Online store traffic but no orders: find where shoppers drop out

Different points of failure require different work. A shorter form will not make an unsuitable product more desirable. More advertising will not remove an unexpected delivery charge. And a payment error can stop even someone who has already decided to buy.

Even when traffic is genuine and comes from a relevant audience, you still need to check whether the specific offer meets expectations. The aim is to connect an observation to a check and a next action, rather than changing the whole store on a hunch.

First, check whether a purchase can be completed

Visit on a phone as an ordinary shopper, without an administrator session. Open a product to which you actually direct visitors. Select an available variant, add it to the basket, choose a supported address or courier collection point and proceed through checkout. If anything stops you, record the exact step and message.

The check does not end at the order button. Establish whether an order was created in the store, whether the selected product and quantity are correct, and whether delivery and the total match what was displayed. Check which confirmations reach the buyer and the team.

Distinguish the order from payment and fulfilment

With cash on delivery, a newly created order is not yet a paid and delivered sale. With a card, an order may exist while payment has failed or remains pending. Conversely, a payment recorded by the provider without a correctly updated store status calls for an integration check. The screen shown after payment does not replace reconciliation of those records.

Use the test mode provided by the platform and payment provider, in an isolated environment where available. Do not enable payment test mode in a live store without an agreed plan. For example, Shopify's test-order documentation warns that customers cannot place live orders while payment providers are in test mode. A controlled real purchase requires permission and arrangements for cancellation, refunds and any fees.

Do not disable security or payment checks to "fix" an order. If the problem concerns the connection to the payment system, give a specialist the time, test-order number, step and error message. Do not send card numbers or other sensitive data.

Check whether the order is missing or merely absent from the report

If genuine orders exist in the store but analytics shows zero, measurement is the first problem. If the store has no orders either, the absence of an analytics event does not, by itself, explain why people are not buying.

In Google Analytics 4, shopping interactions can be measured as separate events. The official GA4 ecommerce guide describes view_item for viewing a product, add_to_cart for adding it to a basket, begin_checkout for starting checkout and purchase for a purchase. These need to be implemented correctly; installing the basic analytics code does not prove that this journey is being measured.

Use a specific order to check what is sent. Google recommends validation through DebugView, where you can inspect events and their parameters. Also check the transaction identifier transaction_id, value, currency and repeated triggering. Different purchases must have different identifiers; reopening the confirmation page should not create another purchase in the report.

Define what the business means by a "successful order" too. Created, paid, delivered and cancelled orders are different states. Analysing drop-off before checkout completion and reporting on completed sales answer different questions.

Relevant traffic does not mean readiness for this particular purchase

Someone may be interested in your category but want a different pack size, application, price range or delivery time. Look beyond who arrives to what they expected before opening the page.

Choose one important category or a small group of products. Compare the advertisement, search result or post with the actual landing page. If the promise is a bundle but the page shows a single product, the visitor has to search again. If an advertisement features a specific variant, check that it is in stock and easy to select.

Separate new visitors from existing customers, channels and devices where the data allows. Compare similar periods and campaigns. Traffic from a product advertisement to a particular item is at a different stage of the buying process from traffic to an educational article.

Site search also provides useful signals. If someone searches for a product you sell but cannot find it because of different naming or a poor filter, the problem is discoverability. Try several realistic shopping tasks: a required size, a compatible model or a particular use. Do not assume that a search with no results means there is no market demand.

The product page needs to resolve doubts before the basket

When products receive visits but few additions to the basket, start by asking whether someone can decide if this is the right product. Check the photos, specifications, variant, availability and price for the actual selection. Attractive copy does not compensate for a missing size or unclear compatibility.

For cosmetics, this might mean quantity, ingredients and manufacturer instructions, differences between ranges, or the need for a complementary product. For equipment, it might mean compatibility and what is included. For clothing, it could be sizing, fit and measurements. Show genuine specifications only; broad claims about "the highest quality" do not answer these questions.

Assess price and trust in the context of the offer

Visitors may compare the whole offer, not just the item's price: quantity, timescale, delivery, service and terms after purchase. If your product costs more, explain the real reason for the difference where one exists. Discounting without a diagnosis can reduce your margin while leaving the same obstacle in place.

Make it easy to establish who the retailer is, how to contact them and where to read the applicable terms. Genuine reviews and real photographs help when they relate to the specific choice. Do not replace missing evidence with fabricated ratings, counters or security badges.

In our Nailchic.bg project, we describe marketing work around the product range and questions people have before choosing: differences between product ranges, combining products and finding an appropriate collection. This Bulgarian store example shows how product explanations connect with communication. It contains no measured funnel demonstrating that a particular change increased orders.

When do delivery terms and the total cost become clear?

Review when buyers learn how they will receive the goods and what the total will be. If an important charge appears only after they enter their details, they are making the final decision with new information. Abandonment at that point may concern the terms rather than a difficult form.

Explain delivery options, timescales, restrictions and how charges are calculated. Where the exact amount depends on the address or method, state that dependency and calculate it at the appropriate step. Do not display an invented fixed charge simply to make the page look easier.

