How to choose an SEO company starts with knowing what separates real expertise from a polished sales pitch.

Most bad SEO hires do not look bad at first. The agency runs onboarding in month one, delivers a detailed audit in month two, and builds a content calendar in month three. Nine months later, you have a folder full of reports, dozens of published articles, and organic traffic that barely moved.

That happens because choosing an SEO company forces you to evaluate skills you hired someone else to provide. Agencies know more about SEO than most buyers, and strong sales teams know how to make almost any strategy sound convincing.

You do not need to become an SEO expert to make the right choice. Instead, evaluate the five factors that decide whether an SEO engagement succeeds or wastes your budget: the problem the agency diagnoses, the people doing the work, your ability to implement their recommendations, the results they measure, and the contract you sign.

You can spot problems in all five areas before you hire anyone.

This guide shows you the five most common ways SEO engagements fail, the questions that expose weak agencies, what SEO should actually cost, and the contract terms that protect your business if the relationship does not work out.

The five ways an SEO engagement dies

1. Wrong diagnosis. They fixed something that was never your constraint.
2. Wrong hands. A senior strategist sold it. A junior carrying 22 accounts delivered it.
3. Wrong side of the bottleneck. They shipped recommendations your team never implemented.
4. Wrong scoreboard. Traffic rose, revenue didn’t, and nobody caught it for six months.
5. Wrong exit. By the time you wanted out, leaving cost more than staying.

Failure three is the one no agency will raise on a sales call, and it quietly accounts for a large share of dead retainers. Hold that thought.

Failure 1: They solved a problem you didn’t have

“SEO” bundles at least five separate disciplines, and almost no agency is excellent at more than two:

• Technical: crawling, indexation, rendering, site speed, migrations, faceted navigation
• Content and architecture: which pages should exist, what they target, how they link
• Authority and digital PR: earning citations from sites with their own audiences
• Local: Google Business Profile, service-area pages, review velocity, citations
• Analytics and conversion: turning sessions into revenue you can defend in a budget meeting

Mismatch your bottleneck to their strength and you will buy competent work that changes nothing.

Run a rough diagnosis yourself first

You don’t need to be right. You need to be specific enough to notice when an agency’s diagnosis contradicts yours. Three checks, twenty minutes:

1. Search site:yourdomain.com. Far fewer results than you have pages? That points at indexation, not content.
2. Open Google Search Console, filter out queries containing your brand name, and sort by impressions. High impressions with a weak click-through rate points at relevance and titles, not authority.
3. Find topics where competitors rank and you have no page at all. That’s a coverage gap, and no volume of link building closes it.

Write down your guess. Seal it.

The test: three fixes and the reasoning

Send every finalist the same one-page brief, then ask for the same thing: spend 30 to 60 minutes with the site and tell us the first three things you would change, and why those three.

A strong answer cites your actual URLs, ranks fixes by expected impact against effort, ties them to revenue rather than rankings, names what they need from your team, and states plainly what they cannot know without data access.

A weak answer is a crawler export. Site health score of 72. Three hundred and forty “errors.” Add alt text. Nothing in it could only have been written about your business.

Then ask the follow-up that does the heavy lifting: “What would you not do for us?” Agencies with a real point of view decline things. Agencies that say yes to everything in the sales process will bill you for everything after it.

Failure 2: The person who sold you disappeared

The strategist running your discovery call is often not the person touching your account. This is the single most common disappointment in the category, and it is trivially easy to check before you sign.

Ask directly, and expect names rather than “a dedicated team”:

• Who writes the content, and can I read three pieces they published in a field like mine?
• Who does technical implementation, and do they work with developers or just file tickets?
• Who does outreach, and what does a placement actually look like?
• Is any of this subcontracted or white-labeled, and to whom?
• How many other accounts does my strategist carry?

