How to Do Keyword Research When Hiring an SEO Consultant or Agency

If you’re searching how to do keyword research, you probably don’t need another generic tool tutorial. You need to know how search demand should be turned into a commercial SEO plan, and how to judge whether a consultant or agency has done that work properly.

Josh Willett Consultancy Ltd works UK-wide as an independent SEO consultancy, drawing on 8+ years of SEO experience, a former front-end development background, and work with 50+ clients across the UK and Europe. This guide shows my keyword research process for turning search data into page decisions, content priorities, and hiring checks.

Why Hiring Changes the Keyword Research Job

The Real Deliverable

Keyword research is not finished when someone gives you a spreadsheet of search terms. The useful output is a decision system that shows which pages to create, which pages to improve, which opportunities to ignore, and which work should happen first based on commercial value, ranking difficulty, and buyer intent.

Most beginner guides teach the mechanics: open a tool, enter seed keywords, export ideas, sort by volume. That’s fine for learning. It’s not enough when you’re paying a consultant or agency to make decisions that affect budget, content, developers, and sales pipeline.

Why The Risk Is Different

Google’s own guidance says hiring an SEO can improve visibility and save time, but can also risk damage to a site and its reputation if the work is irresponsible. The same guidance lists keyword research, technical advice, content development, and SEO training among useful services an SEO may provide, and recommends asking how success will be measured through Google’s SEO hiring guidance (developers.google.com).

That warning matters because keyword choices shape what gets built. A weak keyword research process can send a business into months of content production for pages that don’t fit the buyer journey.

What A Buyer Should Expect

A proper project should produce more than a raw export. It should connect the search market to your business model.

That distinction matters. A list tells you what people search. A keyword strategy tells you what your site should do about it.

The Consultant Test

When I review a search plan, I want to see 5 things before I trust it:

• The terms should connect to products, services, sectors, or buyer problems.
• The current site should be checked before new pages are recommended.
• Search intent should be reviewed in the live results, not assumed from a label.
• Priorities should be tied to commercial value, not just volume.
• The work should be phased so the first month doesn’t try to do everything.

That’s why hiring changes the job. You’re not buying data. You’re buying judgement.

First: Define Commercial Outcomes Before Search Terms

The first step is to decide what the search work is meant to produce. For a B2B service business, that might mean qualified enquiries, audit requests, booked calls, demo requests, partner applications, or better-fit leads from specific industries.

A common mistake is starting with broad topics relevant to the business and assuming the plan will become commercial later. It usually doesn’t. If the project starts as a traffic exercise, the output tends to become a traffic plan.

Set The Commercial Frame

Before any tool work, I’d define the frame in plain business terms. This is where an experienced consultant should slow the process down.

The starting questions are:

• Which services have the best margin or strategic value?
• Which customer types are most likely to buy in the next 3 to 12 months?
• Which locations, sectors, or use cases matter commercially?
• Which existing pages already rank, convert, or attract assisted enquiries?
• Which topics would be a bad fit even if they have demand?

Apply The Business Test

Those questions stop the research becoming disconnected from revenue. They also make it easier to say no later.

The business test should be applied before the keyword research process expands into tools, exports, and keyword lists. If a term can’t be connected to a real service, buyer problem, or strategic market, it should be treated as a weak candidate until there’s a clear reason to keep it.

Turn Outcomes Into Filters

The commercial frame should become a filter used throughout the keyword research process. If a phrase doesn’t connect to a profitable service, a valuable audience, or a strategic market, it needs a stronger reason to stay in the plan.

This is also where a consultant should ask about capacity. There’s no point prioritising a service line the business can’t deliver, a location it doesn’t want to serve, or a lead type the sales team will reject.

Build A Simple Scoring Model

A practical scoring model doesn’t need 20 columns. It needs enough structure to compare opportunities without pretending the keyword data is perfect.

I use this type of scoring because search tools don’t know your margins, sales cycle, or capacity. A phrase with 90 monthly searches can be worth more than one with 9,000 if it attracts the right buyer.

Tie Search Work To A Sales Reality

Where the sale is complex, one page rarely does the whole job. A decision-maker might read an educational guide, compare consultant versus agency, check pricing, review proof, and only then enquire.

That means the search plan should include commercial routes, not isolated pages. A good consultant should be able to explain how a page supports the next commercial step, even when the page itself is informational.

If that reasoning is missing at the start, the later spreadsheet will look tidy while the strategy stays vague.

Then Build Seed Terms From Services, Buyers, And Evidence

Seed terms are the starting inputs used to generate a wider set of search ideas. They should come from your actual offer, your customers’ language, your sales conversations, your current site, and your competitors’ visible search coverage.

