How to Evaluate an SEO Agency or Consultant

SEO is one of the few line items where you can pay for two years before you can prove whether it worked. That delay is exactly what bad vendors hide inside. So the test is not "are they ranking us." The test is whether they think like someone accountable for an outcome, or like someone billing for activity. Those two look identical on an invoice and nothing alike in a meeting.

The good news is that you already evaluate this kind of risk for a living. You know how to tell a consultant who understands your business from one who memorized your deck. SEO is no different. You do not have to audit the work itself. You have to audit the judgment behind it, and judgment is something you can read in a room. The rest of this page is just where to point your attention.

Green flags vs red flags

Most of the signal is in the first conversation, before any work happens. Here is what separates a partner from a vendor.

Two of those deserve a flat veto. Anyone who guarantees a #1 ranking is either lying or does not understand the product they sell, because Google ranks pages on signals no vendor controls. Google's own Search Essentials guidelines state plainly that no one can guarantee a top ranking and warn against companies that claim a "special relationship" with Google. And anyone buying links is selling you a future penalty: link schemes are an explicit violation of those same guidelines. If a vendor pitches either, the meeting is over. You learned what you needed.

The questions that expose competence fast

You do not need to evaluate the answers like an expert. You need to notice whether the answer is specific and grounded, or vague and flattering. A competent vendor gets more precise under questioning. A weak one gets more buzzwordy.

How to read their reporting honestly

Reporting is where good intentions go to get dressed up. The most common trick is reporting on terms you were always going to win. If your company is called Acme and the report leads with "we rank #1 for Acme," that is not SEO, that is a screenshot of your own brand. The same goes for raw impressions, which can climb while clicks, leads, and revenue stay flat, because an impression is just Google showing your link, not anyone choosing it.

Read every report by asking one question: does this number connect to money, or does it stop at activity? "Organic sessions to our pricing and product pages grew, and demo requests from organic rose" connects to money. "We published 40 articles and built 500 links" stops at activity, and worse, that volume can quietly be the thing dragging you down. Strategy looks like a small number of deliberate moves tied to a goal you recognize. Busywork looks like a large number of undifferentiated tasks tied to nothing. When in doubt, ask them to draw the line from a given activity to a business result. If they cannot draw it, it was busywork.

A few honest things a good report contains that a vanity report never does. It names a baseline and compares against it, so a number means something. It segments organic from paid and brand from non-brand, because lumping them together is how a vendor takes credit for traffic your marketing team paid for or your reputation earned. It shows losses as well as wins, because real SEO has both, and a report with no setbacks in it is a report that is hiding them. And it ends with a decision: here is what we learned, here is what we are doing next, here is what we are stopping. If the report is a wall of green arrows with no narrative and no next move, you are being managed, not served.

The AI-era note

A vendor using AI to produce more, faster is not automatically worse, and not automatically better. Speed is not the variable that matters. Judge the thinking and the verification, not the volume or the turnaround. A small team that uses AI to draft and then edits, fact-checks, and pressure-tests every page can be excellent. A team that points a model at a keyword list and publishes the output at scale is running the 2026 version of link spam, and it carries the same risk: Google's guidance treats content made primarily to game rankings as spam regardless of how it was produced. So the AI question to ask is not "do you use AI." It is "show me how you verify what the AI produced before it goes live." A confident, concrete answer is a green flag. A shrug is a red one.

Agency, consultant, or in-house

A quick orientation, not a verdict. An agency gives you breadth and bench depth, but you may be a small account getting junior attention. A consultant gives you senior thinking and direct access, but limited hands and a bus-factor of one. In-house gives you context and control, but is slow to hire and easy to silo. The right answer depends on your stage, budget, and how much SEO drives the business. We cover that tradeoff properly in the resourcing briefing; for now, just know that the flags above apply to all three. A bad consultant and a bad agency fail you the same way: activity with no line to revenue.