Skip to content
busyless
Book a call

B2B SEO strategy

A buying committee searches separately, months apart, for different things

Written by Eugene SuslovLast reviewed 28 August 2026No affiliate links
Sector
Business services
Model
National service
Competition
High
Time to results
9 to 18 months, and the first real signal is not a lead
Typical monthly
$4,000 to $20,000

Key takeaways

  1. 1Volume-led keyword research points the wrong way here. The query that closes a six-figure deal is often searched forty times a month, and the one searched forty thousand times belongs to somebody who will never buy.
  2. 2Gartner puts a typical buying group at six to ten people, each arriving with their own independently gathered information. You are not writing for a persona. You are writing for a committee that never visits your site together.
  3. 3Security and procurement search too, and almost nobody writes for them. A real pricing page and a page about your security posture will outperform most of a blog programme.
  4. 4By the time somebody fills in a form they have already decided something. Treating the conversion as the start of the process is why so much B2B measurement reports the wrong thing.
  5. 5Attribution will not close, and the honest response is to stop pretending. Report how much of the committee your content reaches, not which single touch to credit.

SEO for B2B breaks the assumption that everything else in this directory quietly relies on: that a bigger number means a better keyword. A plumbing emergency is one person with one problem and a decision measured in minutes. A B2B purchase is several people, several months and a decision nobody makes alone.

Gartner's research puts a typical buying group for a complex purchase at six to ten decision makers, each turning up with four or five pieces of information they gathered themselves. Those searches happen at different times, in different words, and the person who searched first is rarely the person who signs.

A B2B SEO strategy therefore has to be built around coverage of a group rather than depth on a person. The champion who found you in March is not the obstacle. The security lead who has never heard of you and gets asked in September is.

This also inverts the keyword arithmetic. Forty searches a month for "[category] implementation timeline" is a rounding error on a report and a live objection inside a deal. The head term with ten thousand searches is mostly students, competitors and people who will never have budget.

The last difference is where the form sits. Buyers spend most of the process without talking to a vendor, so a conversion is a late event rather than an early one. Anything measured from the form fill forward is measuring the last quarter of the job.

Two neighbours are worth naming before going further. Software has its own version of this, driven by review platforms, and industrial suppliers have theirs, driven by specifications. Both hold their own pages in this directory. What follows is about the plural buyer, whatever is being sold.

Who already ranks in b2b

Most B2B SEO work goes wrong before a word is written, at the point where somebody exports a keyword list sorted by volume. What actually sits on the results page is a mix of review platforms, analyst firms, vendor-written listicles and forum threads, and the mix changes completely depending on which member of the committee is typing.

What is on the results page

  • Vendor-written listicles on "best [category] for [use case]", almost always published by a competitor with a content team.
  • Review platform category pages, which frequently outrank every vendor in the category including the market leader.
  • Reddit and specialist forum threads on the practitioner queries, where the answers are blunter than anything a vendor will publish.
  • LinkedIn posts and newsletters surfacing on niche problem queries, particularly where the topic is newer than the category.
  • Analyst content and gated research, ranking on category and market-sizing queries the committee reads before it shortlists.
  • Very thin results on the security, procurement and implementation queries, which is the gap this page keeps returning to.

  • G2

    g2.comClaim it

    Ranks for category and comparison terms across most of software and a growing share of services. The profile is claimable and the category placement is driven by review volume and recency, so a burst of reviews eighteen months ago is worth less than a steady trickle now.

  • Clutch

    clutch.coClaim it

    The equivalent for services, agencies and development shops, and it owns most "best [service] company in [city]" queries. Verified reviews come from interviews rather than forms, which makes them slow to accumulate and harder for a competitor to fake.

  • Gartner and Forrester

    gartner.comClaim it

    You cannot content your way into a quadrant or a wave, and pretending otherwise wastes a year. What you can do is claim and maintain the Peer Insights profile, and read the criteria, because they describe what the committee will be asked to evaluate you against.