Most people make this step too narrow. They start with the service name and miss the situations that trigger a purchase. For example, “SEO consultant” is a service label, but “why has organic traffic dropped” is a problem a buyer may bring to the table.

Use Five Input Sources

A consultant should not rely on one brainstorm. The better approach is to pull starting points from several places and then compare them.

This is where Reddit-style results can be useful without becoming the strategy. A forum thread ranking for a topic tells you searchers may want lived experience, direct answers, and fewer polished marketing claims.

Expand With Tools Carefully

Once the inputs are clear, tools can expand the market. A keyword magic tool, Google Keyword Planner, or phrase-matching report can help uncover adjacent terms, question formats, modifiers, and long-tail demand.

The risk is letting the tool redefine the brief. Good seed keywords should come from business evidence first, then tools should test and expand them.

Separate Labels From Buyer Language

In one review for a B2B services site, the internal team kept using a technically accurate service label. Searchers used a plainer phrase that described the business problem. The fix wasn’t to stuff the page with both.

The fix was to build one page around the buyer phrase, then explain the technical service inside it.

That happens often. Internal language is tidy. Buyer language is messy.

Use Tools After The Inputs Are Clear

Tools are useful once you know what you’re trying to expand. Ahrefs notes that many research tools ask for a seed term and then generate wider keyword ideas, with its free generator returning 20 related terms and 20 related questions per input in the example given by Ahrefs’ guide (ahrefs.com).

Google Keyword Planner, Google Trends, Search Console, Ahrefs, Semrush, Moz, and smaller free tools can all help at this stage. The tool choice matters less than the discipline of starting from business evidence.

If the initial inputs are weak, the export gets bigger without getting better.

Next: Check Demand, Difficulty, And SERP Reality

Search volume data is useful, but it’s not a forecast. It’s an estimate of average searches, not a guaranteed traffic number, and it needs to be checked against the live search engine results page before a page decision is made.

This is where a consultant earns their fee. The tool gives clues. The live results show what Google is currently rewarding.

Read The Metrics Carefully

Ahrefs makes several points buyers should understand. Its guide says search volume counts searches, not people, and that even a number 1 ranking will rarely send traffic equal to more than 30% of that single term’s volume. It also says an average top-ranking page can rank for about 1,000 similar terms.

That changes the decision. You’re not trying to rank for one phrase in isolation. You’re trying to build the best page for a cluster of closely related searches.

A difficulty score is not a verdict. Ahrefs says its KD score is based on the number of unique websites linking to the current top 10 ranking pages. That’s useful, but it won’t tell you whether your experience, offer, or site structure gives you a better angle.

Inspect The Live Results

The results page answers questions a spreadsheet can’t. It shows whether Google is ranking service pages, guides, tools, marketplaces, videos, forum threads, local packs, or category pages.

A fast SERP review should record:

• The review should identify the dominant page type in the top 10.
• The review should compare the number of commercial pages with informational pages.
• The review should note AI Overviews, People Also Ask, video, local, or shopping features.
• The review should assess the strength of the ranking domains.
• The review should check the depth and freshness of the pages currently winning.

Treat The SERP As Evidence

This is where a lot of cheap research fails. The spreadsheet says “blog post”. The results show service pages.

The live results should be treated as stronger evidence than a tool label because they show what searchers are choosing and what Google is already testing. If the SERP rewards comparison pages, local service pages, or forum-led answers, the keyword strategy needs to reflect that reality before a brief is written.

Spot The Mismatch

Sometimes the spreadsheet says “high volume”, but the results show a definition-heavy query with no buyer intent. Sometimes a tool suggests an article, but Google is rewarding product pages, category pages, or local service pages.

Those mismatches should change the page decision. If they don’t, the keyword research process is being driven by software rather than evidence.

Define Search Intent In Practical Terms

Search intent means the purpose behind a search. I define it in work terms: what page would satisfy the searcher enough that they don’t bounce back and click another result?

For hiring decisions, I’d use 4 working categories:

Intent can change fast when a result set shifts. If a forum thread and a newer agency how-to enter the top 10, that suggests users may want practical examples, simpler explanations, and clearer decision support.

Decide What A Better Result Needs

The practical question is not “what label does the tool show?” The practical question is “what would a better result need to include?”

A consultant should be able to answer that before briefing content. If the answer is vague, the final article or service page will usually be vague too.

Once the live results are understood, the next mistake is creating too many pages for the same intent.

