Are your visitors real people?

Not all of them. Imperva, which sells bot protection and publishes a yearly report on bot traffic, found that bots made up more than 53% of all web traffic in 2025, up from 51% the year before. Human traffic was 47%. That's across the web, not your shop, but small shops get their share: scrapers, price checkers, crawlers and fake sign-ups.

Bots don't buy. If they make up a big part of your sessions, your conversion rate looks much worse than it is, and a spike in "traffic" may be nothing but bots.

How to check:

  • On Shopify: since October 2025, sessions reports have a Human or bot session filter. Set it to Human to see the conversion rate of real visitors. From 21 September 2026, Shopify's default sessions view focuses on human visitors where that filter is available. Shopify says sessions, conversion rate and add-to-cart rate may look different after the change, while orders and sales stay the same. So don't compare this autumn with last autumn without the filter: treat the numbers after the change as a new starting point.
  • In Google Analytics 4: known bots and spiders are left out automatically. You can't switch this off, and you can't see how many were removed. Bots that Google doesn't know about still get through.
  • Anywhere: look for sessions that jump on one day, come from a country you don't ship to, arrive as "direct" and leave after one page in under a second. Those are rarely shoppers.

Are your sales being counted?

Sometimes the sales are there and the report doesn't show them. Put two numbers side by side for the last 30 days: the orders in your shop admin, and the purchases in your analytics tool.

  • They match, roughly: good. Carry on.
  • Analytics shows far fewer: the problem is the tracking, not the shop. Common reasons are a purchase event that stopped firing after a theme or checkout change, a payment page that sends people off your site and back, and visitors who don't accept analytics cookies. Fix the tracking before you judge any channel, or you'll cut the ads that were actually selling.
  • Your admin shows few orders too: the sales really aren't happening. Go on to the next check.

Is your conversion rate actually low?

Most visits to any shop don't end in an order. IRP Commerce, which publishes monthly benchmarks for online shops in Great Britain, Northern Ireland and Ireland, put the average conversion rate at 1.97% in September 2026, up from 1.68% a year earlier. That's about two orders for every 100 sessions.

It varies a lot by what you sell:

Average conversion rate by category, UK and Ireland (IRP Commerce, checked 13 November 2026)
CategorySeptember 2026September 2025
All shops1.97%1.68%
Arts and crafts5.76%4.32%
Health and wellbeing2.83%1.93%
Kitchen and home appliances2.59%3.55%
Pet care2.13%2.45%
Fashion, clothing and accessories1.62%1.29%
Food and drink1.28%1.30%
Toys, games and collectables1.07%1.89%
Baby and child0.44%0.69%

The same category can move by a third from one year to the next, so use these as a rough guide, not a target.

Now do the sums for your own shop. At about 2%, 500 human sessions a week should bring about 10 orders. At 150 sessions a week, you'd expect about 3, and a week with none can simply be a quiet week. "No sales" is only a real problem when you have enough real visitors that a handful of orders should have happened, week after week.

If your rate is close to your category, the shop isn't broken. More orders will come from more of the right visitors, or a higher order value, rather than from fixing pages. If it's well below, or far below last year with the same mix of traffic, keep going.

Where do shoppers stop?

Find the step where people leave before you change anything. A shop has four: product page, basket, checkout, order. Each one loses some people, and Baymard's average across 50 studies puts basket abandonment alone at 70%. You're looking for the step that's much worse than the others, or worse than last year.

  • On Shopify: the Conversion rate breakdown report shows sessions, sessions with a basket, sessions that reached checkout and completed orders. We go through it step by step in why no one is buying from your Shopify store: 12 checks.
  • In Google Analytics 4: go to Explore, choose the Funnel exploration template, and add a step for each event your shop sends: viewing a product, adding to basket, starting checkout, purchase.

What each drop usually means:

Where shoppers stop, and where to look first
Where most people leaveUsuallyLook at
Before adding anything to the basketThe wrong visitors, or product pages that don't answer the questionYour traffic sources, then size, fit, delivery date and photos on the product page
In the basketA surprise: delivery cost, delivery date, returnsShow the total and the delivery date before the basket
In checkoutFriction: an account wall, a long form, a missing payment methodPlace a test order on your own phone

The biggest fixable reason in Baymard's US survey is cost: of shoppers who left a basket for a reason other than browsing, 40% said extra costs such as shipping, tax or fees were too high. 20% said delivery was too slow, and 19% didn't trust the site with their card details.

Are they ready to buy yet?

Some of your visitors aren't stuck at all. They're looking. In Baymard's survey of US shoppers, 42% who had left a basket said they were "just browsing" or not ready to buy.

That's especially true now. Black Friday is on Friday 27 November 2026, and in the weeks before it many shoppers find what they want and wait for the price to drop. A shop that is busy and quiet at the same time in mid November is often a shop full of people waiting. We wrote a separate piece on telling browsers, comparers and people waiting for the sale apart, with what to offer each group.

You can't make these visitors buy today, and a discount code for everyone costs margin on the ones who would have bought anyway. What helps is a next step that fits where they are: ideas for someone browsing, the answer for someone comparing, and a way to hear when the sale starts for someone waiting.

A worked example

This is an illustration, not a customer story. A small shop sells kitchen knives on WooCommerce. In early November the owner sees 2,400 sessions in a week and 6 orders, and thinks something is broken.

  1. Real people? A third of the sessions came on two days, from a country the shop doesn't ship to, and left after one page. Without them: about 1,600 sessions.
  2. Sales counted? Analytics shows 6 purchases, the admin shows 6 orders. Fine.
  3. Rate low? 6 orders from 1,600 sessions is under 0.4%, well below the 2.59% for kitchen and home in the table. Something is wrong.
  4. Where do they stop? The funnel shows many people view a knife and few add one to the basket. The product pages are the problem, not checkout.
  5. Ready yet? Reading last November's emails, the same question comes up again and again: "which one should I get for a beginner?"

The fix isn't a discount. It's one line on each product page about who the knife suits, a short guide comparing the three best sellers, and a way to ask a question before buying.

Where Signal fits

Most of these checks need no new tool. Where Signal helps is with the last two: the shoppers who get stuck on a product page, and the ones who aren't ready yet.

A small script on your shop reads why someone is visiting from the pages they read, where they came from, how long they stay and whether they've been before. When the answer would change what to show ("Is it a gift?", "What's holding you back?"), it asks one of your questions, and shows the answer or offer you wrote for that reason. Shoppers can also ask their own question on the product page, and you answer from your inbox. Questions and offers work on any shop that lets you add a script; showing three matching products from your catalogue works with Shopify only, for now.

Its Insights show which channels and pages bring people who buy, with bots and cloud servers kept out of the numbers. It never stores IP addresses. Signal starts at €39 a month. See how it works, or start in shadow mode: it shows what it would have done on your real traffic, without visitors seeing anything.

When this isn't the right approach

  • You have very few visitors. Below a hundred or so real sessions a week, percentages swing too much to read. Talk to the people who did buy, and work on getting more of the right visitors.
  • The product or price is the problem. If people come from every source, on every device, and leave at the product page whatever you change, ask a few of them why. Sometimes no page fix will help.
  • You're reading this in Black Week. Don't rebuild anything. Check the bot filter, place a test order, and put the delivery cost and date on your product pages. Leave the rest for December.