Is direct traffic good or bad for my store?

Direct traffic is usually neither good nor bad. It is GA4's label for visits it could not trace to a source. If it holds buyers who already know you, it is good news. If it hides untagged links, bots or ads you paid for, it is a blind spot that can talk you out of working channels.

What one store's data shows

The usual answer picks a side. Fans say Direct is loyal customers typing your name, so more is better. Sceptics say it is broken tracking, so more is worse. One store's export shows why the share alone cannot settle it.

One store's anonymised GA4 export, 1 January 2024 to 21 August 2026. It holds shares of revenue only: no ad spend, no order counts.

On the Channels sheet, Direct holds 57.7% of revenue in every view. That tells you how big the label is. It does not tell you whether the label is good news.

Two quiet rows show the problem. On the same Channels sheet, Organic Video and AI Assistant both sit at 0.0%. That can mean nobody came from a video or an answer box. It can also mean people watched or read about you there, then typed your name later. A typed visit counts as Direct.

If they typed, that is good demand wearing the wrong name tag. The video earned the visit and Direct collected the credit. The export cannot say which story is true, and GA4 cannot either.

On the Journeys sheet, journeys with 1 touch hold 79.5% of revenue, and those buyers took 0.5 days to buy. Quick, single-visit purchases are what loyal customers look like. They are also what a buyer sent by a friend, a podcast or an unclicked ad looks like.

So the honest reading for this one store is short. Direct is large, and the export cannot grade it. The grade comes from things the export does not hold: landing pages, engagement and dates.

Why does the usual verdict mislead?

Because Direct is three kinds of traffic under one label, and each deserves its own verdict. Google's help names the stakes: whether your campaigns drive traffic, whether credit lands in the right place, and technical faults to fix.

Earned direct is good. GA4 defines Direct as people who arrive through a saved link or by typing your URL. These are people who know where you live. They cost nothing per visit, which is the best price there is.

Borrowed direct is good demand with the wrong credit. Meta's help says its view-through setting counts events within 1 day after someone sees your ad. If that person later types your address, Meta can count the sale while GA4 records a direct visit. GA4's own troubleshooting also lists social media and offline documents among the examples of direct traffic.

Broken direct is bad. GA4's help lists the usual culprits: links without UTM tags, redirects that strip them, URL shorteners and ad blockers. All of them push visits into (direct) / (none). Bots can add to the pile. GA4 drops known bots automatically, but it does not show you how many it removed.

Only the broken kind should shrink. The borrowed kind is the expensive one to misread. A last-click report hands its sales to Direct. The ads that created the demand then look idle, and idle-looking ads are what a tight budget cuts first.

Even earned direct has a trap. If you bid on your own brand name, some people who would have typed your address may click your ad instead. You then pay for a visit that Direct would have brought in for free.

So when is a big Direct share bad news?

When it behaves like broken tracking rather than like customers. Three signs point that way.

• It lands deep. Typed visits tend to start on your home page. Direct visits that start on long product or offer URLs usually came through a link that lost its tags.
• It engages poorly. GA4 calls a session engaged if it lasts 10 seconds or longer, has a key event or views 2 or more pages. Direct far below your site total hints at bots or broken pages.
• It jumps on a change date. A new email tool, checkout or domain can move Direct overnight. Demand does not usually change on the day you change software.

The share itself is not a grade. The same share can belong to a well-loved brand or a leaky tagging plan, and only the behaviour tells them apart.

What to do this week

1. Ask Shopify how much of Direct is borrowed. In Shopify admin, go to Analytics > Reports, filter the Category to Marketing and open Performance by referring channel. In the Attribution menu, read Direct's sales under Last click, then under Last non-direct click. Pass: Direct keeps most of its sales, so most of those buyers had no other channel on record. Fail: a large part moves to paid channels, which opened journeys that Direct finished.
2. Score Direct against the rest of the site. In GA4, open Reports > Acquisition > Traffic acquisition and compare Direct's Engagement rate and Session key event rate with the totals row. Pass: Direct matches or beats the totals, so real shoppers are behind it. Fail: it trails far behind, so look for bots, broken pages or redirect chains before you credit the brand.
3. Put your dates on the chart. In the same report, right click a data point on the line graph and click Add annotation for each campaign launch and each tracking change. Pass: Direct rises when your campaigns run, which is borrowed demand to protect. Fail: it jumps on a tracking change date, which is broken direct to fix first.

Check the homework. Your GA4 Attribution paths export already holds the evidence. Causality Engine reads that one file and shows what each channel caused next to what last-click gave it, in 1 to 2 minutes, for €99 once (excluding VAT), refundable within 30 days. Check the homework

Sources, 1 October 2026: Understand (direct) / (none) traffic (Google); Default channel group (Google); User acquisition report (Google); Known bot-traffic exclusion (Google); Traffic acquisition report (Google); About annotations (Google); About attribution models and attribution settings (Meta); Bot filtering in Shopify analytics and reports (Shopify); Marketing reports (Shopify)

Related answers

• All Direct traffic questions
• How to tell if direct traffic is good or bad, step by step
• Is direct traffic good or bad for an electronics store?
• View-through conversion, defined
• Incrementality testing, explained