If you offer a free-delivery threshold, it must match the actual terms. Check whether a discount, code or different delivery method changes the total. Information in the advertisement, product page, basket and final screen must agree.

Example: a cosmetic product with an unexpected delivery charge

Imagine a cosmetics store where a shopper chooses one product. Its page has a price and a button, but delivery terms are hidden on a general information page. The charge becomes clear only after the shopper enters an address, and they leave.

One reasonable change is to place a short, accurate delivery explanation beside the ordering information, with a link to the details. Then check that the terms remain consistent throughout the journey. If the charge itself is the main objection, showing it earlier will not make it smaller. The offer may need reconsidering, but that is a commercial decision.

This scenario illustrates a possible mechanism. A fall in checkout completion does not prove that delivery is the only cause. Check payment, variants, errors and measurement quality too.

Check the basket and checkout on a phone

Adding to a basket signals interest, not a commitment to buy. Some people use it to compare the total or save a selection for later. To identify a technical or information problem, repeat the specific task and compare what you observe with the data.

Check that the selected variant and quantity are retained, the price updates correctly and a promotional code produces an understandable response. If a gift, bundle or minimum quantity changes the order, the conditions should be visible where the choice is made.

When entering details, test the town and courier collection-point selectors, an address with additional instructions, and returning to the previous step. Courier office collection is a relevant option in the Bulgarian context of these examples. Observe whether an error identifies exactly what needs correcting, the button remains accessible with the keyboard open and no pop-up obscures a required field.

Every requirement placed on the buyer needs a reason

If account creation is compulsory, assess whether it is necessary for your business model. For a standard one-off purchase, check whether the platform supports a suitable guest checkout. In B2B commerce or under special terms, identifying the customer may be part of the service itself. Do not apply the same recommendation mechanically to every store.

Request the data needed for the order and its fulfilment. Too many fields can make completion harder, but removing essential information can cause delivery problems. The aim is for the buyer to complete the order correctly and for the team to fulfil it.

Check each important payment journey separately

A successful cash-on-delivery order does not prove that card payments work. Check the methods you offer and their permitted combinations with delivery options. If payment involves a redirect or bank authentication, follow the return to the store and the resulting status. After an error, buyers should understand whether an order exists and how to continue safely.

Record the device, browser, payment method and exact step at which the difficulty occurs. One successful test does not prove that every variation works. One failed test establishes the specific problem observed, not how many customers have encountered it.

Define the funnel so that comparisons are meaningful

A funnel is a sequence of actions used to examine how far visitors progress. For an initial review, you can follow product viewing, adding to the basket, starting checkout and a confirmed order. This is a diagnostic model; express checkout and other genuine journeys may skip steps.

Do not subtract basket-event counts from product-view event counts and call the difference "lost customers". One person may view several products and add more than one item. Analyse the sequence using comparable users and clearly defined conditions for progressing between steps.

In GA4 Funnel exploration, the distinction between open and closed funnels matters: a closed funnel requires entry at the first step, while an open funnel allows entry later. The order of actions also affects counting. If express checkout bypasses the basket, a mandatory add_to_cart step may show apparent drop-off. Check the actual journeys before configuring the analysis.

Separate phone and desktop journeys, comparing similar products, channels and periods. Define how you attribute a device to the sequence; visitors may return on another device, and recognition may be incomplete. Changes to consent settings, blocked measurement or missing events also limit what is visible. Analytics is not a complete record of every real person.

Start where validated data shows a substantial break in the journey. Then look for an observation that could explain it: a missing variant, unclear delivery terms, an error message or a failed payment. The funnel tells you where to look; it does not reveal the buyer's motive on its own.

Which changes come first on a limited budget?

Start with confirmed obstacles that prevent purchases: a broken button, delivery unavailable in an area you serve, a payment error or an order that is not created. If orders exist but do not appear in reports, fix measurement before making decisions from those reports.

Next, address mismatches affecting important products and real traffic: an unavailable advertised variant, incomplete product information or terms shown too late. Choose a change that clearly relates to the observation, record the date and decide beforehand what you will check afterwards.

Increasing the budget makes sense when the journey works and there is a justified opportunity to reach suitable buyers. At low volumes, do not declare success after one order or expect a reliable A/B test without enough observations. A before-and-after comparison also needs context about campaigns, seasonality and stock availability.

Finally, assess the quality of the outcome: orders actually paid or delivered, cancellations, returns and costs. More orders created do not automatically mean more profit. Do not improve one metric at the expense of the store's ability to fulfil orders or operate profitably.

A checklist to start with

• Complete one important shopping task on a phone.
• Confirm the order, payment status and notifications.
• Check measurement against that specific order.
• Compare the promise with the product, variant and availability.
• Check when delivery terms and the total become clear.
• Choose one verified obstacle and a criterion for checking the fix.

When should you seek help?

Seek specialist help when orders and payments cannot be reconciled, events are unreliable or an identified problem requires a code change. If purchasing works technically but choosing is difficult, the task concerns product information and the user journey.

At WebSEOVision, user journey analysis and investigation of purchase obstacles connect the specific task with the available data and observations. This helps establish what should change, what requires development and how to verify the work.