Map Intent To Pages Before You Write

The right output is not “one search term equals one page”. The right output is a map that decides which terms belong together, which terms need separate pages, and which terms should be handled inside an existing page.

This is where site architecture matters. A consultant with technical SEO experience should be checking indexability, internal links, canonical issues, URL structure, and current content before recommending new writing.

Choose Create, Update, Consolidate, Or Ignore

Every opportunity should land in one of 4 buckets. This keeps the plan usable for marketing teams, writers, and developers.

This bucket system prevents content bloat. It also protects the site from cannibalisation, where multiple pages compete for the same query and weaken each other.

Work From The Current Site

If you’re hiring someone, ask whether they checked your current pages before recommending new ones. This sounds basic. It’s one of the most common gaps I see.

The audit should look at:

• The audit should identify existing URL targeting.
• The audit should check current rankings and impressions.
• The audit should review internal link depth.
• The audit should find duplicate or near-duplicate page intent.
• The audit should decide whether the page already has the authority to be improved instead of replaced.

Scale The Review To The Site

For a site with 50 pages, this review may take a few hours. For an ecommerce or enterprise site with thousands of URLs, it may need a crawler, Search Console export, analytics data, and a mapping sheet.

The scale should match the risk. A small brochure site needs enough review to avoid obvious overlap, while a large site needs a more formal mapping process because one wrong template or category decision can create hundreds of weak or competing URLs.

Use Internal Links As Evidence

Internal links are not just a production detail. They show which pages the business considers important, how authority flows through the site, and whether a new page can realistically be supported.

For example, a guide about hiring SEO support should naturally connect to a relevant service page such as SEO consultancy. If there is no sensible internal link route, the page may be too isolated or too far from the commercial journey.

Avoid One-Page-Per-Phrase Thinking

A single strong page can rank for many related searches if the intent is the same. Creating separate thin pages for tiny variations can leave you with more URLs and less authority per URL.

The better question is: would a searcher expect a different answer?

If the answer is no, group the terms. If the answer is yes, create a separate page with a different job.

Protect The Site From Bloat

Page mapping should make the site simpler to understand, not larger by default. A plan that adds 40 pages without explaining consolidation, pruning, or internal linking is usually avoiding the harder strategic work.

That decision shapes the whole content plan, because writing starts after the page map is settled.

Prioritise With The 80/20 Rule

The 80/20 rule in SEO means a small number of actions usually create most of the measurable gain. In search planning, that means prioritising the pages, fixes, and terms most likely to produce commercial movement instead of spreading effort evenly across every idea.

This is not about doing less work. It’s about sequencing work properly.

Score Opportunities Against Four Factors

I’d usually score each opportunity against 4 factors before turning it into a brief. Each factor can be scored from 0 to 3.

The best early opportunities are rarely the biggest terms. They’re the ones where demand, buyer value, and feasibility overlap.

Use A 90-Day View

A 90-day plan keeps the work grounded. It’s long enough to include research, content, technical fixes, and initial measurement, but short enough to avoid vague annual planning.

A sensible first phase might include:

• The team should fix 3 to 5 pages already close to page 1.
• The team should create 2 to 4 new pages with clear buyer value.
• The team should improve internal links to the highest-value commercial pages.
• The team should clean up obvious duplication or cannibalisation.
• The team should set a baseline for rankings, enquiries, assisted conversions, and non-brand clicks.

Measure Before Expanding

That gives the business something measurable. It also gives the consultant evidence for the next phase.

Measurement should happen before the plan expands because early results can reveal whether the assumptions were right. If a refreshed page earns more qualified clicks, the next phase can build from that pattern. If it doesn’t, the keyword data, page type, offer, or internal links may need review.

Don’t Hide Behind Volume

A high-volume term can be the wrong priority if it attracts students, job seekers, DIY researchers, or buyers outside your market. A low-volume term can be a strong priority if the searcher is describing a service you sell and the results are weak.

This is where I’m usually direct with clients. Traffic that doesn’t create commercial progress is reporting noise.

The 80/20 decision should be visible in the final plan, because a buyer needs to see why some ideas are being delayed.

Make The Trade-Offs Explicit

Prioritisation should show what is being delayed as well as what is being done. That helps the business understand whether the plan is focused or simply incomplete.

A good plan might say, “We are delaying the broad beginner guide because two service pages and one comparison page have clearer buyer value.” That’s a strategic choice, not a lack of ambition.

Validate The Plan Against AI Search And Forum-Led Results

Search behaviour is no longer just ten blue links and a featured snippet. Google’s AI features, forum-heavy result sets, comparison queries, and tool-led pages can all change what a useful result looks like.