  • Capterra and Software Advice

    capterra.comClaim it

    Both sit inside Gartner Digital Markets and both rank on long-tail category and comparison queries. Placement on their lists is partly paid, which is worth knowing before you conclude a competitor has out-marketed you.

  • TrustRadius

    trustradius.comClaim it

    Smaller than G2 and disproportionately read by enterprise evaluators, because the reviews are longer and more structured. Worth the effort specifically if your deals involve a formal evaluation document.

  • Reddit and specialist forums

    reddit.com

    Where practitioners say what they actually think, and increasingly where an answer engine takes its summary from. Not a channel to optimise. The objections raised there are a free list of the pages you have not written.

  • Trade publications

    varies by vertical

    Every B2B market has three or four publications the buyers genuinely read, and they usually rank well on category and regulatory queries. Identify yours by asking customers rather than by looking at domain authority.

  • ThomasNet

    thomasnet.comClaim it

    Still the default supplier discovery route in industrial and manufacturing procurement, and it ranks on component and capability queries a vendor site rarely touches. Relevant only if you sell physical goods or industrial services.

Two columns comparing a keyword list sorted by search volume against one built from lost deals. The volume column surfaces category head terms and definitions; the deals column surfaces pricing, security, implementation and competitor comparisons.
Both columns are real queries. Only one of them contains the questions that ended a deal last quarter.

What people actually search

The clusters below are organised by who is searching rather than by funnel stage, because in this industry those are different axes. A single deal produces queries from six roles, and the four in the middle are where almost nobody publishes anything.

The practitioner with the problem

Informational

> why does our month end close take three weeks

The page that wins it: Genuinely useful problem content that does not require your product

The earliest search and the only one where the buyer does not yet know a category exists. It converts to nothing measurable and it is how you become the vendor the champion already trusts. Nobody funds it, because it looks like a blog.

The champion choosing an approach

Informational

> build vs buy [capability]

The page that wins it: An honest comparison of approaches, including the one where you lose

The buyer has accepted the problem and is deciding how to solve it, not who from. Writing the build-versus-buy page honestly, including when building is right, is what makes the rest of your content credible to somebody who is sceptical.

The evaluator building a shortlist

Commercial investigation

> best [category] for [industry or company size]

The page that wins it: A segment page saying who you are and are not for

The most contested cluster and the one review platforms own. Trying to rank a vendor site for the bare category term is usually futile. Ranking for the qualified version, with a real answer about fit, is not.

The evaluator comparing two vendors

Commercial investigation

> [you] vs [competitor]

The page that wins it: A comparison page that concedes where the competitor is better

Your competitors are already publishing this page about you, and theirs says they win everything. A version that names two things the other vendor does better is the only kind anybody believes, and it is the one that survives being forwarded.

Security, IT and legal

Transactional

> [vendor] soc 2 report data residency

The page that wins it: A trust page with the real artefacts and a named contact

This cluster has almost no competition because vendors put the answers behind an NDA. The searcher is not evaluating you, they are looking for a reason to say no quickly. A page that answers them removes weeks from a cycle.

Finance and procurement

Commercial investigation

> how much does [category] cost per user

The page that wins it: A real pricing page, or an honest explanation of why there is not one

The query every B2B company refuses to answer and every buyer runs anyway. "Contact us" sends them to a competitor's page or to a review site that published a number for you. A range with the variables named beats silence.

The person who will own it afterwards

Commercial investigation

> [category] implementation timeline and resourcing

The page that wins it: An implementation page with real durations and real staffing

The quietest veto in the process. Somebody who will inherit the work is asking whether this becomes their problem. Vendors answer it in sales calls and never in public, so a specific page here is disproportionately valuable.

The champion building an internal case

Informational

> [category] business case template

The page that wins it: Material designed to be forwarded, not to be read on your site

The only cluster where the reader is not the audience. Your page is being pasted into a deck for a finance director who will never visit. That changes what it has to contain: numbers with sources, and no unattributed claims.

