About ShelfGrader

ShelfGrader is a free tool that reads the public pages of an online store the way an AI shopping assistant does, then grades how visible your products are to assistants like ChatGPT, Claude, Gemini and Perplexity. You get a letter grade, the ranked fixes, and the competitor an assistant would pick instead. Checking is free. Having the fixes done is the paid part.

Who is behind it

ShelfGrader is built and run by Dan Kilkelly, founder of Clearway Apps. The legal entity is Kilkelly Enterprises, a sole trader trading as Clearway Apps (ABN 82 862 013 361), based in Western Australia. Clearway Apps builds small, useful software for online businesses. ShelfGrader is one of those products.

We are a small, hands-on team. The person who writes the guides on this site is the person who builds the grader, so the advice and the tool stay in sync.

How the grade works

ShelfGrader fetches the live, public HTML of the store page you submit and checks it against ten signals that decide whether an AI shopping assistant can read, trust and recommend a product. Those include Product schema, a machine-readable price, availability, structured ratings, crawlability and the trust signals an assistant looks for, such as returns and contact details.

We read the page the way an assistant does. Most AI crawlers do not run JavaScript, so a fact that only appears after the page loads scripts is a fact the assistant never sees. We score what is actually in the served HTML, not what a human eventually sees in a browser.

We hold ourselves to the same standard. ShelfGrader's own pages ship their facts and schema in the raw HTML, so the tool that audits structured data has clean structured data itself.

Our editorial and data standards

We built ShelfGrader on a simple idea: stores deserve a straight answer, not slop. That shapes how we write and how we publish data.

• No invented numbers. When we publish a study, the sample is real and the raw results are saved so the figures are checkable. We tell you the sample size and we do not dress a small sample up as a census.
• We flag confidence. If a finding comes from a modest sample or a fast-moving part of the field, we say so rather than implying certainty we do not have.
• We do not fake trust. No invented reviews, no fake ratings, no purchased testimonials. If a number is not real, it does not ship, including on our own site.
• We cite sources. Claims about how AI assistants and search engines behave are based on observed behaviour and credible public research, linked where it matters.
• We keep content current. AI shopping changes quickly, so we refresh guides as the ground shifts and date the updates honestly.

Talk to us

Questions, feedback, press or partnership: admin@clearwayapps.com.au, or see the contact page. You can read more about Clearway Apps at clearwayapps.com.au.