Build With Less AI

From Zero to First Inquiries: A Transparent Look at My AI-Built Store's Conversion Funnel

AI Export Lab · July 2026

As of July 1, 2026, jjradiator.com has 41 clicks from Google Search Console and exactly zero buyer inquiries. This article is about what I built before those 41 clicks arrived — the Request a Quote button, the form, the notification system, and the follow-up process that is currently waiting for its first real test. Most conversion advice is written after someone has data. I am writing this before I have any. When the inquiries start coming in, this article will either look smart or look wrong. Either outcome is useful.

Why I Built the Funnel Before the Traffic

The standard advice for a new B2B website is: get traffic first, optimize conversions later. I did the opposite.

Here is my reasoning: when you have 15,000 product pages, every page is a landing page. Someone searching "16400-0T020 radiator" lands directly on that product. If the page has no obvious way to ask about pricing or stock, that click is wasted — and at 41 clicks total, I cannot afford to waste any.

More importantly, I cannot optimize what I do not measure. If I add the Request a Quote button after 1,000 clicks, I have no baseline. I cannot tell whether the button helped or the traffic just got better. By setting up the funnel at zero, every click from day one has a chance to convert, and I have a complete record of what happened.

This is the same diagnostic principle I wrote about in Why a TCM Practitioner Approaches AI Automation Differently: observe before intervening. You cannot observe what you did not set up.

What the Funnel Actually Looks Like

This is not a sophisticated marketing stack. It is three things:

  1. A "Request a Quote" button on every product page. Positioned above the fold, right next to the product image. It links to a simple contact form. No popup. No chatbot. No multi-step wizard. One button, one form.
  2. A contact form that asks for three things. Name, email, and a message field. The product page auto-fills the subject line with the part number, so I know exactly which product the buyer is asking about. I deliberately did not add phone number, company name, or country fields. Every extra field reduces submission rates. I can ask for more details in the follow-up email.
  3. Email notification when a form is submitted. The form sends to my email. That is it. No CRM. No Slack integration. No lead scoring. When an inquiry comes in, my phone buzzes. I reply within a few hours.

Total setup time: about an hour. Most of that was testing the form submission and making sure the email arrived. The form itself is a simple PHP script that the Cloudflare Worker forwards to — no external service, no monthly fee.

If this setup sounds too simple: it is. But at 41 clicks, I do not need a marketing automation platform. I need to know when someone wants to buy radiators. The rest can be built later, when there is actual demand to automate.

The Numbers That Matter (Even at Zero)

I am tracking three numbers. All of them are currently zero. That is fine:

I am publishing these numbers at zero because anything else would be a success story pretending to be a case study. The next update will have real data, whatever it looks like.

What I Expect to Happen (And Why I Might Be Wrong)

I am writing this section before the data comes in. That way I cannot retroactively pretend I predicted the outcome:

My expectation: The first inquiry will come somewhere between 100 and 300 total clicks. It will be a short, low-effort message — probably just a part number and "price please" or "stock?" It will come from a country I did not expect. The buyer will be a small repair shop or a solo mechanic, not a large distributor.

Why I might be wrong: B2B buyers searching for specific part numbers may have higher purchase intent than I am assuming. The first inquiry could come sooner, from a more serious buyer. Alternatively, the buyers searching for part numbers may be comparison-shopping across multiple suppliers and have no intention of inquiring — they just want to see if I list a price. (I do not list prices. This is deliberate. It forces an inquiry if someone is serious.)

What would make me change the funnel: If I hit 500 clicks with zero inquiries, the form is broken — either the button is in the wrong place, the form asks for too much, or the page does not communicate trust. I would test a shorter form, a more prominent button, or a direct email link instead of a form. But I will not change anything before 500 clicks. Premature optimization is guessing. I would rather have 500 clicks of clean data and one clear change to make than 500 clicks of random tweaks and no idea what worked.

What Most B2B Product Sites Get Wrong

I looked at about 20 small B2B export websites while building mine. Here is what I saw repeatedly:

These are not opinions based on A/B tests. They are patterns I noticed as a buyer myself — I have sourced auto parts for years as part of my family's business. I know what a serious inquiry looks like from the buyer's side, because I have sent them. The form I built is the form I wish I had been given.

The Follow-Up Process (Waiting for Its First Test)

When an inquiry comes in, here is what I plan to do:

  1. Reply within a few hours. Not minutes — I have a day job. But within the same business day. A fast reply signals reliability more than any badge or certificate.
  2. Answer the question directly. If they asked for a price, give the price. If they asked about stock, confirm availability. Do not send a brochure. Do not ask for a phone call. Answer the specific question they asked.
  3. One follow-up after 3 days. If they do not reply, send one short check-in. If they still do not reply, let it go. Desperation is not a sales strategy.
  4. Track everything in a spreadsheet. Product code, inquiry date, country, question type, whether they replied, whether it led to a sale. At low volume, a spreadsheet is better than a CRM — you actually look at it.

This process is intentionally lightweight. At 41 clicks, I do not need automation. I need to personally handle the first 10 inquiries and learn what buyers actually ask about. That knowledge — not software — is what will let me optimize the funnel later.

What I Will Report Next

I will update this article or publish a follow-up when one of three things happens:

  1. The first inquiry arrives. I will share exactly what the buyer asked, how I replied, and whether it led anywhere — with identifying details removed.
  2. 500 clicks with zero inquiries. If the funnel is broken, I will show the data, describe what I changed, and report whether the change worked.
  3. An inquiry turns into a sale. This is the milestone that matters. Not a click. Not a form submission. Money changing hands. When that happens, the conversion funnel stops being theoretical.

If you are building a B2B site and want a second pair of eyes on your inquiry funnel before you start driving traffic, that is literally what I do: Work With Me. I look at B2B product sites and tell you what is likely to convert and what is likely to leak.

Related Articles

→ Back to AI Export Lab — all build notes from this experiment.