That doesn’t mean you need gimmicks. It means the plan should be checked against how results are being assembled now.

Account For AI Search Without Chasing Hype

Google says its AI Overviews and AI Mode can use query fan-out, issuing multiple related searches across subtopics and data sources to develop a response. Google also says there are no extra technical requirements for appearing in AI features beyond being eligible for Search, and recommends foundational SEO work such as crawl access, internal links, visible text, page experience, and matching structured data to visible page content in Google’s AI features guidance (developers.google.com).

For research, that means a page should answer the main query and the adjacent questions a buyer would ask next.

This is why generic overviews are weaker than experience-led guidance. Search systems and buyers both reward specificity.

Respond To Forum Results

When forum-style results rank, the result set is telling you something about trust. Users may prefer blunt answers from people who have tried the thing over polished brand pages that avoid hard choices.

A consultant should respond by adding:

• The page should include clear “what I’d do” recommendations.
• The page should use examples from real delivery work where possible.
• The page should give plain warnings about poor tactics.
• The page should include decision tables that help buyers choose.
• The page should be honest about what the data can prove.

Learn From The Trust Signal

Don’t copy forum answers. Learn from the demand they reveal.

The trust signal is usually the useful part. If people are turning to forums, they may be looking for candour, failure stories, cost context, implementation detail, or someone willing to say when an SEO tactic isn’t worth doing.

Turn Forum Demand Into Better Content

Forum-led results often expose the questions polished pages avoid. These might include cost, ownership, realistic timelines, weak agency deliverables, or what happens when rankings don’t move.

Those are useful clues for the brief. If buyers are looking for candid answers, a safe and generic article won’t satisfy them.

Respond To Agency How-To Results

A newer agency guide in the result set usually covers the expected basics: definitions, tool steps, types of phrases, and broad process. To beat that, your page needs decision support, not another longer version of the same outline.

That means showing how to evaluate the work when you hire someone. It means naming weak deliverables, showing what good looks like, and connecting the process to commercial outcomes.

That’s the gap most guides leave open, and it’s where an independent consultant can add value.

Build The Page Plan And Briefs

The page plan turns research into instructions someone can execute. A weak brief says “write 1,500 words about this topic”. A good brief explains the page’s job, intent, proof, structure, internal links, and success criteria.

If you hire an agency, this is where handovers often break. The strategist does the research, a junior writer gets a thin brief, and the final page reads like a summary of search results.

What Each Brief Should Contain

A usable brief should be clear enough that a writer, editor, developer, or business owner can understand why the page exists.

For this consultancy, that kind of brief would usually tie the page back to SEO consulting, technical SEO, content strategy, or strategic advisory work. The point isn’t to force a sales pitch into every section. The point is to make sure the page has a role in the wider search system.

Define Two Technical Terms

Two terms are worth defining because they affect page planning.

Cannibalisation is when two or more pages on the same site target the same or very similar search intent, causing Google to alternate between them or rank neither as strongly as it could.

SERP intent is the intent Google appears to reward in the current results, based on the page types and formats already ranking.

Explain Why The Terms Matter

Those definitions matter because they prevent a common mistake. More pages can make the site weaker if the pages compete for the same job.

They also help non-SEO stakeholders understand why a consultant might recommend improving one page instead of creating five new ones. That can feel slower at first, but it often produces a cleaner and stronger site.

Keep Writers Close To The Research

Writers need the reasoning, not just the phrase list. If a brief only includes keyword metrics, the writer has to reverse-engineer the strategy and will often fill gaps with generic content.

A good consultant or agency should be able to show the line from search data to outline. If they can’t explain why a section exists, it probably shouldn’t exist.

The brief is where strategy becomes delivery, so the handover quality matters as much as the research quality.

What Should A Consultant Or Agency Hand Over?

A consultant or agency should hand over a decision-led search plan, not just a spreadsheet. The plan should include the data, the reasoning, the recommended page actions, the priority order, and the assumptions behind the work.

This is the easiest way to separate competent work from tool-led work. Tool-led work looks big. Competent work looks usable.

The Minimum Deliverables

For a small or mid-sized B2B site, I’d expect the following as a minimum, because these deliverables show whether the keyword research process has been turned into decisions. Without them, the business may have keyword lists, but it won’t have a reliable production plan.

For larger sites, I’d expect more: crawl data, page templates, indexation checks, taxonomy decisions, content pruning notes, and developer tickets where technical fixes block growth.

The Questions To Ask Before You Sign Off

Ask direct questions. A good provider should be able to answer without hiding behind software.

