Traffic But No Inquiries? Check Who You Write For

Your independent website (your own brand site) gets traffic but no inquiries, and most people first suspect traffic volume, the product, or the price. But one change tends to go unnoticed: the point at which a visitor makes their judgement has moved forward — to somewhere outside your website. They are not coming in to discover you. They are coming in to verify you.

Let's be precise. Traffic without inquiries can be a traffic problem or a conversion problem, and the two call for opposite actions. This article covers only a third case: the pages and the traffic are both fine, but what the visitor needs is not on your website.

First, separate "nobody is coming" from "visitors arrive but don't convert"

"The site gets visits but no inquiries" is the easiest thing to misdiagnose, because the two causes call for opposite actions.

The first: nobody is actually coming. Daily visits sit in single digits, search engines have indexed few pages, and your keywords do not rank. This is a traffic problem, and the work is indexing, keywords and content building.

The second: people arrive but do not convert. Daily visits run into the dozens or even hundreds, and inquiries are zero. This is a conversion problem — and putting more ad spend in now is just tipping more traffic into a leaking bucket.

The way to tell them apart is straightforward: look at bounce rate and average time on page. A high bounce rate with only a few seconds on the page means visitors came in, did not understand what you do, and left. This is also the first thing AI search optimization (GEO — Generative Engine Optimization) has to establish: is the problem at the entry point, or on the page?

Judgement now happens outside your website

In the past, a buyer looking for a supplier followed a linear path: search, scroll the results, open sites one by one, compare, send an inquiry. Your website was the discovery entry point — the first time they learned who you were was on your homepage.

That path has been compressed. More and more buyers explain their requirement to AI first, get a handful of candidates, and only then verify them. In that flow, your website has moved from a "discovery" entry point to a "verification" entry point.

There is measured data on this. A behavioural study published by Pew Research Center in July 2025 tracked 900 US adults across 68,879 Google searches (March 2025):

• When an AI summary appeared on the results page, 8% of users clicked a traditional search result; without a summary, 15% did
• Only 1% clicked through to a source link cited inside the summary
• When a summary appeared, 26% of visits ended on the results page itself (16% without one)

(This is an observational study, not a randomised experiment, so it cannot be read as direct causation — but the direction is clear enough.)

Put into a scene: most of the buyer's question has already been answered on the search results page. They click into your website not because they still do not know the answer, but because they want to confirm "is this company actually reliable".

And in China, the number of people who put questions to AI reached scale long ago. A report from China Internet Network Information Center (CNNIC) shows that as of June 2025, China had 515 million generative AI users, up 106.6% in half a year; among them, 80.9% used generative AI to "answer questions" — the largest share of any use case. In other words, the first thing most people do with AI is ask it something — and sourcing research is, by nature, a string of questions.

"Ask AI first, then decide who to approach" is no longer a trend that needs arguing. For an independent website it means the visitor's identity has changed: they are not hearing about you for the first time — they are here to check up on you. This is also why AI search optimization keeps coming up.

Your website now does one job: being verified

If a visitor comes in to verify, what they are looking for is evidence; and what your website most likely offers is a self-introduction.

That right-hand column is exactly what GEO has to deal with. AI search optimization is not about making the website look better — it replaces the parts that carry only adjectives and no facts with fact blocks that can be extracted and verified.

Because the logic AI uses to decide whether to cite you and the logic a visitor uses to decide whether to contact you are one and the same: it can be read, it matches up, it can be verified.

One business description, two ways to write it

One and the same business description is enough to make the difference clear.

Written for a human to read: "We have advanced production equipment and a professional R&D team; our products are exported worldwide and well received by customers."

Written for AI — and for visitors AI has already screened: "MOQ 500 pieces, sampling 3–5 days, mass production 15 days; 304/316 stainless steel and aluminium alloy; dimensional tolerance ±0.01mm; stainless steel parts machined for customers in Germany and Poland, about 20,000 pieces shipped in 2025."

Not one sentence in the first version is false, and not one sentence can be verified. Every sentence in the second version is a coordinate someone can check — and that is the minimum unit generative engine optimization asks for.

