Website traffic but no sales is one of the most common things founders describe to us, and one of the most misread. The instinct is to blame the website and start redesigning. Sometimes that is right. Often the site is doing its job and the problem sits before it, in who is arriving, or behind it, in an offer that gives them no reason to buy.
This guide works through the checks in the order that wastes the least money: traffic and tracking first, then the offer, then trust and purchase friction. At the end is a decision tree and a way to choose between a deeper diagnosis and a small, bounded test.
If you are still trying to attract visitors in the first place, start with how to get more of the right traffic instead. If nearly all your visitors come from Facebook or Instagram ads, Facebook ads getting clicks but no sales covers the channel-specific checks.
Check whether the traffic and tracking are meaningful
Before asking why visitors are not buying, confirm two things: that they are real visitors, and that purchases are being recorded properly.
Is the tracking telling the truth?
- Place a test order, or pick a recent real one, and confirm it appears in your store platform and in your analytics with the right value.
- Compare the order count in your store with the count in analytics and in any ad platform for the same dates. Large gaps mean at least one of them is wrong.
- Check that your analytics tag loads on the checkout and order confirmation pages, which some setups handle separately from the rest of the site.
- Exclude your own and your team’s visits where your analytics tool allows it.
Your store platform’s order list is the most reliable record of sales you have. Use it as the reference point and treat every other report as an estimate.
Is the traffic real, and is it the right traffic?
Look at where visitors come from and what they do. These are common signs of low-quality traffic:
-
Very short visits from one source.
Many sessions lasting a second or two with one page viewed often mean accidental clicks, bots, or a mismatch between the ad and the page.
-
Unexpected locations.
Visits from countries you do not ship to, or from places with no link to your market.
-
A spike with no cause.
A jump in visits that matches no campaign, post, email or mention.
-
Visitors who never reach a product.
If people land on the homepage from a broad campaign and leave, the traffic may not be looking for what you sell.
Split everything by source. It is common for one source to be selling steadily while another sends visitors who were never going to buy. A single blended conversion rate hides that completely.
Examine the offer and buying intent
If the traffic is real and tracking works, the next question is whether these visitors had a reason to buy, from you, today.
-
Intent.
Someone searching for the exact product is closer to buying than someone who tapped an eye-catching video. Lower conversion from discovery traffic can be normal, so compare sources rather than judging them together.
-
Price and value.
Is the price clear, and does the page show why it is worth it? Search for the same kind of product yourself and see what a shopper sees next to you.
-
A reason to choose you.
If a similar product is available from larger retailers, what makes yours the one to buy? If the answer is only ‘we are cheaper’, or nothing at all, no amount of page polish will fix it.
-
The full cost at checkout.
Shipping, delivery time and minimum order rules are part of the offer. A $40 product with $15 shipping is a $55 decision.
Offer problems are uncomfortable because they are not quick to fix, and they are often mistaken for website problems. A related trap is a store that does sell, but not at a price that works. In January 2026, Apothecary by Justin was paying about $400 to win a customer worth about $55 in gross profit, at a $69 average order. The fix was the offer first, then the ads. The full case has the numbers.
Look for trust and purchase friction
Now look at the site itself, following the path a buyer takes. Where visitors leave tells you where to look first.
| Where they leave | What it may point to | What to check |
|---|---|---|
| Landing page, before viewing a product | Traffic mismatch or an unclear first impression | Does the page match the ad or search that brought them? Is it obvious within seconds what you sell? |
| Product page, without adding to cart | Unanswered questions, weak evidence, price or value doubts | Photos, sizing and materials, reviews, delivery time, returns |
| Cart, before starting checkout | Surprise costs or hesitation | Shipping shown late, minimum order, a discount code box that sends people off to search for codes |
| Checkout, without paying | Friction or doubt at the point of payment | Forced account creation, payment options, errors on mobile, security cues |
Use your analytics funnel or event reports to see these steps. Treat them as prompts to investigate, not proof of a cause.
