A bad provider has never met a budget they didn't like.

If you remember one line from this, make it that one. Everything below is really just a longer way of explaining how to spot it before you have signed anything.

I run a link building agency, which makes this an awkward piece to write honestly. I am describing the questions I would want a prospect to ask me, including the ones that are inconvenient to answer. But the alternative is another agency blog post telling you to check case studies and look for transparency, which helps nobody.

Why this choice punishes you more than most

Every agency decision matters. This one has a specific trap in it.

SEO results do not arrive in the first few days or the first few weeks. Everyone in the industry knows this, and it is genuinely true. Which means a bad provider can pull the wool over your eyes for a long time, and the honest explanation for slow progress and the dishonest one sound identical for the first three months.

You cannot solve that by waiting to see results. By the time the results tell you, you have spent two quarters of budget. You have to solve it on the way in, at the point where you are still just talking to someone.

So the whole thing comes down to the quality of one conversation.

The single best signal: do they try to disqualify you?

Here is what I would watch for above everything else.

A good agency is trying to work out whether they can actually help you. That means they ask about your business rather than your SEO. How do you currently get customers. Which pages convert. What is the sales cycle. What does a closed deal look like.

And then, crucially, they remember the answers. You can tell within two calls whether someone is assembling a picture of your business or running a script. When it is real, they refer back to something you said earlier and build on it. When it is not, every call starts from zero.

What you want to see is them qualifying you, or disqualifying you. There should be a version of your business where they say no. We turn work away when the budget is too low to produce anything meaningful, and we say so plainly, because one link a month serves nobody. It does not work for the client and it does not work for us, and it is better to have that conversation at the start than to take the money and disappoint someone for six months.

An agency with no disqualifying criteria is not being generous with you. They are telling you their process does not depend on fit.

Ask them why

The most useful question is also the bluntest. Why would we work with you?

Not to be difficult, but because how someone answers tells you how they think. You are listening for whether they are willing to be honest about what works and what does not, including the parts that do not flatter their own service.

Nobody legitimate is going to promise you results. If someone guarantees positions or guarantees revenue, that is the end of the conversation. Google says this itself in its own guidance on hiring an SEO: no one can guarantee a number one ranking, and you should be wary of anyone claiming a special relationship with Google.

But there is a wide gap between refusing to guarantee outcomes and refusing to explain the mechanism, and the good providers live in that gap. They will tell you what they expect to happen, why, and what would have to be true for it to work.

What case studies actually prove

Case studies matter. They just prove less than they appear to.

Every agency cherry-picks. The big ones, the good ones, all of them. It is the nature of the business, and you would do the same. Nobody publishes the engagement that went sideways.

Video case studies move me more than written ones, because it is harder to fake a person on camera than a paragraph on a page. But even there, be clear about what you do not know. You do not know the full context, how long it ran, how much they spent, or whether the testimonial came with a discount attached. A client who was given money off in exchange for a video is not lying. They are just not a neutral witness.

And I would not rule out an agency simply because it has no case studies. Everyone starts somewhere, and smaller agencies frequently care far more about the outcome because they cannot afford not to.

Here is how I would weigh each kind of proof they put in front of you.

If you want to see what the format looks like from the outside, we keep a roundup of SaaS link building agencies that includes us, which should tell you how much weight to put on any listicle including its author.

Ask to see the actual links

This is the question most founders forget, and it is the one that cuts through metric arguments fastest.

Ask for a sample of links they have built that are relevant to you. If they cannot show yours, ask for a different client's. They may be limited by NDAs and they may not be able to show you everything, but they should be able to show you something. Some agencies go further and let you pre-approve every placement before it goes live, which is a different structure entirely, but worth knowing exists.

Then look at where those links actually sit. Not the domain rating. Whether the site ranks for the things its own business cares about.

If you are evaluating two agencies and the only real difference between their pitches is the domain rating number they quote, you are not comparing quality. You are comparing who is more comfortable quoting a manipulable metric. Domain Rating is Ahrefs' own logarithmic score based on a site's backlink profile, and because it is calculated from links it can be inflated with a Fiverr service for about five dollars.

For what it is worth, our own published minimum is a domain rating of sixty with a five thousand monthly organic traffic benchmark, and the average of what we actually place lands closer to seventy or seventy-five with thirty to fifty thousand visitors. I am giving you both numbers deliberately, because an agency quoting only its ceiling is quoting the exception, and an agency that will not tell you its floor does not have one. Our thinking on where that bar should sit is in our piece on high authority placements, and the prices attached to each tier are in our cost breakdown.