AI does not grab attitudes; it grabs coordinates. Without coordinates it can only name the competitors who wrote theirs down — and a visitor who lands on a page full of adjectives has no reason to leave a contact detail. This is the fundamental difference between AI search optimization and "making the website pretty": the first solves verifiability, the second solves appearance.

Visitors from AI really do behave differently

Someone will ask: is there data behind this?

Yes, but one distinction has to come first: most of these measurements come from English-language markets and cross-industry samples, and their definitions differ widely (more on that in the next section). The usable sets are these:

• Traffic to websites from AI search engines grew about 16-fold between 2024 and 2026 (SE Ranking research)
• AI platforms currently account for 0.32% of total website traffic — it was 0.24% in 2025 and 0.02% in 2024
• Average time on site for AI-referred visitors is 68% longer than for organic search visitors
• Another study, with a sample of 1,215 domains and 3.3 billion sessions, measured that share at 1.08% (Conductor, 2026 AEO/GEO Benchmarks)

The most valuable item here is not "how much traffic AI brought" but the 68% longer dwell time. It points to a judgement: visitors who arrive from AI are further along in the decision stage. They are not here for a casual look; they came with questions to verify.

This explains the core symptom in this article exactly: if visitors come to verify, and your page offers only a self-introduction, then the more traffic you get, the more you waste. So in AI search optimization the first thing to understand is this — the problem is not how much traffic you have, but whether your content can be treated as evidence.

On "how much traffic AI takes" — why I won't give you a number

Before writing this I went through the research. I had intended to give a figure for "AI has already diverted XX% of website traffic". After checking, I decided not to — the numbers different sources measured differ by nearly two orders of magnitude:

• By one definition, AI search already affects 12%–18% of global referral traffic
• SE Ranking measured AI platforms at 0.32% of all website traffic
• Conductor measured 1.08%
• Chartbeat data shows ChatGPT accounts for only 0.02% of publisher referral traffic

From 0.02% to 18% is a difference of several hundred times. The gap comes from definitions: "referral traffic" or "all traffic", global or a single market, corporate sites or content sites — change the denominator and the conclusion flips completely.

A set of contradictory numbers does not help you decide anything. It only helps someone sell something.

Two things can be confirmed: the direction is growth (multiple independent measurements point upwards); and the absolute share is still small.

And that "small" is exactly what shows something more important: if your independent website is getting no inquiries right now, the reason is probably not that "AI has taken the traffic away". The traffic is still there — the people coming have changed, and so has their purpose. The location of the problem has changed too: from the "entry point" to the "content". That is the stretch GEO has to deal with.

Five things to change

First, explain three things in one sentence above the fold. What you do, how far you can go, and how to make contact next. Do not fill the first screen with drone shots of the factory or a carousel — visitors want a basis for judgement, not a show of scale.

Second, replace adjectives on product pages with numbers. Turn "reliable quality" into a tolerance range, "fast delivery" into sampling and production days, "flexible MOQ" into a minimum order quantity. This is the highest-return step in AI search optimization, because the material already exists in your sales team's heads — it just never made it onto paper.

Third, write certifications and qualifications out in full, and make them checkable. Certification names and numbers are far more useful than "certified to several international standards". This is the category of content AI is most willing to cite — information a third party can verify.

Fourth, structure it. Use the customer's own question as a subheading, put comparisons in tables, and add a short FAQ at the end. What these have in common is that they lower the machine's cost of understanding: subheadings, lists, tables and Q&A can be extracted directly by AI, while long unbroken prose is hard work for it to read.

Fifth, align terminology across every channel. Company information on your website, your marketplace store and public content — founding date, address, scope of business — has to be one single version. AI cross-checks across multiple sources before citing, and when they do not line up it would rather not use you. What is required is consistency of facts, not identical wording; pasting the same paragraph verbatim across the whole web can instead be judged duplicate content.

Your website and your marketplace store are two ends of the same GEO job

One easily confused point needs to be made clear here.

