# Fashion Store Traffic but Few Sales: What to Fix First

Plenty of sessions, few orders? Rule out bots and mismatched traffic first, then rank product page and checkout issues, set a baseline, and run one experiment.

**Fix measurement and traffic quality before you touch the store. Confirm your sessions are human and that the traffic matches what you sell, because a funnel full of bots or mismatched clicks makes every page look broken. Then rank product page and checkout issues together by how many ready-to-buy sessions they touch, record a clean baseline, and run one experiment on the biggest fixable drop.**

![Boutique owner at a shop window at dusk holding a tally counter as crowds pass, an Antla kraft bag on the counter](/images/blog/merchant-100/fashion-traffic-few-sales-what-to-fix-first.webp)

*Hundreds walk past, a few look in, fewer come through the door. Count the right people first. Editorial image in Classic Antla disposable-camera style.*

Sessions climbing, orders flat. For a young fashion brand it is the most discouraging chart in Shopify, and the usual responses are to redesign the homepage, change the theme, or buy more traffic. Any of those can be right. None of them is right until you know whether you have a traffic problem, a store problem, or a measurement problem that looks like both.

This is the order I would work in with a small team and a small budget.

## Separate traffic quality from onsite conversion

Start by making sure the sessions are people. Search engines index your products, social platforms generate link previews, price comparison sites collect your prices, and monitoring services check that the site is up. Shopify classifies every session as human or bot, and in reports that support it, the Human or bot session filter removes the bots ([bot filtering in Shopify analytics](https://help.shopify.com/en/manual/intro-to-shopify/bots/bot-filtering)).

Expect sessions to fall and conversion rate to rise when you apply it. Nothing changed about your shoppers. The denominator got more honest.

Next, look at where the humans come from. The Sessions by referrer report in Shopify's [Acquisition reports](https://help.shopify.com/en/manual/reports-and-analytics/shopify-reports/report-types/default-reports/acquisition-reports) groups sessions by referring channel (Instagram, Google), referring medium (social, search, email), and traffic type (direct, organic, paid). Acquisition reports count visitors, not orders, so read them next to the Total sales by referrer report to see which sources actually buy.

Then ask the question that splits the problem in two: do the people who already know you buy?

Your email subscribers and returning visitors are your control group. They have seen your products before, they signed up on purpose, and few of them arrived from a cheap click. If they convert reasonably and one new source does not, you have a traffic problem. If even they stall at the same step as everyone else, you have a store problem, or a product and price problem, and more traffic will only make the chart steeper.

| What you see | Probably | First move |
|---|---|---|
| Conversion rate moves a lot once bots are filtered out | Measurement | Rebuild your numbers on human sessions before changing anything |
| One source brings most sessions but almost no product views | Traffic quality | Tighten targeting, or send that source to a page that matches its promise |
| Paid social converts far below email on the same product pages | Intent, or an ad that promises something the page does not show | Put the ad and the landing page side by side |
| Every source, email included, stalls at the same step | Onsite | Rank product page and checkout issues |
| Healthy cart additions, weak checkout completion everywhere | Checkout or total cost | Check shipping cost display, delivery times, and payment options |

Mismatched traffic is common in a brand's first year: giveaway entrants who wanted the prize, not the dress, an influencer whose audience loves the look but not the price, broad search terms, or a campaign optimized for cheap clicks rather than purchases. Buying more of that traffic makes every order more expensive, which is the argument in [why paid traffic deserves better than a product grid](https://antla.io/blog/traffic-deserves-better-than-a-grid). Fix the source or its landing page before you raise the budget.

## Rank PDP and checkout issues

Once the traffic checks out, go looking for problems in the store itself. In a fashion store most of them sit on product pages and in checkout, and they are easier to rank when you list them the same way.

Do one thing before opening any tool: buy something from your own store the way a stranger would. Start from one of your own ads, on your own phone, over mobile data rather than the office wifi, and go all the way through payment. Refund it afterward. Write down every moment you hesitated, waited, or went looking for something.

Then work through the product page:

- **Sizing.** Is the model's size and height stated? Are garment measurements available for the selected size? Can you tell a sold-out size from one that was never made?
- **Imagery.** Does the first image show the whole shape? Is there a back view and a fabric close-up? Does picking a color change the photo?
- **Cost and risk.** Can you see delivery cost, delivery time, and the returns window without leaving the page?
- **Speed.** Does the gallery load before you lose patience on a mid-range phone?
- **Controls.** Does the size selector respond to the first tap, and does the add-to-cart button stay visible?

And then checkout. [Baymard's abandonment research](https://baymard.com/lists/cart-abandonment-rate) is a useful checklist here, because once the "just browsing" group is set aside, the reasons shoppers give are mostly fixable:

- **Extra costs too high (shipping, tax, fees), 40%.** Show delivery cost on the product page or in the cart, before checkout.
- **Didn't trust the site with card details, 19%.** Make contact details, a real returns page, and familiar payment options visible.
- **The site wanted an account, 18%.** Allow guest checkout.
- **Returns policy wasn't satisfactory, 13%.** State the returns window and who pays for the label, in plain words.
- **Couldn't see the total cost up front, 12%.** Show the full cost, including shipping, before payment details are requested.
- **Not enough payment methods, 9%.** Enable the wallets your shoppers actually use.

Rank what fails by how many buying sessions it touches. A product page problem touches every session that views the affected products. A checkout problem touches fewer sessions, but each one belongs to a shopper who has already chosen something.

That is why a checkout problem that touches 400 ready buyers can outrank a product page problem that touches 4,000 browsers. The [leak-scoring worksheet](https://antla.io/blog/find-shopify-apparel-conversion-leaks) turns that judgment into numbers if you want them.