Do the journey yourself on a phone, on mobile data, as a first-time visitor would. Then ask two or three people who fit your customer to try it while you watch without helping. Their hesitations are worth more than any checklist. How to improve ecommerce product pages for real shoppers covers what a product page needs to answer.
Trust matters most for a store people have not heard of. Clear contact details, a readable returns policy, real reviews and honest delivery times do more than design flourishes. If the site does turn out to be the cause, conversion optimisation that starts with the cause explains how that work should be scoped, including when a rebuild is and is not justified.
Choose a diagnosis or a bounded test
Once you have worked through the checks, you are usually in one of three places: you have found a clear, cheap fix; you have a likely cause that needs testing; or you still do not know. The decision tree below helps you work out which, and what to do next.
Traffic but no sales: what to check first
- 1
Do orders in your store platform roughly match what analytics reports for the same dates?
Yes. Tracking is usable. Move on to traffic quality.
No. Fix tracking first. Every later decision depends on it.
- 2
Is most traffic from real visitors, in places you sell to, staying longer than a few seconds?
Yes. Move on to buying intent.
No. Pause or fix the sources sending poor traffic before judging the site.
- 3
Do visitors from your highest-intent sources (people searching for the product, email subscribers, returning visitors) buy at least sometimes?
Yes. The store can sell. Look at why other sources convert less, starting with the match between ad and landing page.
No. Look at the offer, price and trust before the page design.
- 4
Do visitors reach product pages and add items to the cart?
Yes. The drop comes later. Check cart and checkout for surprise costs and friction.
No. Check the offer, the evidence on the product page and whether it answers shoppers’ questions.
- 5
Have you had enough orders to tell a real change from chance?
Yes. Test one change at a time and compare periods of the same length.
No. Rely on qualitative evidence, such as watching real shoppers and reading customer questions, and make changes you would make anyway rather than reading small swings as results.
A warning about small numbers
Conversion rate is orders divided by visits. When orders are few, that rate is not a stable fact about your store. It is a rough estimate that jumps with every single order.
Two fortnights, two extra orders
A store has 600 visits and 3 orders in a fortnight, a conversion rate of 0.5%. The founder changes the product photos. The next fortnight brings 600 visits and 5 orders, or about 0.83%.
On paper that is a two-thirds improvement. In practice it is two orders. A different mix of visitors, a payday, one gift buyer or plain chance could explain it. The new photos may have helped, or may have done nothing.
With this few orders, the honest reading is ‘no clear evidence either way’. Hypothetical figures for illustration.
This is also why a benchmark is not a diagnosis. Knowing that other stores convert at some rate tells you nothing about why yours does not, and at low volumes you may not yet know your own rate. Use a benchmark, if at all, as a prompt to look, never as the answer.
At low volumes, lean on evidence that does not need large numbers: watching people use the site, reading the questions customers send, seeing where analytics shows people leave, and fixing plain faults such as broken pages or hidden shipping costs.
Diagnosis or test?
-
Choose a bounded test
when you have one likely cause and a change you can make cheaply, such as showing shipping costs on the product page. Set the budget, the time and the measure before you start, and write them down.
-
Choose a fuller diagnosis
when the checks point in several directions, when changes you have already made did not help, or when the offer and the economics are in question. A diagnosis looks at the business, the site and the traffic together rather than one at a time.
Either way, change one thing at a time where you can. Three changes in the same week leave you unable to say which one mattered. If several people have already given you different answers, why three providers found three different problems explains how that happens, and what to fix first sets out the order I work in. More guides are in the resource hub.
You are on the list. The next one lands on Friday.
About the author
Mareike NiedermeierMareike Niedermeier is the founder and CEO of Sales Savvy Online and the creator of Survivable Distance, a theory of ecommerce viability. Since 2019 her team has worked with more than 200 ecommerce brands. The Australian Business Journal named her a Top Ten Marketer in Australia.
More about Mareike →