"We don't have a lot of people coming to our site. It's all phone calls and emails." That's how a company defended their website to me: a contract machine shop making production parts for OEMs, whose buyers are purchasing agents and design engineers who email drawings for quotes. If it sounds like yours, check before you decide the website doesn't matter. Sometimes a website gets traffic but no leads, and nobody inside has looked at the traffic. Your own analytics can give a first answer in a few minutes: how many sessions arrive, how many are direct visits that land on the homepage, and how many homepage visitors ever reach the contact or quote page. Those first figures include your own staff, so treat them as an upper bound, and re-run the check once a staff filter has a few weeks of clean data. It will tell you whether the belief that hardly anyone visits is a fact or a feeling.
What did one company's analytics actually show?
When they said it, I asked who had analytics access. After the call, they connected us, and we looked. The traffic was large in relation to the calls they said they were getting. Most of it was direct traffic to the homepage, which I read as people who already knew of the company. And from the homepage to the contact page, the numbers fell away sharply.
How I'd sum up what we saw: "You're seeing a dramatic drop-off from that to the contact. There's something there." In that case, the homepage didn't really explain what the business did. That's unnecessary friction.
Why is my direct traffic so high, and what does it mean?
It's easy to misread.
Google defines the Direct channel in Google Analytics 4 as visits that arrive "via a saved link or by entering your URL" (Google Analytics Help, Default channel group). That isn't someone who typed your company name into Google and clicked the result. The same page files that visit under Organic Search, the channel for non-ad links in search results. In session reports, a returning visitor's direct visit goes to their last non-direct source (Google on scopes), so Direct is mostly visits GA4 can't trace to any source.
Google is also plain that "(direct) / (none)" means traffic "that doesn't have a clear referral source." Visits from links in PDFs and Word documents land there, untagged links can lose their source, and ad blockers can push visits into it too (Google Analytics Help, Understand (direct) / (none) traffic). So can someone who reads an AI answer and types your address later, as we note in AEO vs SEO. Clicks straight from an AI assistant's answer are different: GA4 files those under its own AI Assistant channel when the referrer matches a known assistant, while clicks from Google's AI Overviews and AI Mode count as Organic Search (Default channel group).
So the honest reading is a mix: people with your address but no traceable earlier visit, your own team, and visits with lost sources. Direct on its own says nothing about intent.
How do you run the same check on your own Google Analytics 4?
Each point links to Google's own steps. Use the same dates for every report, and keep them inside the funnel's reach (point 4).
- Your own staff and existing customers are in Direct, or in whatever channel last brought them in. Staff whose browser opens on the website land there, and so can customers who always type your address. You can exclude office traffic by IP address with an internal traffic filter (Google's guide to filtering internal traffic), but filters only work going forward: "Data filters do not affect historical data" (Google on data filters). An exclude filter is also permanent, because the excluded data is never processed; Google offers a Testing state to validate it first. An office IP filter won't catch staff on phones or at home, so even after a few weeks, treat homepage figures as an upper bound, Direct most of all.
- Compare visits with the calls and emails you actually get. The Traffic acquisition report shows sessions by channel: Direct, Organic Search, Referral and the rest (Google's guide to Traffic acquisition). Look at Engaged sessions next to the total. Google counts a session as engaged if it lasted 10 seconds or longer, had a key event, or had two or more page views. Put both beside the calls and emails you got over the same dates. You're only checking whether visits outnumber them by a wide margin, not matching them.
- Look at which page people land on, and by which channel. The Landing page report shows the first page of each session, and Google suggests adding Session source / medium as a secondary dimension to see the source behind each landing page (Google's guide to the Landing page report). That shows how much of Direct starts on the homepage.
- Compare how many homepage visitors ever reach the contact or quote page. A Funnel exploration with two steps, homepage then contact page, does this (Google's Funnel exploration guide). It counts users, and a user can only enter the funnel once, so compare its two steps with each other rather than with the session totals. Apply a new-users segment and compare it with everyone, so staff and returning customers don't dominate the drop-off. It also only reaches back as far as the property's data retention: that setting "only affects explorations and funnel reports", and dates older than it aren't visible (Google on data retention). Retention defaults to two months; more in what it means when you can't see your own analytics. Phone and email links in your header bypass the contact page entirely, so the funnel misses them.
What can't the numbers tell you?
Analytics counts sessions and users, not people with names. It can't tell you whether those homepage visitors were buyers, job applicants, suppliers checking you out or customers looking up a phone number. It won't show the visitor who read the homepage, picked up the phone and called. Phone calls don't show up at all unless you've set up call tracking. Which website numbers are worth tracking covers call tracking and conversion goals, if neither is set up yet.
So a steep drop from homepage to contact page isn't proof that people left unhappy. Some of them called. What it does tell you is that, if staff are accounted for and visits still outnumber calls and emails by a wide margin, a lack of visitors isn't the problem. People are coming. The question becomes what they find.
Is a big drop-off from the homepage a reason to act?
Not on its own. The way I'd put it: "Is it going to kill your business? Probably not in that case, but if you see that, why would you not want to fix that? That's just one piece of data."
It's one piece of data from your own account, and it turns the belief that the website doesn't matter into a narrower question. Most of the time, someone recommending you passes on your name, not your web address. The person they told lands on the homepage knowing who you are but not yet what you do. Can they tell what you do and how to reach you?
For that company, the first part was the problem. The homepage didn't say what the business did, which we've written about in why your homepage doesn't say what you do. If you want an outside read on yours, our homepage benchmark scores it on eight checks, two of which are exactly this: does it say what you do, and is there an obvious next step. Either way, you'll be deciding from your own numbers instead of from the feeling that nobody's looking.
