Questions to ask before you hire an SEO consultant

If you are choosing between SEO consultants, the first decision is not who looks most convincing. It is what kind of help you actually need. A strategic review, a technical diagnosis, content guidance and local search support are not the same job, and a consultant who is strong in one may be weak in another.

That is why the best questions to ask before hiring an SEO consultant are not about secret tactics or bold forecasts. They are about scope, evidence, method and uncertainty. You want to know what they will inspect, what they can verify, what they will leave as a hypothesis, and how you will tell whether the work was useful.

Start with the decision you actually need to make

Before you compare suppliers, define the problem in plain terms.

If your site has lost visibility, has pages disappearing from search, or has recently changed platform, templates or URL structure, you may need technical diagnosis first. That means checking crawling, indexing, canonicals, redirects, internal links, robots directives and server behaviour before anyone starts rewriting copy.

If your site is indexed but does not seem to earn impressions for the topics you expect, you may need strategic advice or content guidance. That is a different brief. The consultant should be looking at search demand, page intent, duplication, weak coverage, internal linking and whether the site explains the subject clearly enough for users and search systems.

If your issue is map pack visibility, business profile confusion, location pages or inconsistent local signals, then local search help is the right comparison group. Google states that local ranking depends on relevance, distance and prominence, and that position is not something you can buy from Google directly (Google local ranking factors).

If your concern is whether your business is being surfaced in AI answers, cited by assistants, or discovered through newer interfaces, ask for a separate explanation of AI search evidence. A consultant should distinguish between a crawler visit, a citation, an impression and a click. These are not interchangeable. Google’s published guidance says existing Search fundamentals still apply to its AI features, and no special AI schema or text file is required (Google AI features guidance).

A useful first question is simply this:

“What problem do you think we are trying to solve, and what type of SEO work does that actually require?”

If the answer is vague, or if every problem is treated as the same package, you already know something important.

At The SEO Blog, we usually suggest writing the brief in one sentence before you speak to anyone. For example:

• “We lost non-brand traffic after a migration.”
• “Our service pages are indexed but rarely shown.”
• “We appear in organic results but not local pack results.”
• “We want to understand whether AI search is sending any discoverable visibility.”

That sentence keeps the conversation grounded. It also stops a sales process drifting straight into deliverables you may not need.

What to ask about their diagnosis process

A credible SEO consultant should be able to explain how they diagnose before they prescribe. That matters because many recommendations sound plausible whether or not they address the real cause.

Ask these questions early:

“What evidence will you inspect before recommending fixes?”

Listen for specific sources, not just “we’ll audit the site”. For most sites, you would expect some combination of:

• Google Search Console performance data
• Google Search Console page indexing information
• crawl findings from a site crawler
• robots.txt and meta robots checks
• canonical and redirect checks
• internal linking review
• template and page type sampling
• server response checks
• for local work, Google Business Profile and location page review

If they mention Google Search Console, ask the follow-up:

“Which reports in Google Search Console matter most for our issue, and what would you be looking for?”

A sound answer might include the Performance report for queries, pages, dates and device patterns, and the Page indexing report to understand why URLs are excluded or indexed. Google’s own documentation is clear that excluded URLs need interpretation, not blanket “fixing”, because not every excluded page should be indexed (Google Page indexing report).

If the consultant jumps straight from “some pages are excluded” to “we must force everything into the index”, that is a warning sign. Diagnosis should distinguish between intentional exclusion, duplication, canonical consolidation and genuine indexing failure.

“Will you inspect crawling behaviour, and how?”

You are listening for practical understanding of how search engines discover and revisit pages. A consultant does not need to make a drama out of log files on every project, but they should be able to explain when crawl evidence matters. They may use a crawler to model internal discovery paths, inspect orphaned pages, identify redirect chains, or compare XML sitemaps with indexable URLs.

If they mention bot activity, they should keep the bots straight. Googlebot is not the same thing as OAI-SearchBot, and OAI-SearchBot is not the same thing as GPTBot. OpenAI’s own bot documentation separates search access from model training controls (OpenAI crawler documentation). That distinction matters if somebody starts making claims about AI visibility based on the wrong crawler.