Roughly 8 to 10 clients per strategist is workable. Twenty should trigger hard questions about capacity. In a Databox survey of 48 agencies, nearly 70% said each account manager handled fewer than 10 clients, while only a little over 10% assigned 15 or more clients to one manager. Among respondents, 6 to 10 clients was one of the most commonly recommended ranges, although workload varied substantially with account complexity and team support.

White-label delivery is not automatically disqualifying. Hiding it is. And if they won’t introduce you to the practitioner before signature, treat that as a soft no.

Failure 3: Nobody on your side could implement the work

Here is the failure mode the pitch deck never covers. An audit fixes nothing. Content briefs fix nothing. A prioritized roadmap sitting in a shared drive because your developer is heads-down on the checkout rebuild until Q3 fixes nothing.

Before the first call, audit your own capacity honestly:

• Developer time. How many hours a month can you actually get, and who approves the ticket?
• Subject-matter access. Will your best salesperson sit for a 30-minute interview, monthly?
• Approval speed. How long does a draft sit before someone publishes it? Two days or five weeks?
• Platform constraints. Can you edit templates, title tags, and URL structure, or does your platform generate product pages you cannot modify?
• Legal or compliance review. Add weeks to every content cycle if the answer is yes.

Now use that as a filter. An agency that only advises is the wrong fit for a company with no execution muscle, no matter how sharp the advice. An agency that wants to write 12 articles a month is the wrong fit if your compliance team can clear three.

The right question to ask every finalist: “What do you need from us every month, in hours?” Anyone who answers “nothing” is planning to publish content nobody briefed and nobody reviewed. Vagueness here becomes blame in month six.

Failure 4: You measured the wrong thing for six months

Record your baseline before the first sales call

Once an agency starts reporting, their numbers become the only numbers. Get yours first. Spend an hour capturing:

• Search Console: clicks, impressions, and average position for the last 16 months, exported
• The branded versus non-branded split (filter out queries containing your company name; what’s left is the number that matters)
• GA4 organic sessions and organic conversions, month by month, for 12 months
• Your top 20 landing pages by organic entrance
• Your economics: lead value, close rate, average order value or lifetime value
• The 10 queries you want to win, written down before anyone suggests easier ones

Screenshot it. Date it. File it.

Demand a scoreboard, not a dashboard

Impressions, keyword counts, and “keywords improved” (which can include a move from position 87 to 62) are the favorite metrics of agencies with nothing to show. Ask for a redacted sample of a real monthly report from a live account, not a template.

Good reporting leads with non-branded organic clicks, visibility for the query clusters that map to money, conversions or pipeline, and a short section on next month’s priorities and why. It reads like a decision document.

Settle attribution in advance: which leading indicators you will accept in months one through four (crawl health, indexation, non-brand impressions, clicks to commercial pages) and which lagging indicators you will judge by in month nine (revenue, qualified leads, assisted conversions). In a poll of 3,680 SEO practitioners, Ahrefs found that three to six months was the most common window for SEO to begin showing results. Google is more cautious: it says some search-related changes can take several months to be reflected. Treat month nine as a reasonable checkpoint for business impact, not a guaranteed deadline.

Ask about AI search, then check whether they’ve measured it

AI Overviews and assistant-based answers have changed how clicks get distributed on informational queries. Pew Research Center found that AI Overviews appeared on 18% of Google searches in its March 2025 browsing study; when an AI Overview appeared, users clicked a traditional search result on just 8% of visits, versus 15% when one did not. A separate Ahrefs analysis of 300,000 keywords using December 2025 Search Console data found that an AI Overview was associated with a 58% lower click-through rate for the number-one organic result.

You are not listening for buzzwords. You are listening for a specific, testable claim: how it changes their recommendation for your site, and how they track whether your pages get cited inside AI answers rather than only counting blue links. An agency quoting conference talking points sounds identical to one that has run the numbers in client accounts, right up until you ask which accounts.

Be skeptical of anyone selling “AI optimization” as a separate line item with its own retainer.

