TL;DR: Most SEO agencies overpromise and underdeliver. I've hired them, I've competed against them, and I've cleaned up after them. Here's the checklist I wish I'd had — 10 red flags, 10 green flags, the exact questions to ask, and when to skip the agency altogether.
I talk to business owners every week who've been burned by SEO agencies. The stories are depressingly similar: they paid $2,000-5,000/month for 6-12 months, got some traffic reports that looked impressive, but saw zero impact on leads or revenue.
The SEO agency market is massive and largely unregulated. Anyone can call themselves an SEO expert. And because results take months, a bad agency can string you along for a very long time before you realize nothing is working.
Let me give you a few examples from real conversations — names changed, obviously. A SaaS founder in Munich hired an agency that spent four months "building backlinks." When he finally asked to see the links, they pointed to a network of sites that all had the same WordPress theme, the same hosting provider, and suspiciously similar About pages. It was a PBN. He was paying 3,000 euros a month for links that would eventually get him penalized. He didn't know what a PBN was. Why would he? He's a founder, not an SEO.
Another one: an ecommerce brand paid an agency $4,500/month for "full-service SEO." The deliverables each month? A 10-page PDF report showing impressions going up and a list of "optimized pages." When I looked at the "optimizations," they'd changed three meta descriptions per month. That's it. Three meta descriptions. For $4,500.
That's why I put this together — a structured way to evaluate agencies before you sign anything. I've organized it into red flags (run away), green flags (good sign), and the specific questions that separate real practitioners from smooth talkers.
| # | Red Flag | Why It's a Problem |
|---|---|---|
| 1 | Guaranteed #1 rankings | No one can guarantee this. Google's algorithm considers 200+ factors. Anyone who promises specific rankings is lying or using black-hat techniques that'll get you penalized. I once sat in a pitch where the agency owner literally said "we guarantee page 1 in 90 days or your money back." They did get the client to page 1 — for a keyword nobody searched for. Technically not a lie. Practically useless. |
| 2 | "Secret formula" or proprietary method | SEO isn't secret. The best practices are public knowledge. Agencies that hide behind mystery are usually hiding the fact they don't do much. |
| 3 | Results in 30 days | Real SEO takes 3-6 months minimum. If they promise fast results, they're either doing something risky (PBNs, link schemes) or they're measuring vanity metrics. |
| 4 | Reports focused on impressions, not revenue | Impressions are easy to inflate. If the reports show impressions growth but skip conversion data, leads, and actual revenue impact — that's a red flag. One agency I reviewed was reporting "3 million impressions per month" for a client. Sounded amazing until I looked at the keywords: they were ranking for "what is SEO" and "how does Google work." Informational queries with zero commercial intent for a B2B SaaS selling to enterprise. Those impressions generated exactly zero leads. |
| 5 | One-size-fits-all packages | Your SaaS company and the local bakery down the street have completely different SEO needs. If the agency sells the same package to everyone, they're not doing real strategy. |
| 6 | 12-month contract lock-in | Good agencies don't need to trap you. A 3-month minimum is reasonable (SEO takes time), but 12 months with no exit clause is a warning sign. |
| 7 | They don't ask about your business | If the sales call is all about their services and they never ask about your revenue model, target customers, or competitive landscape — they're selling a template, not a strategy. The best agency pitch I ever received spent the first 40 minutes asking us questions and the last 20 minutes showing us anything. The worst spent 55 minutes on their slide deck and 5 minutes asking "any questions?" |
| 8 | Can't show their own rankings | An SEO agency that doesn't rank for SEO-related terms is like a personal trainer who's out of shape. Their own website is their resume. |
| 9 | No transparency on link sources | If they build backlinks but won't tell you where they come from, they're probably using link farms, PBNs, or other schemes that'll catch up with you. See the Munich PBN story above. Ask for a sample of 10 links they've built for other clients. If they refuse, that tells you everything. |
| 10 | Communication goes dark after signing | Responsive during sales, invisible after payment. If you can't get a meeting or a status update, they've moved on to the next sale. |
| # | Green Flag | What It Tells You |
|---|---|---|
| 1 | Case studies with revenue impact | Not just "traffic up 200%" but "leads increased from 50 to 180/month" or "organic revenue grew by $120k." They understand the business, not just the metrics. |