“How do you separate symptoms from causes?”

This is one of the most revealing questions before hiring an SEO consultant. A serious answer might sound like this:

• ranking drops can be page-specific or site-wide
• traffic drops can come from demand changes, tracking changes or SERP changes
• indexing issues can be intentional, accidental or irrelevant
• content underperformance can be caused by weak alignment, poor structure, duplication or low discoverability
• local visibility can vary by searcher location, category relevance and business profile completeness

A weak answer sounds like “we optimise everything”.

“What access do you need, and what can you assess without it?”

This helps you judge both competence and process. Some consultants ask for full CMS, hosting and analytics access before they have even defined the task. Others can begin with a limited review and then explain what additional access is justified. In the UK, that is also a practical governance issue. If you are giving access to business systems, you should know why, for how long, and what outputs you will receive.

If you want a clearer picture of what can and cannot be inferred from Search Console alone, our guide to Search Console evidence and its limits is a useful companion.

How to compare evidence, not promises

Once two consultants appear competent, compare how they support claims.

Rankings

If someone says they can improve rankings, ask:

“For which pages, which queries, and what evidence would show progress?”

Rankings are not a single number. They vary by query, location, device, SERP layout and date. Google’s own starter guidance does not promise rankings for following best practice, and neither should a consultant (Google SEO Starter Guide).

Better evidence looks like:

• specific page groups
• query classes, such as service terms or informational terms
• date comparisons with context
• explanation of what changed on the site

Poor evidence looks like a screenshot of one favourable query with no baseline.

Traffic

If they claim they will increase traffic, ask:

“Which channel and which measurement source?”

Organic clicks in Google Search Console are not the same as sessions in analytics. Branded and non-branded traffic should usually be separated. A good consultant will also warn you that more impressions can coincide with lower click-through rates if the site starts appearing for broader searches.

AI visibility

Claims about AI search need especially careful handling because evidence is still patchy and easy to overstate.

Ask:

“What do you mean by AI visibility, and how would we verify any change?”

A useful answer should separate:

• crawler access by systems such as OAI-SearchBot or GPTBot
• citations in AI answers
• referral traffic from AI products
• indirect effects such as more branded searches after exposure elsewhere

These are different signals. A bot visit does not prove citation. A citation does not prove clicks. And a rise in branded search does not prove causation from AI interfaces. If you want a grounded overview, see our AI search and visibility guide and our article on AI search citations and what they do and do not prove.

Local search

If the offer includes local search, ask how they will assess relevance, distance and prominence in your actual trading area. If they talk as though local position can simply be “boosted” by buying a package, they are skipping over how local search works. You should expect discussion of categories, business information consistency, location landing pages, reviews as a business process rather than a numbers game, and the limits of rank tracking for location-sensitive results.

Primary guidance

One simple test works across all these areas:

“Which published guidance supports this recommendation, and where are you making an informed judgement rather than citing a rule?”

That question does not require the consultant to recite documentation. It checks whether they can distinguish established platform guidance from their own interpretation. That is often the difference between a careful adviser and a slogan merchant.

Questions that expose risky or vague SEO offers

Some offers reveal their weakness as soon as you ask for detail.

“Do you guarantee rankings?”

A guarantee is the clearest red flag. No consultant controls Google’s results, local pack placement, AI citations or competitor behaviour. They may guarantee work, response times or reporting cadence. They should not guarantee outcome positions.

“Are links included, and if so, what exactly are they?”

If the answer is “a package of authority links” with no explanation of source, placement, relevance or editorial control, treat it cautiously. Paid link schemes are often sold in language designed to avoid saying what they are. Ask whether links are earned editorially, sponsored, directory-based, digital PR-led, or placed on owned networks. If they cannot answer plainly, that is the answer.

“Who owns the work if we stop?”

This matters more than many buyers realise. Ask who owns:

• written copy
• research documents
• dashboards
• crawl exports
• schema implementations
• page changes
• accounts and platform access

If the consultant builds everything inside their private systems and you leave with little more than a slide deck, you may be buying dependency rather than help.