Six committee roles listed down the side with the assets that serve each one set out across from them. The first three rows carry two or three assets each. Security, finance and the person who inherits the work carry none.
Whoever is talking to you this week is one of six. The empty rows are the colleagues who will decide without ever reading anything of yours.

What the rules change

There is no B2B regulator, so what constrains this work is general law that happens to bite hard on the way B2B companies market. Every item names its source. None of it is legal advice and the dated items move.

1

Business contact data is personal information in California

California Consumer Privacy Act as amended by the California Privacy Rights Act; the business-to-business exemption lapsed on 1 January 2023

What it means

Until the start of 2023, information collected in a business context about employees of other companies sat outside most of the statute. That exemption expired. A work email address, a job title and a company name belonging to a California resident are now personal information with the ordinary rights attached to them.

So do this

Check that the privacy notice, the rights request process and the opt-out actually cover your lead database and not just website visitors. Most B2B privacy notices were written before the exemption lapsed and were never revisited.

2

GDPR never had a business exemption at all

General Data Protection Regulation, and the ePrivacy rules implemented separately by each member state

What it means

A named person at a company is a data subject regardless of the context. Legitimate interests can be a lawful basis for B2B marketing, but it has to be assessed and documented rather than assumed, and the cookie and tracking rules apply to a business visitor exactly as they do to a consumer.

So do this

Write the legitimate interests assessment down before the campaign rather than after a complaint, and treat the consent banner as a real gate rather than a notice. If it fires the tags before a choice is made, it is not doing the job.

3

Testimonials and reviews are now covered by a specific rule

Federal Trade Commission Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465, effective 21 October 2024

What it means

The rule prohibits fake or misrepresented reviews, buying reviews conditioned on expressing a particular sentiment, and undisclosed insider reviews. It applies to business customers as well as consumers, and it reaches incentivised reviews on third-party platforms as well as testimonials on your own site.

So do this

Never condition an incentive on a positive review, and disclose the material connection wherever an employee, investor or partner has left one. Then audit what your own team has posted on the review platforms, because that is where the exposure usually is.

4

Commercial email has no business carve-out

CAN-SPAM Act and the FTC's implementing rule

What it means

The requirements apply to commercial messages sent to a business address in the same way as to a personal one: a working opt-out honoured promptly, an accurate subject line and header, and a valid physical postal address. There is no exception for a message sent from one company to another.

So do this

Make the opt-out one click and honour it across every sending system rather than just the one that sent the message. The usual failure is a suppression list that exists in marketing automation and not in the sales tool.

Proving expertise

The buying committee is looking for reasons to reject you and each member has a different one. Expertise signals here are less about authorship credentials and more about whether specific, checkable claims exist in public.

  • Named customers with named people, at companies a stranger can verify, rather than anonymised logos and "a leading provider".
  • Numbers with the method attached: what was measured, over what period, against what baseline, and who at the customer confirmed it.
  • A security page carrying the real artefacts, the sub-processor list, the data residency options and a named contact.
  • Public pricing, or a public explanation of the variables when there genuinely is not a list price.
  • Documentation and implementation guides outside the login wall, which is the strongest possible evidence that a product is real.
  • Named authors with operational experience in the buyer's function, not a marketing byline on a technical subject.
  • Comparison pages that concede something, which is the single most reliable trust signal in this category.

How to build a B2B SEO strategy

Ninety days will not move a cycle that runs for a year, so the sequence below front-loads the pages that unblock deals already in progress. A B2B SEO strategy that starts with a blog calendar leaves the security and pricing questions unanswered for another two quarters.

