What conversion rate means
Conversion rate is the share of visits that end in the action you want, most often a purchase. The calculation is orders divided by visits, written as a percentage. If 2,000 people visit in a month and 40 place an order, the conversion rate is 2%.
You can measure other conversions as well: email sign-ups, add-to-cart clicks or account creations. The one most owners care about is the order, and it is useful to say which one you mean when you compare numbers.
Be careful with benchmarks. Conversion rates differ widely with the product, price, traffic source, device and country. A store selling expensive items to first-time visitors will usually convert at a lower rate than one selling cheap items to returning customers. Comparing yourself with an average from somewhere else is less helpful than comparing yourself to your own past results.
Why one number is not enough
A single store-wide rate hides where the problem is. Break the journey into steps and look at each one. Here is a hypothetical month for the same store:
| Step | People | Share of previous step |
|---|---|---|
| Visits | 2,000 | |
| Viewed a product | 600 | 30% |
| Added to cart | 150 | 25% |
| Started checkout | 90 | 60% |
| Completed order | 40 | about 44% |
The overall rate is 40 out of 2,000, or 2%. The table shows more. Only 30% of visitors reached a product at all, which points to the home page, navigation or traffic quality. Fewer than half of the people who started checkout finished, which points to checkout. The first problem may be worth more attention, since the effect carries through every later step.
These figures are invented to show the method. Your own analytics will show the real picture.
Diagnosing drop-offs
Work step by step, and ask what is likely to stop a shopper at each point.
Arrival. If many visitors leave at once, check the match between what attracted them and what they see. An advert promising one thing and landing on another produces quick exits. Also check page speed and how the page looks on a phone.
Browsing. Shoppers who cannot find products leave. Review site search, category names and filters.
Product pages. Look for missing information: size, material, delivery time, return terms. Weak photos, thin descriptions and a lack of customer reviews are common causes of hesitation. See whether the price is clear and whether the call to action is easy to spot.
Cart. If people add items but do not start checkout, consider shipping cost visibility, unclear totals and distractions on the page.
Checkout. Long forms, forced accounts, limited payment methods, errors and surprise costs are frequent causes. Failed card payments also count as lost orders.
Use more than numbers when you can. Watch a few real or volunteer users try to buy something and note where they pause. Read customer emails and chat logs for repeated questions. Five people struggling at the same spot is a strong clue.
Segment before you conclude
Split the same funnel by device, traffic source and new versus returning visitors. A store that converts reasonably on computers but poorly on phones has a mobile problem, not a general one. A channel that brings plenty of visits and no orders may be attracting the wrong audience, in which case the fix is in the marketing and not the store.
Running simple tests
Once you have a suspect, you can test a fix. The idea is simple: show one version to some visitors and another to the rest, and compare. A deliberate approach to split testing is covered elsewhere, but the basics for a small store are:
- Pick one change. If you change the headline, the image and the button together, you will not know what mattered.
- State a reason. “Adding delivery times near the price will reduce hesitation” is a testable idea. “Try a green button” is a guess.
- Choose one measure before you start, such as add-to-cart rate on that page.
- Run it long enough. Small stores have few orders, so results swing widely by chance. Running a test for a few days and declaring a winner is a common mistake. Cover at least a couple of full weeks so weekday and weekend behavior are both included, and collect enough events that the difference is not just noise.
- Avoid overlapping tests on the same page or the same step.
- Record the result, including when nothing changed.
If your traffic is so low that a fair test would take months, a before-and-after comparison is weaker but sometimes all you have. Treat it cautiously, because seasons, promotions and traffic mix change in the meantime. For very low traffic, concentrate on fixing evident problems found by watching users and reading feedback instead of testing.
Fixes that are rarely controversial
Some changes are widely seen as reducing friction without needing a test:
- Showing the full cost early so there are no unexpected fees.
- Offering guest checkout.
- Removing unnecessary form fields.
- Making sure pages load promptly on a phone.
- Displaying clear contact details and return terms.
- Offering payment methods your customers actually use.
Conversion is not the only goal
A higher conversion rate does not always mean more profit. A deep discount can lift conversions while reducing profit per order. Likewise, a cheaper traffic source that converts poorly may still be good if the visitors are inexpensive. Keep average order value, profit margin and returns in view, and judge changes by what they do for the business and not for one percentage.