Are they testing on themselves, or on you?

A question I like: do they research and test these things themselves, or is your site going to be the test?

There is nothing wrong with an agency learning. There is something wrong with an agency learning on your budget without telling you. If they can point to their own research, their own site, their own experiments, you are buying informed judgement. If every answer is theoretical, you are funding their education.

Related, and often overlooked: find out who you will actually be working with. I prefer to work with the founder of an agency where possible. With a larger agency that is impossible, and that is fine, but you want someone involved rather than someone who hands you off. The person selling to you has just spent hours learning your business. If they disappear after signature, that knowledge goes with them.

How they answer the measurement question

Ask how results will be measured and how ROI will be tracked. The answer tells you almost everything.

A bad answer promises a number and a date. By month three you will have this much traffic, or this many demos, or this many signups.

A good answer explains how the work connects to pipeline, and is honest about where attribution gets murky. Here is how we measure ours.

Page-level keyword positions. If we build links to a specific page and that page moves, we can reasonably attribute the movement. This is actually easier for a link-only engagement than for a full service one, because we are not touching everything else at the same time.

AI citation tracking. We take a baseline set of buying prompts and re-run them over time to see whether the brand appears more often. This is newer and it is becoming a bigger part of what we report on.

Share of voice and trajectory. How frequently you appear in your category, and whether the direction is consistent. Not one month's number.

Inquiries from organic search. Are they happening, and are they increasing.

Whether the client stays. The most honest metric in the industry. Nobody keeps paying for months and years if it is not working, and retention is harder to fake than a screenshot.

Each of those has a ceiling, and an agency worth hiring will tell you where it is.

The AI citation column is the newest of these and the one most likely to be sold badly. If you want the evidence behind why we track it against rankings rather than traffic, it is in our citation gap research.

What you will not get from me is a revenue claim. I genuinely dislike it when agencies say they did SEO for a company and now that company has four times the revenue. There are so many factors feeding that number. The product changed, the sales team changed, the market moved, they raised money, they hired. Claiming the lot is not confidence. It is a category error, and anyone in the industry reading it knows that.

One of those answers is worth knowing in advance: most agency inventory comes from exchange networks, and what a partner pool actually is tells you what you are really buying.

The questions, in one place

If you take nothing else into the call, take these.

1. Why would we work with you?
2. What kind of client do you turn down?
3. Can I see a sample of links you have built, ideally for someone like me?
4. What is your floor for quality, and what is your actual average?
5. Do you test these approaches on your own properties, or will mine be the first?
6. Who will I actually be working with after I sign?
7. How will we measure this, and at what point should I expect to see something?
8. What would make you tell me this is not working?

That last one is the one I would listen hardest to. An agency that has never considered how it would know it was failing has not thought very hard about how it succeeds.

A closing note on the awkward bit

But the providers who are good at this want these questions, because they are the ones where the honest answer is also the impressive one. The providers who are not good at it want you to compare domain rating averages and look at a case study.

If a single question from that list gets a defensive answer, that is your signal. It is much cheaper to find out on a call than in month six.

And if you are still at the stage of deciding whether to hire anyone at all rather than run this yourself, that is a different question, and I have answered it separately in the honest maths on buying versus building.

You are also welcome to point all eight questions at us. What we actually deliver on a link building retainer is written out in public, including the monthly volume and the quality minimum, so you can check our answers against the list before you ever get on a call.

Frequently asked questions

What questions should I ask a link building agency?

Ask why you should work with them, what kind of client they turn down, for a sample of links relevant to you, their quality floor and their actual average, whether they test on their own properties, who you will work with after signing, and what would make them tell you it is not working.

Should I trust an agency's case studies?

Treat them as evidence, not proof. Every agency cherry picks its best results. Video case studies are harder to fake than written ones, but you still do not know the budget, the timeline, or whether the testimonial came with a discount attached.

Is domain rating a good way to compare agencies?

No. Domain rating can be inflated cheaply and says nothing about relevance. If the only difference between two pitches is the DR number quoted, ask instead whether the linking sites rank for the terms their own businesses care about.

How should a link building agency measure results?

Page level keyword positions on the pages being targeted, AI citation tracking against a baseline set of buying prompts, share of voice, and inquiries from organic search. Be sceptical of any agency attributing revenue multiples to SEO alone.