General interview questions tell you whether a company is organized and easy to work with. They do not tell you whether the people answering actually understand search. To test for depth, you need questions that are hard to answer with a rehearsed script. The goal is to listen for reasoning, not vocabulary. A shallow answer names tools and tactics. An expert answer connects a specific symptom to a testable cause and explains the tradeoffs involved.

Ask how they would prioritize a list of problems

A useful probe is this: “If an audit turned up ten technical issues, how would you decide which three to fix first?” This question works because there is no single correct answer, so a memorized response falls apart quickly. A shallow answer ranks issues by a generic severity label or simply works down the list. An expert answer ties prioritization to business impact. They will ask which pages drive revenue or leads, weigh how many URLs each issue affects, consider how much effort a fix requires, and explain that a small problem on an important template can matter more than a large problem on pages no one visits. The reasoning, not the ranking, is what reveals expertise.

Ask them to diagnose a symptom

Give them a realistic scenario and ask for their thinking out loud. For example: “Impressions for a section of the site dropped, but rankings for the main keywords look stable. Where would you look?” A shallow answer jumps straight to a conclusion, often blaming a recent algorithm update or recommending more content. An expert answer treats it as an investigation. They will mention checking Google Search Console for crawl and indexing changes, reviewing canonical tags, HTTP status codes, sitemaps, internal linking, and JavaScript rendering, and cross-checking a fresh crawl against Search Console data and, where needed, server logs. The willingness to gather evidence before naming a cause is the signal.

Ask what they would not be sure about

Expertise includes knowing the limits of what can be known. Ask how they would forecast results, or how confident they are that a change will improve rankings. A shallow answer promises specific outcomes or timelines with certainty. An expert answer explains that search results depend on competitors and on factors outside anyone’s control, so they work in ranges and probabilities, test changes, and measure rather than guarantee. Comfort with uncertainty is a stronger sign of competence than confidence is.

Ask about a recent change in search

Search shifts constantly, so current knowledge is a fair test. In early 2026, Google rolled out a Discover update in February and both a spam update and a broad core update in late March. The spam update widened enforcement against scaled, low-value AI content, expired domain manipulation, and site reputation abuse. The core update placed more weight on topic-level expertise, original value, and user experience signals. You do not need to memorize this yourself. Ask the company what has changed in search recently and how it affects their approach. An expert can describe a specific recent update and explain a concrete adjustment they made because of it. A shallow answer speaks in vague terms about Google “always changing” without naming anything.

Ask them to show their own results

A direct test is to ask how the company’s own website performs for the terms it competes on, and to walk you through it. Companies that practice what they sell can do this comfortably. Companies that cannot, or that become uncomfortable with the question, are telling you something useful.

What to listen for

Across all of these questions, the pattern is consistent. Shallow expertise sounds like a checklist: tactics named in order, outcomes promised, updates blamed in general. Real expertise sounds like a diagnosis: questions asked back to you, evidence gathered before conclusions, tradeoffs made explicit, and honesty about what is uncertain. Ask questions that have no scripted answer, then judge the quality of the thinking.

An SEO company can offer informed projections about future performance, but it cannot predict it with precision, and it should never guarantee a specific result. Honest forecasting is useful for planning a budget and setting expectations. Honest forecasting also means being clear about what the numbers can and cannot tell you.

Why precise predictions are unreliable

The core reason is simple: no SEO company controls Google. Search results are produced by ranking systems that Google owns, updates frequently, and does not fully disclose. Google runs broad core updates several times a year, along with many smaller changes, and any of these can shift a site’s visibility within days. A forecast made in one quarter can be undercut by an algorithm change in the next.

Other factors are also outside an agency’s control. Competitors publish new content, earn links, and improve their own sites. User behavior changes. The way Google displays results changes, too: a growing share of searches end without a click to any website because the answer appears directly on the results page or inside an AI-generated summary. A page can hold a strong ranking position and still send less traffic than the same position would have sent a few years ago.

There is also a limit built into forecasting itself. Most projections are based on past patterns, such as how a site’s traffic has grown historically or how similar keywords tend to perform. A pattern-based model cannot account for things it has never seen, including a major site overhaul, a new competitor, or a change in how a search topic is presented. The further out a forecast reaches, the wider its margin of error becomes.

