Most agency sales calls are a discussion of what the agency says it will do for you. That is the least reliable information available, because it costs nothing to say.

There is a much better signal sitting in front of you, and you can check it before the call: the agency's own website. It is the one property they control completely, with no client to blame, no budget constraint they did not set themselves, and no approval process slowing them down. Whatever standard they hold themselves to there is the ceiling of what they will hold for you.

Here is how to run that audit in about twenty minutes, ordered so the most revealing checks come first.

1. Try to verify one testimonial

Take the named businesses in their testimonials and search for them. Not the person, the company. A real local service business has a Google Business Profile, a website, reviews, or at minimum a Secretary of State registration in the state they claim.

If a named business does not surface anywhere, you have learned two things. First, the testimonial is unverifiable. Second, and more important, the agency was willing to publish a specific attributed claim without expecting anyone to check.

Note what you are testing here. A vague testimonial, "Mike R. HVAC owner", is weak proof but it is not falsifiable. A specific one (a named company, a named city, and a claimed booking rate) is a factual assertion. Specific and unverifiable is worse than vague, because it means someone chose the specificity for credibility without earning it.

This also matters legally, which tells you something about the operation's risk tolerance. The FTC's Endorsement Guides, at 16 CFR Part 255, govern testimonials in advertising, and the Commission's Rule on the Use of Consumer Reviews and Testimonials (effective October 2024) makes fake and fabricated reviews subject to civil penalties. An agency that has not thought about this on its own site has not thought about it on yours either.

2. Count the sources under their statistics

Agency sites are dense with numbers. "97% of customers research online." "$42 returned for every $1 on email." "3x more booked jobs."

Look for a source next to each one. Then look at what kind of number it is, because there are three very different categories and they get deliberately blurred:

Industry statistics are claims about the market. These should carry a named, linkable source. If they do not, the agency either does not know where the figure came from or would rather you did not look. The $42-per-$1 email figure is the classic example — it traces to a 2019 DMA/Litmus survey of marketers reporting on themselves, and it gets republished endlessly with no methodology attached.

Scope commitments are promises about what the agency will do: channels managed, time to launch, response time. These need no external source — they are contractual. Check they appear in the actual agreement.

Outcome claims are assertions about results clients got. "3x more booked jobs" with no client attached is the one to push on. Ask: three times what baseline, over what period, for which client, and can you speak to them? A real result has a name attached and a client willing to confirm it.

The tell is not that an agency uses statistics. It is whether the three categories are kept honestly separate.

3. View source on their product screenshots

If the site shows a dashboard, look at the URL in the browser chrome of the screenshot. Then try loading it.

Plenty of agencies show a product mockup, and there is nothing wrong with an illustration, as long as it is labeled as one. What you are checking is whether they present a mockup as a live product. If the screenshot shows `app.theiragency.com` and that hostname does not resolve, they are showing software they do not ship, usually populated with invented client names and numbers.

An illustration labeled "illustrative" is fine. An unlabeled mockup of a product that does not exist is a statement about how they handle the gap between what they have and what they claim.

4. Click their client login

This one takes four seconds and is remarkably informative.

If they sell a "proprietary platform" or "custom CRM build" and the sign-in link goes to a third-party SaaS login page, you now know what the platform actually is. That is not automatically disqualifying, white-labeling a mature platform is a completely legitimate way to deliver, and often better than a half-built proprietary tool. Most good agencies do it.

What matters is whether they told you. If the site says proprietary and the link says otherwise, the issue is not the technology choice. It is that you found out by clicking instead of by being told.

5. Read their title tags

Right-click, view source, find the `<title>`. Then check a few pages, not just the homepage.

You are looking for duplicated brand names ("Case Studies | Agency Name | Agency Name") which happens when a template appends the site name to a title that already contains it. Also look for identical titles across different pages, and titles over about 60 characters that get truncated in results.

This is not a serious ranking problem. It is a competence signal, and it is one of the first things any SEO audit tool reports. An SEO agency with a doubled title tag on its own site has either never run their own audit or ran it and did not act. Ask which.

6. Check the sitemap against reality

Load `/sitemap.xml`. Spot-check a handful of URLs.

You are looking for URLs that 404, URLs that redirect rather than resolving directly, and (on agencies selling local SEO) programmatic city pages. If they have generated a page for every city crossed with every service, check two of them side by side. If the only difference is the city name swapped into otherwise identical copy, that is the exact thin-content pattern their own audit deck will criticize in your competitors.

7. Look for a physical address, then check whether it matches

Find the address in the footer. Then find their Google Business Profile and compare, character for character.

NAP consistency (name, address, phone) is a local ranking factor and it is the single most common finding in any local SEO audit. If the footer says one thing, the GBP says another, and the area code belongs to a third state, they have the defect they sell the fix for.

A remote agency with no physical address is fine, incidentally, as long as it is handled consistently: no LocalBusiness schema implying a storefront, and no city-specific claims they cannot support.

8. Run their own audit tool on them

Many agencies offer a free audit. Some publish the criteria. Take their criteria and apply it to their site.

If they will not share the criteria, run PageSpeed Insights on their homepage on mobile and see where they land. Mobile is the honest test; desktop scores flatter everyone. A marketing agency scoring in the seventies on its own homepage is telling you what their build quality looks like when nobody is checking.

What a good result actually looks like

None of this requires an agency to be perfect. Sites accumulate problems, and a defect you can point to is not disqualifying on its own.

What separates the good ones is the answer when you raise it. An agency that says "you are right, that shipped as placeholder copy and it is coming down today" is one you can work with. An agency that explains why a fabricated testimonial is industry standard is showing you how they will handle it when something goes wrong on your account.

Ask the question. The answer is the audit.

The uncomfortable corollary

If you run this checklist, run it on everyone, including whoever sent you the checklist.

An argument that an agency's own website is the best available evidence of what they will build for you is a genuinely good argument. It is also one that cuts toward whoever's site holds up better under it. Anyone making that argument should expect to be measured by it first, and should have already done the work.