| 2 | Clear discovery phase before pricing | They audit your site, understand your market, and then quote a price. If they give a number before looking at your site, it's not customized. |
| 3 | Honest about timelines | "You'll see initial movement in 3-4 months, meaningful results in 6-9 months." That's honesty. SEO is a compounding game, not a sprint. |
| 4 | Transparent reporting with dashboard access | Real-time access to rankings, traffic, and deliverables. Not a PDF that arrives two weeks late. Bonus points if they use live dashboards you can check anytime. |
| 5 | Long-term client retention | Ask how long their average client stays. 2+ years is a strong signal. If clients leave after 6 months, something is off. |
| 6 | They explain what they do (and don't do) | Clear deliverable outlines per month. They're upfront about what's included and where costs might increase. No hidden fees or surprise charges. |
| 7 | They push back on bad ideas | If you ask for something that won't work (like targeting an impossibly competitive keyword), they tell you why and suggest an alternative. Yes-men don't get results. |
| 8 | Industry expertise (or willingness to learn) | Ideally they've worked in your vertical before. If not, they should be asking deep questions about your business, customers, and competitive landscape. |
| 9 | They talk about your sales process | The best agencies connect SEO work to closed revenue — which keywords correlated with actual deals, which pages led to demos or sign-ups. |
| 10 | They use (and recommend) tools, not just services | Good agencies supplement their work with the right tools and teach you to use them. They build capability, not dependency. |
Here's my checklist of questions that separate real agencies from pretenders. Print this out, bring it to the meeting. Watch their faces when you pull it out — that reaction alone tells you something.
SEO pricing is all over the place. Here's what the market actually looks like in 2026:
| Service Level | Monthly Cost | What You Get | Best For |
|---|---|---|---|
| Freelancer | $500-1,500/mo | Technical audit, keyword research, on-page optimization, basic reporting | Small businesses, local SEO |
| Mid-tier agency | $2,500-5,000/mo | Full strategy, content creation, link building, technical SEO, monthly reporting | Growing businesses, moderate competition |
| Premium agency | $5,000-10,000/mo | Dedicated team, content strategy, digital PR, advanced technical, competitive analysis | Established companies, competitive markets |
| Enterprise | $10,000-20,000+/mo | Full department outsourcing, multi-market, international SEO, executive reporting | Enterprise, multi-location, international |
Reality check
The national average for agency retainers in 2026 sits around $3,500-4,500/month. If someone offers "full-service SEO" for $500/month, you're getting a template, not a strategy. You get what you pay for. For more context, read my breakdown on how much SEO actually costs.
Not everyone needs an agency. Here's how to decide:
| Factor | DIY | SEO Tool (like SEOJuice) | Agency |
|---|---|---|---|
| Monthly cost | $0 (your time) | $35-200/mo | $2,500-10,000+/mo |
| Time investment | 10-20 hrs/week | 2-5 hrs/week | 1-2 hrs/week (meetings) |
| Expertise needed | High | Low-medium | Low |
| Customization | Full control | Tool-guided priorities | Fully managed |
| Best for | SEO-savvy founders, small budgets | In-house marketers, freelancers, small agencies | Businesses without SEO expertise, large budgets |
| Scalability | Limited by your time | Scales with pages, automated workflows | Scales with budget |
My honest take: most businesses under $5M in revenue don't need an agency. They need a good SEO tool, a basic understanding of SEO fundamentals, and 5 hours a week of focused effort. The right tool will tell you exactly what to work on and automate the repetitive stuff. Yes, I'm biased — I built one of those tools. But I also ran an agency before that, and I know what most of them deliver at the $3K/month price point. A tool plus your own effort will get you further.
Agencies make sense when you need deep expertise (international SEO, complex technical migrations), heavy content production (10+ articles/month), or you simply don't have anyone in-house who can do the work.
Give them 6 months for meaningful results. But expect clear progress indicators by month 3: improved crawl health, more indexed pages, some keyword movement. If nothing has changed by month 3, demand an explanation and a revised plan.
Specialist, almost always. An agency that does SEO, PPC, social media, web design, and branding is spreading itself thin. The best results come from agencies that focus on SEO or even a sub-discipline like technical SEO or content-led SEO.
Below $1,000/month for agency work? Almost never. At that price point, you're getting templated work, outsourced link building, and minimal strategy. You'd get better results using a good SEO tool and doing it yourself.