What a reasonable projection looks like

A credible SEO company will present projections as a range rather than a single confident number, and it will state the assumptions behind that range. A reasonable projection is usually built from observable inputs: current keyword positions, existing search demand for those terms, realistic click-through rates by position, and the site’s track record of progress. From there, the agency can describe a plausible band of outcomes, often with conservative, moderate, and optimistic cases.

Good projections are tied to a specific plan of work. If the forecast assumes a set of technical fixes, a content schedule, and a link-building effort, those assumptions should be written down so you can check progress against them. The projection should also come with a timeframe. Meaningful SEO results typically take months, not weeks, and a forecast that promises fast, large gains deserves scrutiny.

A reasonable projection focuses on direction and magnitude rather than exact figures. It is meant to answer questions like “Is this investment likely to pay off, and roughly when?” It is not meant to lock in a guaranteed traffic number or ranking position.

Guarantees are a warning sign

If an SEO company guarantees a number one ranking, a fixed amount of traffic, or a specific revenue figure, treat it as a red flag. Google itself states that no one can guarantee a top ranking. A guarantee usually means one of two things: the company is overpromising to win the contract, or it is targeting easy, low-value terms that were never in real competition. Neither serves your business.

What to ask for instead

Rather than a guarantee, ask the company to explain how it builds its projections, what data it uses, and which assumptions the numbers depend on. Ask how it will report progress, how often it revisits the forecast as conditions change, and what it will do if results fall short of the moderate case. A company that forecasts honestly will welcome those questions, show you the reasoning behind its estimates, and update them as real data comes in.

The most reliable signal of future performance is not a promised number. It is a sound diagnosis of your site, a prioritized plan of work, and a clear record of consistent progress against that plan over time.

When an SEO company evaluates a backlink, it is asking one core question: would this link exist if search engines did not? A link that a real editor or writer chose to include because the content was useful carries weight. A link placed mainly to influence rankings carries little, and may carry risk. To make that judgment, an agency looks at several signals together rather than relying on any single score.

Relevance of the linking site and page

Relevance is usually the first thing an SEO company checks, and many practitioners treat it as the most important factor. A link is stronger when the page it sits on, and the site as a whole, covers a topic related to your business. A link from an article that genuinely connects to your subject reinforces what your page is about. An unrelated link, even from a well-known site, sends a weaker and sometimes confusing signal. The agency reviews the surrounding content, the site’s general focus, and whether the link appears in a context where a reader would naturally find it useful.

Authority and trust of the referring site

Authority is the second consideration. Search engines pass value through links, so a link from a site that has earned trust over time tends to count for more. SEO companies often reference third-party metrics such as Domain Rating or Domain Authority as a quick gauge. These are estimates produced by tools, not figures from Google, so a careful agency treats them as a starting point rather than a verdict. Trust matters as much as raw authority: a site with a clean history and a real audience is preferable to one that simply has a high score.

Editorial links versus paid or manipulative links

A central part of the assessment is whether a link was earned editorially or placed through payment or a link scheme. Editorial links are included by a person who decided the content was worth referencing. Paid placements, link exchanges, and links from low-quality networks are different. They are not automatically harmful, but if they are meant to influence rankings they should carry a “sponsored” or “nofollow” attribute. Search engines have grown more capable at detecting unearned links through pattern analysis, so an SEO company looks for signs of manipulation: footprints repeated across many sites, link sections unrelated to the content, or pages built mainly to host links.

Anchor text profile

Anchor text, the visible words in a link, is reviewed at the profile level rather than one link at a time. Anchors that relate naturally to the destination topic send a helpful signal. A healthy profile is varied: it includes brand mentions, plain URLs, partial-match phrases, and generic wording. A profile dominated by exact-match keyword anchors is a common warning sign of manipulation, so an agency checks the overall mix and flags patterns that look engineered.

Traffic and real audience of the referring page

Finally, an SEO company considers whether the linking page reaches actual people. A link on a page with real visitors can send referral traffic and reflects a site that search engines treat as active and credible. A link on a page with no traffic offers little and may indicate a site that exists only to publish links.

Why quality beats quantity

A small number of relevant, editorially earned links from trusted sites generally does more for rankings than a large volume of weak or unrelated ones. Low-quality links add risk without adding value, and cleaning them up later takes time. A competent SEO company therefore measures links against these signals before pursuing them, and reports on link quality in those terms rather than simply counting how many links were acquired.