How to Evaluate an SEO Marketing Agency Before Signing a Contract

Why Most SEO Agency Hires Fail

Buyers usually pick an SEO agency the way they'd pick a wedding photographer. They flip through the portfolio, maybe take a reference call, then go with whoever seemed sharp and personable on the sales call. Six months later they're paying $4,000 a month for a monthly report with rankings on keywords that don't drive revenue.

The failure pattern is almost always the same. The agency was competent at what they showed you on the sales call. That work has close to zero overlap with what your specific business actually needs. Nobody checked for the overlap during vetting because vetting was set up to answer the wrong question. The question the buyer asked was "is this agency good?" The question that would have caught the problem was "does this agency know how to run the specific play my business needs to win?"

The rest of this guide is organized around answering that second question. Sections move from the least revealing signals (credentials, awards) to the most revealing ones (technical fit interview, reporting sample, contract terms). Skip ahead if you already know a section applies to your situation. If you're just starting the search, the Learn hub has adjacent pieces on SEO services scoping that pair well with this one.

The Credentials Check: What Actually Matters

Certifications and awards are the easiest thing to evaluate and, on their own, the least predictive of whether an agency will do good work for you. That doesn't make them worthless. It makes them a floor, not a ceiling.

Useful signals in the credentials layer:

• Years operating. Agencies under three years old are still figuring out delivery. Agencies over ten years old have survived at least one algorithm cycle. Neither disqualifies anyone. Both should inform how much verification you do at other layers.
• Team size and where the work gets done. Ask who owns your account day-to-day, whether content is written in-house or offshored, and whether the technical work is billable hours from a specialist or included in the retainer. Answers vary, but vague answers are a signal.
• Google Partner and Meta Business Partner status. Table stakes for paid media agencies. For pure SEO shops, useful only if you'd expect them to run paid alongside organic.
• Certifications from industry vendors. Semrush, Ahrefs, HubSpot, and Moz all offer them. They confirm the team knows the tools. They don't confirm the team can do the strategy.
• Industry recognition. Clutch top-rated status, Search Engine Land Awards, US Search Awards. Useful as a filter for agencies that operate at a professional level. Not a substitute for the fit interview later in this guide.

What credentials should push you to verify at the next layer: how the agency handled a Google core update that hurt a client, how they respond when a technical deploy breaks a canonical tag, how they built the AI search visibility practice that most agencies now claim to offer.

How to Read a Case Study (What the Deck Isn't Showing You)

Case studies are marketing artifacts. They are edited to close deals. Reading them well means treating them the way an equity analyst reads an earnings deck: look at what's on the page, then ask why anything obvious is missing.

Here are the four line items that separate a case study from a rankings screenshot.

Traffic without revenue is a common tell. It usually means the agency ranked the client for high-volume keywords that don't convert, and the client either didn't notice or didn't want the case study rewritten. When an agency won't share revenue data even under NDA, the honest read is that revenue didn't move the way traffic did.

A subtler tell: no algorithm updates mentioned in a case study spanning a year or more. Google runs multiple core updates per year. An agency that engaged for twelve months and didn't have to navigate at least one update is either working on a site the algorithm doesn't care about, or is editing the history.

What to ask when reviewing case studies:

1. Which of your clients most resembles our business in vertical, size, and business model? Can I speak with them?
2. Show me a case study where the initial strategy didn't work. What did you change and how long did that take?
3. What percentage of your case study wins are still your clients today? (Churn tells you more than the wins.)
4. Can you show me the same client's data in Google Search Console during the engagement, not just the agency dashboard?

The Technical Fit Interview

This is the layer buyers most often skip and the layer that most reliably predicts fit. The technical fit interview is a 30 to 60 minute conversation where you put a live problem from your business in front of the agency and watch how they think.

You are not asking them to solve it on the call. You are watching for how they frame the problem and what they ask about before they answer. The goal is to see whether their instincts match the shape of your business.

Set up the interview

Pick a real, current issue from your site. Examples:

• A category page that used to rank on page one and now sits on page three
• A cluster of blog posts cannibalizing each other for the same query
• An indexation problem where Search Console shows 40% "crawled, not indexed"
• A drop in impressions on a specific topic cluster after a recent Google update
• A situation where AI Overviews are pulling answers but organic traffic is falling

Send the agency the problem 48 hours before the call. Give them access to Search Console read-only for the affected properties. Ask them to come prepared with initial hypotheses, not a solution.

What to watch for on the call

An agency worth hiring will spend the first ten minutes asking questions. What is the business model on that page? What's the current traffic and revenue baseline? What competitive movements have you noticed? What was the last technical change to the template? What does the internal linking structure look like from adjacent pages?

An agency that opens with recommendations before asking questions is either running a template playbook or performing certainty for the sales call. Both are red flags for the actual engagement.

The strong signal isn't the answer. It's the shape of the diagnostic path. Do they consider technical, content, and off-page factors? Do they check assumptions before making them? Do they name the specific data they'd want to see next? Do they have an opinion about how AI search behavior intersects with the problem, or do they treat it as separate from traditional SEO?

Reporting and Attribution

Ask for a sample report from a current client, redacted for confidentiality. What you get back tells you almost everything about how the agency thinks.

A poor report is a stack of screenshots: ranking positions on the left, an organic traffic chart in the middle, a work log at the bottom. It answers the question "did the agency do the tasks" and avoids the question "did the tasks work."

