Would you pursue a relevant small site with little traffic data?

Here is a prospecting exercise: a small specialist publication has a closely relevant audience and thoughtful articles, but the available traffic estimates are sparse. A larger general site has stronger reported metrics and almost no overlap with the topic.

I would investigate the specialist site first. The questions would be who it serves, whether its recent work is maintained, how it credits sources, and whether the proposed reference would help a reader.

A metric can help prioritise investigation, but it cannot explain why a particular link belongs in an article. I would avoid turning a score threshold into the entire quality test.

What would make you reject the small site?

Sparse estimates should be marked “uncertain,” rather than converted into a confident zero. Different tools may have little visibility into a specialist site's actual readership.

I would record the evidence available and what remains unknown. That lets the team compare prospects without pretending every row has the same quality of measurement.

I would reject an approach that has no useful reason for contacting the editor. Relevance does not create an obligation to publish a link.

Look for an actual contribution: an original source, a correction with evidence, or an example that improves a specific article. If the pitch is only “please link to us,” the prospect's size does not rescue it.

My rejection checklist would focus on misleading content, irrelevant placement, unclear ownership, or a proposal built around buying an unqualified ranking link.

For a promising small site, I would write down the exact reader benefit before adding it to outreach. That gives us a meaningful next step and a clear reason to stop if there is no fit.

A small specialist directory may be worth investigating if it serves an actual audience and maintains its entries. I would try its browse path before looking at a score.

Also check whether business details can be corrected later. A useful reference should remain accurate after publication, not simply exist on the delivery date.