AI is a tool, not a strategy. Agencies using AI to scale content production or automate analysis can be efficient. But if "AI-powered" means they're using ChatGPT to generate thin content at scale, that's a red flag. Ask specifically how they use AI and what human oversight exists.
Absolutely. In fact, I recommend it. Your own tools give you independent visibility into what's actually happening. You're not relying solely on the agency's reporting. It keeps them honest and helps you learn.
Before switching, have an honest conversation. Share specific concerns with data. If they can't explain the strategy, show progress on agreed KPIs, or adjust their approach, it's time to move on. Document everything and make sure you own all assets before leaving.
tbh the checklist's emphasis on finding the "right SEO partner" and transparency hit home — always ask for a sample keyword strategy and two client refs in your niche before signing. ngl I did a 30-day paid trial once and added a contract clause for "no black-hat" + rollback; saved us from a messy penalty later. anyone else require dashboard access or monthly KPI exports?
100% — demand dashboard access. I always ask for GA4 + Search Console read-only, a Looker Studio dashboard + monthly raw CSVs for rankings, traffic, backlinks and CTR. Add SLA: exports within 5 days, rollback/penalty indemnity — if they refuse, walk. Watch for overnight spikes from spammy links (red flag). #SEO — what KPIs do you lock in?
Useful checklist but it feels high‑level — calling out “transparency, strategy, expertise and ethics” and warning about agencies that “guarantee instant results or low prices” is necessary but not sufficient. Insist on a baseline audit (crawl + log analysis, Lighthouse CI, Search Console + GA4 exports), a sample playbook with concrete tactics, explicit KPIs/SLA and a contractual rollback/clawback for risky link-building. Also ask how they validate causality — do they run randomized A/B tests or time‑series models, and how do their changes scale across multi‑language, high‑crawl‑budget sites?
100% — love this level of specificity. A few concrete additions I’d expect to see if I were vetting an agency (tbh most dodge this stuff):
- Baseline audit (concrete): run a full crawl with Screaming Frog/DeepCrawl + server log analysis (BigQuery + logs or Loggly) to map crawl patterns, Lighthouse CI for core web vitals baseline, export Search Console + GA4 to BigQuery for historical signal. Deliverable: CSVs + a prioritized findings spreadsheet (priority, effort, risk, owner, ETA).
- Sample playbook (what it should actually contain): exact tactics per issue type (e.g., fix duplicate titles → rename pattern, update templates, test on N sample pages), expected timeline per tactic (dev → staging → prod → observation window), roll‑out strategy (canary by URL group, language, or subdomain). Ask for a real example playbook used on a past client (redact PII).
- KPIs / SLA (practical examples): tracked weekly — organic clicks, impressions, average position for target keyword set, pages crawled/day, indexation rate, Core Web Vitals percentiles, revenue per organic session. SLA: incident response within 48h for critical drops, mitigation plan within 5 business days, monthly reporting. Tie payment milestones to KPI windows (e.g., baseline → 3 months → 6 months).
- Rollback / clawback: define thresholds (e.g., >10% drop in organic traffic for >14 days across >X pages) that trigger immediate rollback and partial refund proportional to sustained loss. Include who owns dev time/costs to revert.
- Validating causality (real talk): A/B tests are ideal but tough at site-wide SEO scale. Practical options:
- URL split tests: change X% of similar pages and hold the rest as control (we did this on ~300 product pages and waited 6–10 weeks for signal).
- Staggered rollouts / canary releases by region/language and compare with controls.
- Time‑series / causal models: CausalImpact, Prophet, synthetic control — good for attribution if you’ve got a decent pre‑period and control series.
- Beware confounders (seasonality, external campaigns). Always pair statistical tests with qualitative checks (indexing, render logs).
- Scaling across multi‑lang, high crawl‑budget sites: use consistent hreflang + language mapping, avoid duplicate content via correct canonicals, test on a low-risk subdirectory first, prioritize sitemaps + paginated XMLs, use crawl-delay or server-side throttling during heavy tests, and ensure CDN + server config are stable to not skew CWV.
Personal anecdote: I once insisted on a URL-split playbook with a 10% holdout. Agency pushed generic promises until we forced a staged rollout; the split test gave a clean +12% uplift in 8 weeks. Saved us from a bad full-site push.
If you’re negotiating with an agency, ask them to provide a redacted sample audit + one real playbook and a proposed causal test plan for your site. What kind of site are you thinking about (size, vertical)? That changes the testing strategy.
no credit card required
