How to vet a link prospect without ranking promises

If someone offers us a link with ranking promises attached, the promise is the first thing to ignore. A worthwhile prospect is not one that claims authority, page power or quick movement in search. It is one that makes editorial sense, reaches a real audience, and can exist on the web without needing a sales script to justify it.

That is the practical approach to evaluating a link opportunity without buying ranking promises. We look at whether the proposed link belongs on that page, whether it helps the reader, whether search engines can actually access the page, and whether the request contains signals that the placement is manufactured rather than editorial. Useful evidence exists, but it is narrower than many pitches suggest.

What counts as a worthwhile link opportunity

A good link opportunity starts with relevance. Not keyword overlap, and not a tool score, but genuine topical and audience relevance. If a page discusses a subject our page materially helps with, a link may be sensible. If the connection only works after a long explanation, it usually is not.

Google’s own starter guidance is a better benchmark than third party authority metrics. It points back to useful, accessible content and descriptive links, while avoiding any suggestion that rankings can be guaranteed Google SEO Starter Guide. That matters because many link pitches are built around a promise that no publisher can honestly make.

For us, three criteria matter most.

First, editorial fit. Would the link still make sense if nobody were measuring SEO benefit? Could a site editor defend it as a useful citation, source, example or further reading link? If the answer is no, the placement is weak before we even inspect the domain.

Second, page level relevance. A respectable site can still offer a poor page. We care about the exact URL where the link might appear, not just the logo in the header. A gardening blog with one article about analytics might be fine for analytics software only if that article is well written, current and genuinely about the subject. Site wide reputation does not rescue a mismatched page.

Third, compliance with search guidance. Google treats links as part of how the web is discovered and understood, but not as tradable ranking guarantees. Existing search fundamentals also continue to apply to AI features, and Google does not require a special AI schema file or text file for those features Google AI features. So when a prospect claims they can sell us ranking uplift, AI visibility, or placement in AI overviews purely through one link, they are stepping beyond what the evidence supports.

This is also where we separate a publication decision from a buying decision. At The SEO Blog, we are teaching readers how to assess evidence. We are not treating every outreach email as an opportunity to accumulate links. If a prospect does not improve the reader’s journey, we do not need a metric to tell us to walk away.

A five-step check before you reply

Before replying to a link request, we use a simple five step screen. It is quick enough for a small team and specific enough to stop us being swayed by vague promises.

1. Check the proposed linking page itself

Open the exact page. Read it as a reader first.

Ask:

• Is the page about the same subject area as ours?
• Is the article coherent, current and clearly edited?
• Would a reference to our page genuinely help the reader finish the task or understand the topic better?
• Does the page already contain a clutter of commercial anchors, casino style topics, or irrelevant outbound links?

Look for signs of mass placement. These include generic list posts that exist only to host links, contributor pages with no editorial standards, and articles where every paragraph sends the reader elsewhere. If the page feels assembled for link inventory, that is usually what it is.

Also check whether the page can be found on the site through normal navigation, topic hubs or internal links. A page buried in an author archive with no internal prominence can still be indexed, but it often signals weak editorial value.

2. Check the page they want to link to on your site

A link can only be as defensible as the destination. If the prospect proposes a deep commercial page that does not answer the apparent topic, the fit may be wrong even if their page is decent.

We ask whether the destination page:

• directly answers the topic discussed on the linking page
• contains evidence, explanation or resources worth citing
• is indexable and accessible
• uses a clear title and heading structure
• would still deserve a link if rankings were not part of the conversation

Sometimes the best move is to offer a different destination. If someone wants to link to a service page from an informational article, a guide or explainer may be the better editorial fit. That is often safer and more useful.

If you need a quick internal sense check before deciding, our guide on checking your site for free before buying SEO tools covers a sensible starting process.

3. Check indexing clues, not just domain claims

A page does not help much if search engines cannot access it or choose not to index it. That does not mean every unindexed page is a problem, because some pages should be excluded, but it does mean we should verify the basics.

Useful checks include:

• a site: search for the exact or near exact title
• whether the page appears to have a cached or recently visible presence in search results
• whether the URL is blocked, canonicalised elsewhere, or obviously marked noindex
• whether your own destination page is indexable and not accidentally excluded

For your own pages, Google Search Console is the best direct source for indexing status and crawl information. Google’s indexing documentation is clear that excluded URLs need investigation by reason, and that not every excluded page should be indexed Google Page indexing report.

If you have server logs, crawler access can add useful context. A visit from Googlebot shows crawl activity. It does not prove rankings, traffic quality or business impact. The same caution applies to AI related crawlers. OAI-SearchBot and GPTBot are not the same thing, and search access versus model training controls are separate choices in OpenAI’s documentation OpenAI crawler documentation. We should not let a prospect blur those distinctions.

4. Check audience fit

This is the step most link pitches skip, and it is often the most important. Who actually reads the page?

A local trade association blog, a specialist software publication and a general lifestyle site may all have decent looking sites. Only one may have readers who would sensibly click through to our page.

Check:

• the publication’s subject focus
• whether comments, bylines or article style suggest a real readership
• whether the article’s intent matches our page’s intent
• whether the geography fits, especially for UK specific topics

For UK businesses, audience fit can also mean legal and practical context. A page written for a US market may not be a good linking environment for advice on UK search reporting, local visibility or website management. Where a rule differs between the UK and the wider EU, we should say so clearly. In this topic, the main issue is less legal divergence and more audience context. A prospect aimed at EU wide regulation news may still be a poor fit for a practical UK guide on search measurement.

5. Check risk signals in the outreach itself

The email often tells us as much as the site.