Failure 5: Leaving cost more than staying

Ask what you keep when the relationship ends. The answer should be everything. Confirm before signing:

• Analytics, Search Console, and Google Business Profile live in your accounts, with the agency added as a user. Never the reverse. Keep verified owner status on Search Console yourself.
• Content lives in your CMS, along with the keyword research and briefs you paid for.
• Links point to your domain, not to a leased subdomain, microsite, or agency-controlled property.
• Any dashboard is a convenience, not the only place your data exists.
• IP transfers on payment, in writing.

Agencies that host your site on a proprietary platform, or build links from a network they own, aren’t building you an asset. They’re building leverage.

Then negotiate the exit. Most agencies will move on at least three of these:

What SEO actually costs, and what the money buys

Rates vary by market and specialization, but these bands hold up reasonably well as buying ranges. Clutch’s 2026 pricing data, drawn from more than 65,000 SEO companies and client project reviews, puts typical agency retainers at $2,000 to $20,000 per month and the most common agency hourly rate at $100 to $149. A separate Ahrefs survey of 439 SEO providers found agencies averaged $3,209 per month, versus $1,349 for freelancers; local SEO averaged $1,557 per month, while providers serving worldwide markets averaged $3,474.

Two pieces of arithmetic before you judge any quote.

The hours math. Divide the retainer by a blended hourly rate:

$2,500 ÷ $125/hour = 20 hours per month

Twenty hours is about one day a week covering strategy, technical work, content, links, and reporting for your entire account. Once you see a quote in hours, cheap retainers stop looking like bargains. Underfunded SEO doesn’t move slowly. It doesn’t move.

The breakeven math. If a customer is worth $3,000 in gross profit and the retainer is $4,000 a month, you need roughly 1.4 extra customers a month to break even, and that should hold within 9 to 12 months. Run this before the first call and you will negotiate from a much stronger position.

Two billing questions most buyers forget: are content and link costs inside the retainer or passed through at cost plus margin, and are writers in-house, freelance, or offshore? Those answers explain most of the gap between two proposals that look identical on paper.

Skip performance-based pricing. It pushes the agency toward whatever moves the agreed metric fastest, which is rarely what builds a durable asset, and it ends in arguments about attribution.

Agency, freelancer, or in-house?

Freelancer or consultant ($75 to $200/hr, or a small monthly retainer). Best when you have writers and developers and what you lack is direction. Senior thinking, no account management layer. Risk: one person, capped hours, no cover when they’re sick.

Boutique agency (roughly 3 to 20 people). You’ll talk to someone senior, and small shops often specialize: local service businesses, ecommerce, B2B SaaS. Specialization beats size. An agency that has run 30 ecommerce migrations knows things a generalist doesn’t. Risk: a thin bench in whichever discipline you need most.

Full-service agency. Worth it for multi-location, multi-language, or enterprise complexity, or when SEO has to coordinate with paid, PR, and creative. Risk: juniors behind a senior sales team.

In-house hire plus a consultant on retainer. Cheapest over three years, slowest to start. Best when organic search will be a permanent channel rather than a campaign, because an internal owner knows your margins and which customers are actually profitable.

Many companies land on a hybrid: one internal owner who sets priorities, with outside specialists supplying capacity.

Red flags that should end the call

• Guaranteed rankings, guaranteed position one, or guaranteed traffic volumes. Nobody controls Google’s index. Google says flatly: “No one can guarantee a #1 ranking on Google,” and warns businesses to be wary of SEOs that guarantee rankings or claim a special relationship with Google. Google also notes that it does not guarantee that a page will be crawled, indexed, or served in search results.
• Any claim of a special relationship or contact at Google.
• A proprietary method they won’t describe in plain language.
• A cold-outreach “audit” with your name merged into a template and no mention of a specific page or competitor.
• Links sold by the unit, in packages, with no publisher list you can review.
• Deliverable quotas set before anyone diagnosed anything.
• Accounts created under the agency’s ownership rather than yours.
• Reports that lead with branded traffic or raw impressions.
• No questions about your margins, your sales cycle, or who approves content.
• Urgency tactics. Discounts expiring Friday belong to timeshare sales, not a 12-month strategic engagement.