The on-platform end: the platform itself is becoming AI-driven, and whether your store profile and product information are structured and complete directly affects whether AI can read them in. This step can be done in the back office.

The off-platform end: that is the independent website this article is about. When buyers ask questions in AI, what AI reads are public web pages — your own website, media coverage, discussions in industry communities. The platform cannot manage this end; you have to build it yourself.

The two ends are not a choice; they are two positions of the same job. Do only the on-platform end and you may already be off the candidate list when a buyer asks AI off-platform; do only the off-platform end and the on-platform information is too thin to support AI's judgement that "this company can do it". Generative engine optimization was never a channel question — it is a repositioning of information.

There is also a hierarchy of evidence: self-evidence (specification pages on your own website), third-party evidence (customer cases, discussions in industry communities) and authoritative evidence (certification bodies' public records, third-party test reports). Authoritative evidence sets the ceiling on citations; without it, the efficiency of the other levels is noticeably constrained. Self-evidence alone is just blowing your own trumpet.

Shenzhen Yihang Network Technology Co., Ltd. had always worked on Alibaba.com-related services and expanded into GEO in 2024; these two things connect in our practice — one end is how to write your store information, the other is how to build off-platform sources such as your own website. We have served a good number of export businesses in the Pearl River Delta and the Yangtze River Delta. Our method is published as public material at https://www.chyihang.com/en/geo/ — go and have a look.

Three questions about independent website content that need straight answers

My website doesn't get much traffic anyway. Should I still work on content first?

Yes — but the order changes. If daily visits are in single digits, fix indexing and keywords first; that is a traffic problem. But content is the base of the same job: a page AI can read is also a page search engines index more easily.

If I publish MOQ and sampling days, won't that scare customers off?

The cost of not writing them is higher. When a buyer asks AI, AI can only give the numbers it can find; if you hide yours, you are handing that customer to the competitor who wrote theirs down. Spelling out your limits actually filters out inquiries that were never a fit. This is also the layer of AI search optimization hardest to replace — what it demands is truth, not sales talk.

How long before I see a change?

No promises on timeline, no promises on outcome. Any claim that you will "be in AI recommendations within 90 days" lacks verifiable support. The openly observed pattern: on time-sensitive platforms, content weight concentrates in the first week or two; on some platforms it takes about three months; evergreen content keeps getting cited. GEO is continuous content engineering, not one batch of articles and done.

A closing note

An independent website with traffic but no inquiries is the easiest thing to misdiagnose as a traffic problem.

But the real change is this: visitors no longer get to know you through your website — they confirm you through it. By the time they arrive, AI has already answered a round of their questions. What they want is evidence.

So the problem is not how much traffic you have, but whether what you write can be treated as evidence. Writing the answers out is something you do for visitors and for AI at the same time — the logic by which both judge you is one and the same.

Have you ever opened your website analytics and looked at how long visitors stayed? Tell us in the comments.

References

1. Pew Research Center, "Google users are less likely to click on links when an AI summary appears in the results," 22 July 2025 (sample: 900 US adults, 68,879 Google searches). pewresearch.org
2. 中国互联网络信息中心 (CNNIC — China Internet Network Information Center), Report on the Development of Generative AI Applications (2025), released 18 October 2025. cnnic.net.cn
3. SE Ranking, "Analysis of Top AI Search Engines: Who Is Catching Up to ChatGPT?" (AI referral traffic research covering 2024–2026). seranking.com/blog/ai-traffic-research-study
4. Conductor, 2026 AEO/GEO Benchmarks Report (sample: 1,215 corporate domains, 3.3 billion sessions, May–September 2025). conductor.com

Data and scope note

The behavioural research and traffic statistics cited in this article come from third-party research organisations, with samples drawn mainly from English-language markets and cross-industry corporate websites. They are used to illustrate structural change in how information is obtained and how visitors behave. They do not mean that any specific industry or individual company will necessarily receive inquiries or orders. The source composition and recommendation mechanisms of each AI platform are not officially published, and behaviour may change as platforms iterate. Before acting on this, a company should still assess its own products, target markets and content base independently.

#ExportWebsite #Export #AISearch #GEO