Warning signs include:

• guaranteed ranking claims
• offers to place multiple keyword rich anchors across several sites
• insistence on exact match anchor text
• unwillingness to name the exact page before payment
• mention of private networks, partner sites, or “our inventory”
• claims about AI visibility that collapse crawler access, citations and rankings into one promise

We also treat reciprocal link suggestions cautiously. There can be legitimate editorial cross referencing between related sites, but an automated link exchange scheme is not the same thing. If the proposal only exists as a trade, not because each page genuinely helps the other audience, we decline.

Worked example of a borderline prospect

Consider a realistic case.

A software comparison site emails us offering a link from an article about “how small businesses measure website performance”. The domain looks respectable. The design is clean. Articles have authors. The site appears in search. They propose linking to one of our measurement articles in exchange for a fee and suggest anchor text around Search Console reporting.

At first glance, this is not an obvious scam. It is a borderline prospect.

We start with the page. The article is broadly relevant. It discusses analytics tools, reporting dashboards and KPIs. A link to a page explaining what evidence Search Console can and cannot provide could help readers. So the topical fit is plausible.

Now the closer checks.

The article is two years old and lightly updated. It has several outbound links, most of them to commercial software pages. Some are useful. Some feel inserted. The section where our link would sit is a short paragraph about “SEO visibility tools”, but the article is mainly about general web analytics. That weakens the fit. Our page would help only if the copy around it clearly introduced search specific measurement.

Next, the destination. A page such as Search Console evidence and limitations is a better editorial target than a homepage or service page because it matches the measurement theme and explains evidence boundaries. So the destination can be made sensible.

Then indexing clues. The article is indexed. The site is crawlable. Nothing obvious suggests a blocked page. Good. But that only tells us the page exists in search systems. It does not tell us whether readers use it, whether the link will be clicked, or whether any ranking effect follows.

Audience fit is mixed. The publication serves small business buyers comparing tools. Our article serves readers trying to understand what Search Console can prove. There is overlap, but not perfect overlap. A careful editor could make it work. A sales broker could also force it.

The decisive step is risk. The outreach email promises “measurable ranking movement within weeks” and offers upgraded placement on “higher authority partner pages” for an extra fee. That language undermines the otherwise respectable appearance. It suggests the site may be selling placements as inventory rather than making editorial decisions.

In this case we would not accept the offer on the original terms. We might negotiate only if the publisher is willing to remove the ranking claims, identify the exact URL, let the editorial team decide the surrounding copy, and link to the most relevant informational page with natural anchor text. If they insist on a promise or a forced anchor, we decline.

That is what a borderline prospect looks like in practice. The site itself is not the whole answer. The page, placement, destination and sales behaviour matter.

What evidence helps and what it cannot prove

Useful evidence exists, but it has limits.

Helpful evidence includes:

• whether the linking page is indexed
• whether your destination page is indexable
• whether Googlebot can access the destination page
• impressions and clicks in Google Search Console after publication
• referral traffic from the linking page, if any
• changes in crawl frequency or discovery patterns

These signals help us understand visibility and access. They are especially useful when combined with a proper before and after investigation rather than a single screenshot. Our article on how to investigate a change in search performance is the right next step if something appears to move.

What they cannot prove is just as important.

A crawler visit does not prove a citation. A citation does not prove a click. A click does not prove an enquiry. An enquiry does not prove the link caused the sale. A ranking change after a link does not prove the link was the cause.

This matters because many pitches bundle all of these together. The same overclaim appears in AI search language. Access by OAI-SearchBot is not the same as GPTBot training access. Neither one guarantees mention in AI outputs. Google’s AI features documentation is also explicit that existing search fundamentals apply, and no special AI text file or AI schema is required Google AI features.

So when we are evaluating a link opportunity without buying ranking promises, we should keep the evidence ladder intact. We can observe impressions, clicks, referral visits and crawler access. We cannot honestly jump from those to guaranteed rankings, sales, or “AI visibility” in the abstract.

If you want a grounded way to think about ranking evidence itself, our guide to tracking rankings without fooling yourself complements this article well.

When to decline, negotiate or proceed

A final decision becomes easier once we separate three outcomes.

Decline

We decline when the link would not make editorial sense, when the page is low quality, when the audience is wrong, or when the outreach relies on guarantees and manipulative anchors.

Typical decline reasons:

• the page topic does not really match ours
• the site appears to sell placements at scale
• the proposed anchor text is forced
• the destination page is not the best fit and they will not consider a better one
• the pitch makes unsupported claims about rankings, sales or AI visibility

A short no is usually enough. We do not need to debate authority metrics with someone selling a promise.

Negotiate

We negotiate when the site and page could be relevant, but the proposed placement needs tightening.

Useful negotiating points:

• ask for the exact URL
• suggest a more relevant destination page
• ask for surrounding copy that accurately introduces the linked resource
• request natural anchor text that describes the destination
• remove any ranking or AI outcome claims from the discussion
• confirm the placement is editorially reviewed, not inserted into a bulk network

This is the safest path for borderline opportunities. It turns a vague offer into a clear editorial decision.

Proceed

We proceed when all of the following are true:

• the linking page is genuinely relevant
• the destination page genuinely helps the reader
• the placement is understandable without SEO jargon
• the page is accessible and indexable, where indexing is appropriate
• the audience fit is credible
• no one is promising outcomes they cannot prove

Even then, we proceed for the value of the placement itself, not because someone has promised a ranking result. The best links are often the least theatrical. They look like normal citations because that is what they are.

That is the practical standard we use at The SEO Blog. It keeps the decision grounded in relevance, evidence and editorial fit. It also protects us from wasting time on placements that sound impressive but collapse under inspection. If a prospect cannot stand without a ranking promise, it is usually not much of a prospect at all.