How to Tell If Your Shopify Traffic Is Actually Bots

If your sessions are up but your orders aren’t, this is the first thing worth checking — not your product, not your price, not your ads. Your traffic itself.

Here’s how to actually check, without needing a fancy tool.

The five real signs

1. High concentration on one device, one country, one referrer. Real customers are messy. They come from a mix of phones and desktops, a mix of places, a mix of sources. If a single day’s traffic is 95%+ one device type, one country, and one referrer all at once, that’s not a normal customer pattern — it’s a hallmark of a single automated source hitting your store repeatedly.
2. Zero engagement, every time. Look at average session length and pages per session for the traffic in question. Real shoppers browse — even a “no” browses. A session that lands and leaves in under a few seconds, over and over, on volume, isn’t someone deciding not to buy. It’s a script that already did what it came to do (usually: nothing you’d care about).
3. A conversion rate that’s suspiciously exactly zero. Every real store has some conversion rate, even a bad one. A specific day, referrer, or traffic segment sitting at a flat 0% orders across hundreds of sessions is a red flag, not bad luck.
4. Traffic that spikes with no real-world cause. A genuine traffic spike usually has a reason — an ad went live, a post went viral, a sale started. A spike with none of that, especially one that repeats on a schedule, is often a crawler or scraper running its own cycle, not people discovering your store.
5. Your platform’s own “human” label doesn’t match the above. This is the one most people miss entirely. Shopify (and most platforms) already try to label sessions as human or bot — but that label is based on a narrower set of signals than the four checks above, and it can miss real, ongoing bot activity for weeks at a time. Don’t assume “human” means human. Check the pattern yourself.

Why this actually matters

This isn’t a vanity metric problem. Every real number downstream inherits the distortion:

• Conversion rate looks artificially worse, because the denominator is padded with sessions that were never going to buy anything.
• Ad performance gets misjudged, since real, converting traffic gets diluted by noise in the same reports.
• Product and pricing decisions get made off a false signal — you can end up “fixing” a product page that was never actually the problem.

What to actually do about it

Short of building your own detection pipeline (which is a real, ongoing engineering project, not a weekend script), the practical options are:

1. Check manually, periodically — pull your device/referrer/country breakdown for a day that looks off, and eyeball it against the five signs above.
2. Don’t trust a single “human/bot” label from one platform alone — cross-check it against your own funnel data (did anything happen after the session started).
3. Use a tool built specifically to catch what your platform’s own classifier misses — which is the actual reason we started building yoursignalstack.

We built yoursignalstack because we kept running into exactly this problem on our own store, and no existing tool was honest about how often it happens. It’s not launched yet — but if you want to know the moment it is, leave your email below.

Get notified when yoursignalstack launches →