15 Questions to Ask Before You Hire Anyone for SEO

SEO is an industry with no licence, no required qualification, and no standard unit of delivery. Anyone can call themselves an expert on Monday and invoice on Friday.

That is not automatically a problem: plenty of unregulated professions produce excellent work. It does mean the buyer carries the entire burden of evaluation, usually while knowing less about the subject than the person selling to them.

The usual advice is to check case studies and testimonials. That advice is close to useless: case studies are unverifiable by design, and testimonials are selected by the person publishing them. These fifteen questions work better, and none of them requires technical expertise to interpret.

1. What did you find when you looked at my site?

The single highest-value question, and the one to ask first.

A competent provider will name something specific: your product pages share templated titles, your blog sits on a subdomain sharing no authority with the main site, four pages compete for one query, your content renders client-side so Google sees an empty page.

A weak one will describe their methodology: “We begin with a comprehensive audit, then move to on-page optimization, then build authority.”

The difference is audible to a non-expert. Describing a real finding requires having looked, which requires competence.

2. What would you look at first, and why?

A follow-up that catches anyone who dodged the first one. A competent provider names a specific likely constraint. A weak one describes a process.

3. Under what circumstances would you tell me not to do this?

There are real answers. If your margins cannot support the acquisition cost. If the searches do not exist at meaningful volume in your market. If your conversion problem sits somewhere search cannot reach: a broken checkout, a phone nobody answers, a price that is not competitive. If your domain carries a history that needs resolving first.

A provider who cannot name a single scenario has never turned down work, which tells you exactly how their recommendations are generated.

This question also changes the dynamic. It signals that you are evaluating judgement rather than enthusiasm, and tends to produce more honest answers to everything after it.

4. Do I own the Search Console, Analytics and Business Profile accounts?

The least negotiable item on this list.

Every property should be created under your ownership with access granted to the provider, never the reverse. When a provider holds these accounts, leaving costs you your entire measurement history, the exact data any future provider would need.

The correct answer is immediate and unambiguous. Hesitation here is disqualifying on its own.

5. Who actually does the work, and can I speak to them before signing?

At an agency, the person who wins the account is rarely the person delivering. That is not a scandal (it is how the economics function), but you should know which product you are buying, because pricing overlaps enough that businesses routinely compare quotes for entirely different things.

An evasive answer means you are buying an account manager and a templated deliverable.

6. What happens to my project if you become unavailable?

The mirror-image risk when hiring an individual rather than a firm. One person gets ill, takes a larger contract, or moves on.

There are good answers: a documented handover, a named collaborator, work delivered in a form your own team could continue. Treat a vague answer as the answer.

7. What is the metric, and when should I expect to see it?

“Improved rankings” is not a metric. Which queries, from what position to what position, measured how, reported where?

The strongest version is a named business metric (qualified enquiries from organic search, or organic revenue) with a stated timeline. The honest range for most sites: technical fixes show within weeks, content and authority work needs three to six months to read clearly.

Anyone guaranteeing a specific position by a fixed date is either misleading you or planning to target queries so obscure that ranking for them means nothing.

8. Can I see a real report before I commission one?

Ask for a redacted sample from an actual engagement.

You are checking whether it connects activity to outcome. “We rewrote these six page titles because they were competing for the same query, and here is what happened to impressions on those pages over the following six weeks” is a report. A list of tasks performed is an invoice with formatting.

Watch specifically for rank tables presented without click data. Positions can improve while traffic falls: click-through rate at positions one to five declined roughly 17.92% year over year per GrowthSRC’s 200,000-keyword study, as AI Overviews and SERP features absorbed clicks. A report showing rising positions and flat enquiries is telling you the targeting is wrong, and a provider who does not say so is not doing the job.

9. What is your position on directory submissions and bought links?

In most markets, bulk directory submission has been worthless for a decade. In thin markets it is actively harmful, because the available directories are largely low quality or spam and submitting to two hundred of them manufactures exactly the footprint that link spam systems detect.

The honest answer in 2026 is that directories are worthless at best. Enthusiasm for volume link acquisition is a signal about what is being sold.

10. Would you ever recommend deleting content?

Healthy SEO work involves subtraction. Thin pages consolidated, duplicate pages competing for one query merged, content that no longer serves anyone removed.