Use this checklist:

• Which 5 opportunities would you do first, and why?
• Which ideas did you reject?
• Which existing pages are at risk of overlap?
• What changed after you reviewed the live results?
• Which assumptions could be wrong?
• What needs technical work before content work?
• How will we know after 90 days whether the plan is working?

Look For Rejected Ideas

The rejected ideas are especially useful. If every idea makes the plan, the provider hasn’t made strategic choices.

A strong keyword strategy should have a visible reject pile because rejection proves the provider has applied judgement. Ask to see examples of terms removed for poor intent, weak buyer value, duplication with existing pages, unrealistic competition, or lack of fit with your sales priorities.

Ask For The Reasoning

A useful handover should include the “why” behind the recommendations. You don’t need every minor data point, but you do need enough reasoning to approve budget with confidence.

If the consultant can explain why one low-volume commercial page outranks a high-volume guide in the priority list, the plan is probably grounded. If they can only point back to the tool export, it isn’t.

What Good Communication Looks Like

Google recommends asking SEOs whether they’ll explain the changes they make and the reasoning behind them. That’s a sensible hiring standard.

I’d expect a provider to explain the trade-offs in plain English. For example: “This term has lower volume, but the buyer value is higher and the current top results are weak, so I’d do it before the larger informational topic.”

That kind of sentence is worth more than a 20-column export with no recommendation.

How To Review The Work Before You Pay For More Content

Review the research before the writing queue starts. It’s cheaper to fix the plan at the spreadsheet stage than to rewrite 12 pages later.

I’d treat the first review as a quality gate. The provider should walk you through the logic, not just the deliverables.

Use A 10-Point Review

Score the work against the items below. Anything under 7 needs fixing before production starts.

A score of 10 doesn’t mean the plan is guaranteed to work. SEO doesn’t work like that. It means the reasoning is visible enough to test.

Watch For Common Failure Patterns

The same problems show up again and again.

The main failure patterns are:

• Volume chasing happens when high-demand terms are prioritised with no buyer value.
• Tool obedience happens when difficulty scores replace judgement.
• Page sprawl happens when every variation becomes a new URL.
• No technical review happens when content is planned for a site that can’t support it.
• Thin briefs happen when writers get phrases but no angle.
• No measurement plan happens when success is discussed after the fact.

Treat These As Structural Problems

None of these are small issues. They change what gets built.

Treat them as structural problems because they affect the whole production system, not just one document. If the research is chasing volume, ignoring current pages, or skipping technical constraints, every brief that follows will inherit the same flaw.

Check The First Brief Before Production

Before approving a full content batch, review one brief in detail. It should show the target cluster, the page job, the intended reader, the internal links, the proof needed, and the reason the page deserves to exist.

This small review catches most problems early. If the first brief is thin, the next ten usually will be too.

Decide What You’ll Do Yourself

Some businesses can do the first pass themselves. If you have Search Console access, a small site, and a clear offer, you can build a starting map before hiring help.

The limit is judgement. You may be able to find search demand, but you may still need an experienced SEO to decide whether to create a new page, improve an existing one, change the site architecture, or avoid the topic.

That’s usually the point where hiring makes sense.

FAQ

These answers cover the buyer questions that usually sit around the main process. They include the beginner workflow, 80/20 prioritisation, research methods, and the point where DIY work should become specialist support.

How To Do Keyword Research For Beginners?

Start by listing your services, customer problems, and existing pages. Expand those ideas with Search Console, Google Keyword Planner, Google Trends, and one paid SEO platform if available. Then group terms by intent, check the live results, and prioritise pages that connect to revenue rather than traffic alone.

What Is The 80 20 Rule Of SEO?

The 80 20 rule of SEO means a minority of actions often produce most measurable gains. In practice, that means fixing high-value pages close to page 1, improving internal links, and creating pages for buyer-intent searches before spreading budget across low-value informational content.

How Can I Research Keywords?

You can research keywords by combining customer language, Search Console data, competitor page analysis, tool exports, and live result checks. Don’t stop at volume. Check intent, difficulty, page type, commercial value, and whether your site already has a page that could target the same demand.

Can I Do SEO By Myself?

Yes, you can do SEO yourself if the site is small, the market is not too competitive, and you have time to learn the basics. Hire a consultant when technical issues, content prioritisation, migration risk, international targeting, or commercial decision-making become too costly to guess.

Josh Willett Consultancy Ltd works with UK businesses that want search research turned into a clear SEO plan, not a spreadsheet nobody uses. Book a strategy call with Josh in London to review your current search opportunities and decide what should happen first.