ask them, at the product page, what it costs to deliver and when it arrives, and see if they can answer
- Ask what they would do if it turned up broken, and see if they can find that answer
- Let them complete a real checkout, and watch which field they pause on
What those fifteen minutes buy is the one thing your analytics structurally cannot give you: a reason.
Fix them in this order
- The free trust items. Delivery window and refund policy on every product page. An afternoon, no cost, and it lowers chargeback risk too.
- The cart and checkout group. Small, but concrete and fast, and chapter 3.3 gives you the exact list.
- Discovery. Larger and slower, and the biggest prize on most stores. Start by finding out what people search for and what returns nothing.
- Traffic quality. Only once the store is known good, because otherwise you cannot tell a bad audience from a bad page.
- Design. Last, and only against something specific a real person told you.
Common questions
How do I know which half is my problem?
Is my store just ugly?
How many product reviews do I actually need?
Should I add a discount to fix this?
One number pays for the rest
What a single order is worth decides what every other fix can afford. Chapter 4.1 is where that number moves.
Go to chapter 4.1