One of these is a conversation. Three is a decision.

The 10 questions to send in writing

Comparing written answers side by side is far more revealing than comparing charisma on Zoom. Send the same list to every finalist.

1. Based on what you’ve seen, what’s the single biggest constraint on our organic growth, and what evidence supports that?
2. What would you fix in the first 30 days, and why that first?
3. Who does the day-to-day work, how many accounts do they carry, and can I meet them before signing?
4. Walk me through your last five link placements. Real URLs.
5. What do you need from my team every month, in hours and approvals?
6. Which KPI are you willing to be judged by at month nine?
7. Show me a client whose traffic grew but revenue didn’t. What happened?
8. How do you track whether our pages get cited in AI Overviews and assistants, and what would you change because of it?
9. Who owns the content, links, accounts, and research if we part ways?
10. May I speak with a client who left you?

Question 10 is rare and extremely revealing. When you get a reference on the phone, ask one thing: what happened in month four, when the honeymoon ended?

Score them, don’t feel them out

Rate each finalist 1 to 5, multiply by the weight, total the columns.

Notice what isn’t on the list. Add price only to break a tie between two agencies within a few points of each other. Choosing an SEO company on price is how most businesses end up choosing twice.

What “working” looks like at 30, 90, and 180 days

Days 1 to 30. Baseline confirmed against your own export. Technical issues found and prioritized. Tracking verified. A written roadmap you can read without a translator. No traffic change yet, and anyone promising one is guessing.

Days 31 to 90. Fixes shipped to production, not just recommended. Crawl errors dropping, indexation improving, first content live, early movement on low-competition terms. Communication should be proactive rather than reactive.

Months 4 to 6. Non-branded impressions and clicks trending up. Target queries moving from page two toward page one. First attributable conversions on the pages they built. Three to six months is a reasonable window for meaningful movement on an established site, although new domains, competitive markets, and migrations can take materially longer. In a poll of 3,680 SEO practitioners, Ahrefs found that three to six months was the most common timeframe for SEO to begin showing results.

Months 7 to 12. Compounding traffic, revenue attribution, and a defensible argument for expanding or reallocating budget.

Legitimate things do interfere: seasonality, core updates, a migration, a competitor’s spending spree. The tell is timing. Good agencies flag those risks before they land and revise the forecast. Weak ones discover them in month 11, on the slide explaining the shortfall.

If month three arrives with no roadmap and no baseline, the problem is not that SEO takes time. The problem is the agency.

The takeaway

You will never out-expert an SEO company on SEO, and you don’t have to. Choosing an SEO company well means vetting for the five ways this fails: a wrong diagnosis, invisible delivery staff, work your team can’t implement, a scoreboard that hides bad results, and a contract that punishes you for leaving.

Do three things this week. Export your last 16 months of Search Console data and file it. Write a one-page brief with your real numbers in it. Send that brief plus the 10 questions above to three companies and ask each for the first three things they would fix.

Their written answers will tell you more in twenty minutes than any pitch deck will in an hour. Then buy a paid audit from whoever scores highest, and don’t sign anything longer than 90 days on a first engagement.

Related Articles:

• Best Local SEO Company: How to Vet One in 60 Minutes
• What to Look for When Hiring an SEO Company (9 Tests)
• Is Hiring a SEO Company Worth It? 5 Ways It Fails
• How to Tell If Your SEO Company Is Working (7 Checks)
• Why Hire an SEO Company (And When You Shouldn’t)
• How to Fire an SEO Company Without Losing Your Site

Last Updated on 1 month ago by Alipio Umiten IV