A good report is organized around metrics that map to business outcomes. It shows organic traffic segmented by intent and page type. It shows conversions or leads from organic with a defined attribution window. It shows position changes weighted by keyword commercial value, not raw ranking count. It names the algorithm activity for the reporting period. It ends with a decision, not a summary: what the agency plans to do next month, why, and what they'll stop doing.

The attribution question

How the agency handles attribution is the fastest way to separate strategists from technicians. The AI search environment has made this harder in specific ways: LLM citations don't show in Google Analytics, AI Overviews suppress click-through on top-of-funnel content, and users increasingly complete pre-purchase research inside ChatGPT or Perplexity before ever landing on your site.

Ask the agency:

1. What attribution model do you use for organic traffic that touches a lead? (First touch, last touch, position-based, custom.)
2. How do you account for branded search that comes from off-site awareness activity you're driving?
3. How do you measure impact when AI search platforms are citing our content but not sending traffic?
4. What tools connect our CRM to search data, and who owns keeping that connected?

An agency without answers here isn't disqualified. An agency that gives you rehearsed answers without specifics about your stack is running a script.

Contract Terms That Matter

Most SEO contracts look similar on the surface. The differences hide in five clauses that determine what you can do if the engagement isn't working.

The content and data ownership clauses matter more than most buyers realize. An agency that owns the analytics accounts, the Search Console access, and the drafts of every article they've written can effectively hold your SEO work hostage. Never sign a contract that puts your data on their side of the wall.

On pricing structure: pure retainer works, and so do hybrid retainer-plus-performance-bonus models. What doesn't work is pure performance pricing on organic search. The timelines are too long and the attribution is too messy. Worse, the incentives push toward tactics that rank fast and don't compound. If an agency offers you a pure pay-for-rankings deal, they're either desperate or planning to bill for services they haven't described.

Red Flags That Should End the Conversation

Some signals mean the engagement will go badly regardless of how good the sales rep is. Walk away when you see them.

1. Guaranteed page-one rankings. No agency can guarantee this. The ones that promise it are targeting long-tail keywords with no commercial value, or they're running the private blog network playbook that Google has been penalizing for a decade.
2. No named strategist on the account. If the pitch is "our team will work on your account," ask who specifically owns the strategy. If the answer is a rotating pool, you're buying task execution, not strategic thinking.
3. Reliance on rank tracking as the primary metric. Rank tracking is a diagnostic. It is not an outcome. Any agency that leads with rank tracking is optimizing for the wrong thing.
4. Refusal to share client references from your industry. Understandable if you're competitive with a current client. Suspicious if the answer is "we can't share references" without a specific reason.
5. Content produced offshore without editorial review. The cost math is obvious. The output is the problem. Google's Helpful Content system and the AI content detection layer have made this playbook actively negative for rankings.
6. Vague answers about link acquisition. Ask specifically how they build links. If the answer includes phrases like "we have relationships" without naming outlets or process, they're either buying links or doing PBN work. Both create liability. Genuine link acquisition is a repeatable process the agency can describe in detail.
7. No local proof for local work. If you need local SEO, the agency needs to show you businesses they've ranked in map packs in geographies with similar competition to yours. Generic SEO experience doesn't transfer to local SEO automatically.

15 Questions to Ask Before You Sign

Print this list. Take it into every discovery call. Score answers as concrete, plausible, or evasive.

About their work

1. Which of your current clients most resembles our business? Can I speak with them for 20 minutes?
2. Show me a client engagement that didn't hit the initial targets. What did you change?
3. What's your client retention rate at 12 months? At 24 months?
4. Who specifically will own strategy on our account? How much of their time do we get?

About the work they'd do for us

5. Based on our site and market, what would you focus on in the first 90 days? Why those things?
6. What technical audit tools do you use, and can you show me a sample audit output?
7. How do you approach AI search visibility compared to traditional SEO?
8. What percentage of the retainer goes to content, technical work, and link acquisition?
9. How do you handle it when we disagree on a strategic call?

About reporting and accountability

10. Can I see a sample report from a current client, redacted?
11. What metrics do you report on monthly? Which of those tie directly to revenue?
12. How do you measure impact when AI platforms cite our content but don't send clicks?
13. What's the cadence for strategic reviews versus tactical updates?

About the contract

14. What's the exit clause if we're not seeing results by month six?
15. Who owns the content, the data, and the tool access if we end the engagement?

Putting It All Together

A useful vetting process runs in five phases, roughly in this order:

The compressed version, if you only have time for two things: run the technical fit interview and read the exit clause carefully. An agency that shows up sharp on a real diagnostic conversation and gives you a clean out at month six has almost no way to trap you. Everything else is a subset of those two.

If you're currently mid-evaluation and want a second read on the proposals you're seeing, we do proposal reviews at no cost. Even when we don't end up being the right fit, buyers usually walk away with better negotiating leverage on the agency they do pick. That's the point of the exercise: your budget deserves the vetting either way.

Author

Michael Timi

Partner & Marketing Manager, eMac Media

Drives strategic partnerships and revenue growth through high-impact marketing initiatives, business development, and lead generation.

Editor

Princess Pitts

Director of Communications Strategy, eMac Media

Specializes in editorial strategy, content governance, and brand communications at scale.