A provider whose recommendations are exclusively additive has a business model reason for that. It also matters more than it used to: commodity content that can be assembled from general knowledge has lost most of its value as AI systems answer those queries directly, and a site carrying forty pages of generic advice is diluting the pages that could rank.

11. What do you need from me, and what happens if I do not provide it?

Engagements fail from the client side as often as the provider side, and a provider who has thought about this will tell you plainly.

The usual requirements: account access granted early, one named person who can approve a change, a developer or explicit permission to work without one, and honesty about site history, such as a previous provider who bought links, a secondhand domain, a migration that went badly. None of it is disqualifying and all of it changes the diagnosis.

A provider who has never had to ask for these has not run many engagements.

12. What does month seven look like?

Most proposals describe the first ninety days in detail and go vague afterwards. That is where the money actually gets spent.

Ask what the work becomes once the obvious fixes are done. A good answer describes a shift, from technical remediation to content and authority, or from broad coverage to conversion. A weak answer describes the same monthly activity continuing indefinitely, which is the definition of a retainer that has stopped being diagnostic.

13. How do you decide what NOT to do?

Every site has more possible improvements than budget. A crawl of a mid-sized site surfaces hundreds of issues, and most of them do not matter.

The skill being purchased is not the ability to find problems: a tool does that in twenty minutes. It is knowing which four of the four hundred actually constrain this specific site, and having the discipline to ignore the rest.

A provider who describes their process as fixing everything the audit found is describing a very expensive way to change nothing. Ask how they prioritise, and listen for whether the answer references your business or a generic severity score.

14. What happens in the first two weeks?

Specific enough that vagueness is obvious.

A good answer: access granted and verified, a full crawl, a baseline captured, Search Console and Analytics reviewed, and a written diagnosis by the end of the fortnight. Something concrete arrives before the first invoice.

A weak answer describes onboarding and discovery calls. If the first month produces no finding you could act on independently, you are funding ramp-up rather than work.

15. Can I end this, and what do I keep?

The exit terms tell you how confident a provider is in the work.

Month-to-month means the engagement has to keep justifying itself, which is the correct pressure. Long lock-ins protect the provider from the consequences of underperforming, though there is an honest counterargument worth accepting: search genuinely takes three to six months to read clearly, so a provider asking you to commit to that timeline in principle is being reasonable. Asking you to commit contractually is different.

More important than the notice period: what do you retain? Documentation of what was done and why, the content, the accounts, and the data. A provider whose departure leaves you unable to explain your own site’s configuration has built in a dependency.

How to use these

You will not ask all fifteen. Ask 1, 3, 4 and 8: the finding, the decline scenario, account ownership, and a real report. Those four separate providers more reliably than any amount of portfolio browsing, and they take about forty minutes.

Two notes on interpretation.

Price is a weak signal, but the extremes are informative. Below a certain threshold the arithmetic does not permit real work: whatever the local rate, a very cheap retainer buys a handful of hours, and a handful of hours goes on producing something that looks like a deliverable rather than covering diagnosis. At the other end, large retainers with six-month minimums are defensible for genuinely complex sites and poor value for a ten-page business site that needs its indexing fixed.

“Do they rank for their own terms” is the most-cited test and the least reliable. Ranking for “SEO agency [city]” shows a firm can rank a site in a niche where competitors are equally small agencies. It says little about whether they can grow a manufacturing business, and plenty of good practitioners are too busy with client work to compete on it.

Where you buy matters less than what you ask

One structural point worth stating plainly: senior expertise is now available at rates that were not accessible five years ago, sold directly rather than through intermediaries. For diagnostic work, which is knowledge-bound rather than labour-bound, there is no structural reason the output differs by geography.

A business weighing whether to hire seo expert support at Nepali or Eastern European rates rather than a Western agency retainer is making a genuinely reasonable comparison. Independent consultants in these markets publish monthly rates well below Western agency pricing for comparable seniority, and some publish rates openly precisely because the industry’s pricing opacity is a market failure that costs buyers more than it costs sellers.

The caveat applies with full force: variance in emerging markets is wider, which means these questions matter more, not less. A seo consultant pokhara or in Kraków or Manila who answers them clearly is a better purchase than a London agency that cannot. One who cannot answer them is a worse purchase at any price.

The questions do not change with the postcode. Only the invoice does.