“What is included in implementation, and what remains advice only?”

Some consultants diagnose well but do not implement. That is not a problem if it is clear. It becomes a problem when an offer bundles strategy, recommendations and implementation language so loosely that you cannot tell what will actually happen.

“How will you report uncertainty?”

This is one of the strongest filters. Good SEO work often involves uncertainty. A professional consultant should be able to say:

• what they know
• what they suspect
• what they need to test
• what may take time to observe
• what cannot be attributed confidently

If every statement is certain and every recommendation is urgent, caution is sensible.

“What happens if the initial diagnosis is wrong?”

This sounds awkward, but it is practical. You want to hear that they will revisit assumptions, inspect new evidence and change course. Not that they will keep delivering the same monthly package regardless.

For a broader framework on responsible interpretation, our piece on how to investigate a change in search performance shows the kind of evidence trail worth following before drawing conclusions.

A practical scorecard for choosing between two consultants

You do not need a giant procurement process to make a sensible choice. A one-page scorecard is usually enough.

Create five columns:

• Criterion
• Consultant A
• Consultant B
• Evidence seen
• Notes and uncertainty

Then score each criterion from 1 to 5, where 1 means unclear or weak and 5 means clear and well supported.

1. Fit to the actual problem

Did they identify whether you need technical diagnosis, content guidance, strategy or local search help?

A strong score goes to the consultant who narrowed the brief sensibly, even if that means a smaller initial scope.

2. Diagnosis method

Did they explain what they will inspect in Google Search Console, crawling, indexing and page-level evidence before making recommendations?

A strong score goes to the consultant who described process in concrete terms.

3. Evidence standard

Did they support claims with examples, data sources and published guidance where relevant?

A strong score goes to the consultant who distinguished evidence from assumption.

4. Scope clarity

Can you tell what is included in the first phase, what outputs you will receive, and what is not included?

A strong score goes to the consultant whose scope could be handed to a colleague without interpretation.

5. Reporting and communication

Will you receive a dashboard, written findings, implementation tickets, calls, or all three? How often? In what format?

A strong score goes to the consultant who explains how decisions will be documented, not just discussed.

6. Ownership and access

Do you retain access to accounts, documents and implemented assets?

A strong score goes to the consultant who avoids unnecessary lock-in.

7. Treatment of uncertainty

Did they acknowledge what cannot yet be proved, especially around AI visibility, local variation and attribution?

A strong score goes to the consultant who is comfortable saying “we need to test that”.

8. Risk signals

Did they avoid guarantees, vague link packages, inflated claims and bundled promises that cannot be checked?

A strong score goes to the consultant with the fewest avoidable red flags.

Here is a simple worked example.

Consultant A says your problem is “low authority”, proposes content, technical fixes, local SEO and backlinks in one monthly package, and promises page one improvements. They mention Google Search Console but do not specify which reports they will use. They say AI visibility will improve because they “submit to AI engines”. Score that low on diagnosis, evidence and risk.

Consultant B says your recent redesign may have affected internal linking and indexable page templates, asks for temporary Google Search Console access, samples affected URLs, explains that they will compare indexed versus intended pages, and says AI visibility cannot be promised but crawler access and citations can be monitored separately. They do not promise rankings. Score that higher, even if the proposal sounds less dramatic.

That is often the awkward truth in these buying decisions. The better consultant may sound less exciting because they are describing real uncertainty instead of selling certainty.

If you want a structured way to record your current baseline before speaking to suppliers, our search measurement baseline workbook can help you gather the evidence you will want during those conversations.

The aim is not to turn an editorial guide into a quotation page. That is why The SEO Blog keeps this supplier assessment separate from Suffolk Web, the service business under the same Fleeta Limited ownership. Here, our job is to help you ask better questions, read answers more critically, and recognise when a recommendation is grounded in evidence rather than theatre.

The best questions to ask before hiring an SEO consultant are the ones that force method into the open. What are you diagnosing, with which evidence, under what assumptions, and how will we know the difference between a fix, a test and a guess? If you get clear answers to those, the choice usually becomes much easier.