SEO agency red flags, including two of our own.

Judge structure, not style.

Most red flag lists describe personality. Pushy salespeople, jargon, slick decks. Those are irritating and they are weak predictors, because plenty of good practitioners present badly and plenty of bad ones present beautifully.

The signals worth acting on are structural. They are properties of the offer rather than of the person, they are checkable before money changes hands, and each one has a specific mechanism behind it rather than a bad feeling.

Five below. Each includes the mechanism, because a warning you understand is one you can apply to situations this page did not anticipate.

What is the single clearest red flag when hiring an SEO agency?

Guarantees, and pricing that hides.

A guarantee of specific rankings is not usually a lie. It is a bet that you will not read the exclusions, which typically limit the promise to terms nobody searches or to a window that quietly expires. The useful move is not to argue but to ask for the failure clause in writing.

Pricing withheld until a call is a negotiation structure. The explanation, that every business is unique, is true and is not the reason. Hidden numbers prevent comparison, and comparison compresses margin.

Neither is disqualifying on its own. Both tell you what the firm optimises for, and that is worth knowing before the relationship rather than during it.

Link volume, and reporting with nothing to check.

Bulk link promises are the highest risk item on this list, because the consequence lands on your domain rather than the agency’s. Purchased, exchanged, and networked links are named directly in Google’s spam policies. The only correct question is where the links come from, and vagueness in the answer is the answer.

Reporting without a baseline is subtler and more common. If nobody documented where you started, every subsequent number is unfalsifiable, including the good ones. That ambiguity usually serves the party producing the report.

Both are checkable before you commit. Ask where links come from. Ask what your baseline is and who records it. Two questions, two minutes.

Contracts that hold what the work should.

A long term with a buried notice period moves the pressure from the work to the paperwork. When leaving is difficult, the incentive to keep earning your business every month weakens, and scope tends to drift downward while the invoice holds steady.

Long terms are defensible where a build justifies them, because a project with a defined end reasonably carries a defined commitment. They are harder to justify for ongoing work, which should have to earn its renewal continuously.

Read the cancellation clause before the scope. It is the shortest section and the most revealing.

What red flags apply to Uncharted SEO itself?

The failure modes that apply here.

A warning list that exempts its author is marketing. Two apply to this practice specifically and you should weigh them.

Capacity is finite. Founder led delivery removes the handoffs where projects usually fail, and it means there is no bench. A period of high demand can push a build back. Ask what the queue looks like before signing, and treat a vague answer from us the same way you would from anyone else.

Month to month cuts both ways. It protects you from a bad engagement and it also means work that compounds can be stopped before it compounds. A client who leaves at month two has paid for foundations and collected none of the return. We say this during the audit rather than after, because the alternative is a client who feels misled.

The signals that a firm is probably real.

Published pricing, because it invites the comparison most firms structure their sales process to avoid.

Live client links rather than screenshots, because they let a stranger check the work without asking permission.

A firm that qualifies you out. An agency willing to say a program is not worth it yet is protecting both parties from an engagement that would fail, and the firms that never turn work away are the ones whose clients leave at month four.

Reporting that includes flat months. It is the strongest available signal that the rest of the reporting is real, because a fabricated report has no reason to contain a bad month.

What to do when the engagement is underway and you are unsure.

Ask for your baseline and your current position against it. An agency doing the work can produce both quickly. One that never recorded a baseline will not be able to, which tells you what the reporting has been measured against.

Then ask what shipped in the last ninety days, in nouns. Pages, fixes, links with named sources. If the answer is activity rather than artifacts, the engagement may be producing reports rather than results.

Before you leave, secure your assets. Confirm ownership of hosting, registrar, analytics, and Search Console while the relationship is still cordial. That is a far easier conversation before a cancellation than after one.

What you should own regardless of who you hire.

Your hosting and your domain registrar, in accounts registered to you. This sounds obvious and is the single most common thing agencies quietly hold, which turns leaving into a negotiation rather than a decision.

Your analytics and Search Console properties, with you as an owner rather than a delegated user. If access was granted to you rather than created by you, it can be withdrawn.

Your content and your site, built on standard tools any competent professional can maintain. A proprietary platform makes leaving expensive by design, and that expense is a feature for whoever built it.

Ask about all three at the start. The answers are short, and an agency that hesitates on any of them has told you what the exit looks like.

What a bad engagement actually costs.

The obvious cost is the fee, and it is the smallest part. A year of a thin retainer at a modest monthly rate is recoverable. What is harder to recover is the year.

Search compounds, which cuts both ways. Twelve months of correct foundational work produces positions that keep earning. Twelve months of reporting theatre produces nothing to build on, and the competitor who started at the same time is now a year ahead.

The worst outcome is active damage. Purchased links, scaled thin content, or a migration that discarded earned equity all leave you behind where you started, and unwinding that takes longer than the original work would have.

This is why the checks matter more than the price. The difference between a good and a bad engagement is not the monthly fee, it is whether the year was spent or wasted.

What honest reporting looks like, so you can recognise it.

Everything reads against a baseline recorded before any work ships. That single document is what turns a monthly report from an assertion into something checkable, and its absence is the most common reason a client cannot tell whether they were served well.

AI search adds a wrinkle worth knowing about. Ask the same question twice and the cited sources can differ, so a single check proves very little. The method that survives that is a fixed panel: agreed questions, rerun on a schedule, logged with dates. Fixed questions cannot be quietly swapped for easier ones.

Single months are noise and direction across months is signal. Both get reported, including the flat ones. A report showing progress every single month is describing something other than reality, and recognising that is a useful skill whoever you hire.

What a real engagement produces early, at any agency.

Month one should produce two artifacts: a documented picture of what your buyers actually search, and a baseline recording where you stand today. If month one produces a strategy deck and nothing else, ask what happens to it.

Months two and three should produce shipped work you can open in a browser. Pages that exist, technical faults closed, profile and entity facts made consistent. Work that produces no artifact is difficult to distinguish from work that did not happen.

After that the program compounds, and the monthly read should tie back to the baseline rather than to a fresh set of numbers each time. Use this as a checklist against whoever you hire, including us.

What this practice will not do, whatever the incentive.

No purchased links, no exchanges, and no private network placements. All three are named directly in Google’s spam policies as link schemes, and the consequence attaches to your domain rather than to us.

Two companion pages cover the rest of this decision: How to choose an SEO agency and SEO agency vs freelancer. Our own numbers are on the pricing page, the roster is on the portfolio, and the proof page sets out the evidence standard.

No guaranteed rankings and no guaranteed timelines. Nobody controls a search engine. What can be committed to is the work, the measurement, and the honesty of the reporting, and those are the only commitments worth having.

No content published at volume for its own sake. Scaled thin content is named in the same spam policies, and it is also simply a waste of the budget it consumes.

No client kept by paperwork. Month to month, no notice period, no early termination fee. If the work stops earning its place you should be able to act on that the same month.

An agency writing the buyer’s guide is a conflict of interest.

There is an obvious problem with an agency publishing advice on choosing an agency. Every criterion can be selected to favour the author, and most pages of this kind are exactly that exercise.

The defence is to use criteria that are checkable independently and that we could fail. Whether a firm ranks for what it sells, whether it publishes a price, whether its clients are verifiable, and whether it records a baseline are all things you can confirm without asking us anything.

Read this page assuming it is self serving, then run the checks on us first. If we fail one, that is more useful to you than the page is.

Straight answers, in depth.