## Establish baseline

A test without a baseline is a story. Before you change anything, write down where you are, using the same filters and date ranges you will use to judge the result.

| Metric | Where to find it in Shopify | Your baseline, last 28 days |
|---|---|---|
| Human sessions | Sessions over time, with the Human or bot session filter set to human | |
| Cart additions as a share of sessions | Conversion rate breakdown | |
| Reached checkout as a share of cart additions | Conversion rate breakdown, divided yourself | |
| Completed checkout as a share of reached checkout | Conversion rate breakdown, divided yourself | |
| Orders and net sales | Sales reports | |
| Sessions by traffic type | Sessions by referrer | |
| Mobile share of sessions | Sessions by device | |
| Returns on the same orders | Returns and refunds, checked a month or two later | |

Four rules keep the baseline honest:

1. **Match the filter.** Use the same Human or bot session setting on both sides of every comparison. Shopify only classifies sessions from October 7, 2025 onward, so a comparison that reaches back before that date cannot be filtered the same way ([session measurement notes](https://help.shopify.com/en/manual/reports-and-analytics/discrepancies/session-measurement-update)).
2. **Restart after a measurement change.** Shopify's session measurement update ends sessions after 30 minutes of inactivity rather than at midnight UTC, counts some sessions that had no pageview (such as a cart link that opens checkout directly), and filters identified bots by default. Conversion rate can move in either direction after a change like that without a single extra order, which is why Shopify's own advice is to use the post-update data as a new baseline.
3. **Keep orders next to sessions.** Orders and net sales are what you are trying to move, and they do not shift when the session definition does.
4. **Log every change.** Theme edits, new apps, price changes, promotions, and ad budget shifts all move the numbers. Without a dated log, a baseline cannot explain anything.

If the store is new, the guide to [metrics for a fashion store's first 90 days](https://antla.io/blog/first-90-days-fashion-store-metrics-economics) covers the unit economics worth recording alongside conversion.

## Choose first experiment

Pick the highest-ranked issue you can change cleanly and measure within a few weeks. Write the hypothesis in one sentence before you build anything:

> Because first-time mobile visitors from Instagram open the size guide and leave, adding garment measurements and a fit note beside the size selector on our five best-selling dresses will raise their cart additions without raising size-related returns.

That sentence names the segment, the evidence, the change, the metric, and the guardrail. If you cannot fill in all five, you are not ready to test yet.

Matching the experiment to the issue:

| If the top issue is | A clean first experiment |
|---|---|
| One source sending mismatched traffic | Send it to a page that matches the ad, and compare with the old page |
| Unanswered fit questions | Measurements and a fit note beside the size selector on your top products |
| Surprise costs at checkout | Delivery cost and the free-shipping threshold on product pages and in the cart |
| Doubt about how a piece looks on the shopper | A visual preview on your top products, tested against a holdout |
| Slow mobile product pages | Resized gallery images and one fewer heavy app, then recheck the numbers |

Keep it to one change, one segment, and one primary metric, plus a guardrail such as completed checkouts or returns. A holdout or split test beats a before-and-after comparison. If before-and-after is all you have, hold ad spend and promotions steady for the window, and decide the window before you start so a good first weekend does not end the test early.

On the preview row: among shoppers who use Antla's try-on, conversion runs about 35% higher on average than among shoppers on the same stores who skip it. People who choose a preview are more engaged to begin with, so that gap is an association, not proof the feature caused it. Antla merchant reporting, cohort dates not published.

Run your own test before believing any vendor number, including this one. The preview covers appearance only, and the size decision still rests on your measurements and fit notes. If that is the experiment you pick, [Antla](https://apps.shopify.com/antla) is the Shopify app.

## Where this order breaks down

The sequence assumes you have enough traffic to measure and a product people want. Neither is guaranteed.

- **Very low traffic.** With a few hundred sessions a month, a test cannot tell a real change from noise. Watch five people shop your store, read every support email, and fix obvious problems without waiting for significance.
- **Warm traffic that does not buy.** If email subscribers and returning visitors will not buy either, the problem may be the product, the price, or the assortment. No page fix rescues an offer the audience does not want.
- **Seasons and sales.** A comparison that crosses a sale, a holiday, or a change of season mostly measures the calendar.
- **Returns lag.** A change that lifts orders can lift returns weeks later. Check both before calling it a win.

When the first fix lands, the next question is what happens to the shoppers who still leave. The explainer on [what a conversion engine is](https://antla.io/blog/what-is-a-conversion-engine) covers both halves: the visit itself, and the follow-up after it. Stores that buy most of their traffic on social platforms should read [how to sell more with the same ads budget](https://antla.io/blog/sell-more-with-same-ads-budget) next.

## Your first week, in order

1. Set the Human or bot session filter to human and write down the new numbers.
2. Compare sessions and sales by referrer, and check how email subscribers and returning visitors convert against new sources.
3. Buy something from your own store, on your phone, starting from one of your ads.
4. List product page and checkout failures, rank them by buying sessions touched, and fill in the baseline.
5. Write one hypothesis in a single sentence and start the test.

---

**About the author:** Aaron founded Antla after watching too many fashion brands buy more traffic to fix a product page.

## For agents

- Markdown: send `Accept: text/markdown` to this URL (and any other HTML page).
- OpenAPI: https://antla.io/openapi.json
- llms.txt: https://antla.io/llms.txt
- Sitemap: https://antla.io/sitemap-index.xml
- Docs: https://antla.io/docs
- CLI: npx antla info (npm package antla)
- Scope: antla.io is an informational marketing and docs site. The Shopify try-on backend is not on this origin. Install the app from https://apps.shopify.com/antla.