  1. 1

    Weeks 1 to 3

    Get the queries out of the deals you already have

    • Read the last twenty lost deals and write down the objection that killed each one
    • Ask sales which questions get asked on every call, and which ones they answer by sending a PDF
    • List every role that touched the last five closed deals, and what each of them needed to know
    • Check what a competitor's comparison page says about you, and whether it is accurate
    • Claim and correct the review platform profiles that matter in your category

    You end up with
    A query list derived from real deals rather than from a volume export, with a role attached to each entry

  2. 2

    Weeks 2 to 8

    Publish the answers that are currently in sales calls

    • Build the trust page: security posture, certifications, sub-processors, data residency, a named contact
    • Publish pricing, or publish the variables and the ranges and say plainly why there is no list price
    • Write the implementation page with real durations, real staffing and the failure modes
    • Write the comparison pages for your two most common competitors, conceding something real in each
    • Take the three PDFs sales sends most often and republish them as pages

    You end up with
    Four to six pages that shorten a live cycle, all of which answer questions your team is already answering by hand

  3. 3

    Weeks 6 to 20

    Cover the roles nobody writes for

    • Write the build-versus-buy page honestly, including the case where building is the right call
    • Build segment pages that say who you are not for, by industry, size or maturity
    • Publish the business-case material designed to be forwarded, with sources on every number
    • Start the problem-level content for the practitioner who does not yet know a category exists
    • Get named customers on the record with checkable detail, replacing the anonymised logos

    You end up with
    Coverage across the committee rather than depth on one persona, and material that works when it is pasted into somebody else's deck

  4. 4

    Weeks 16 to 40

    Report the committee, not the click

    • Segment Search Console by role rather than by funnel stage, so the security and pricing clusters are visible separately
    • Report how many of the committee-role clusters a given account has touched, where you can see it
    • Track pipeline influenced with the attribution model stated on the page, and the gap stated beside it
    • Compare cycle length on deals that touched the trust and pricing pages against those that did not

    You end up with
    A report a revenue leader can act on, which never claims a single touch caused a deal that took eleven months

Technical fixes with the best payoff

B2B websites fail in a small number of characteristic ways, and almost all of them come from the same instinct: that information is a bargaining chip. Each item below is a place where that instinct costs more than it collects.

  • Everything worth reading is behind a form

    A day per asset

    The whitepaper, the benchmark report and the buyer's guide are all gated, so none of them can rank and none can be quoted by an answer engine. The gate collects a few hundred addresses and forfeits the entire audience that would never fill in a form.

    Publish the content as a page and gate something else: a template, a calculator, a live session. The paper earns the visibility and the tool earns the address. Do this with the highest-value asset first and measure both sides.

  • The resource library has one URL

    A sprint

    A filterable grid of case studies and papers, rendered client-side, where selecting a filter never changes the address. Dozens of genuinely useful documents exist and none of them has a page, a title or a chance of ranking.

    Give every asset a real page with its own URL and its own summary, and make the filters produce real addresses. The library becomes several hundred indexable pages rather than one.

  • Documentation sits behind the login

    A decision, then a sprint

    Product documentation is the most specific, most trusted content a B2B company owns, and it is routinely locked to customers. Evaluators searching for whether the product does a particular thing find a marketing page instead.

    Open the documentation. If some of it is genuinely sensitive, open the rest. This is usually the single largest untapped source of qualified traffic on a B2B domain.

  • Solutions pages built from one template with the noun swapped

    A day of deleting, then real writing

    Twelve industry pages that differ only in the header image and the word in the H1. They read as filler to a buyer and they look like scaled content to a search engine, which is a policy problem rather than a quality one.

    Keep the segments where you have a genuinely different answer, a real customer and a specific objection. Delete the rest. Four honest segment pages beat twelve interchangeable ones.

  • The pricing page says contact us

    A day, and a conversation with the CFO

    The query runs anyway, and it lands on a competitor page, a review site or a forum thread where somebody has posted a number that may be wrong. Refusing to answer does not prevent the answer, it just means somebody else gives it.

    Publish the model even when you cannot publish a number: what drives the price, a realistic range, what a typical first year looks like. Name the variables. This is a commercial decision that marketing cannot take alone.

  • The blog and the product live on different hosts

    A sprint, plus a nervous month

    Content on a separate subdomain or a hosted platform accumulates its own reputation and passes very little of it to the pages that convert. It is a decision usually made for convenience by a team that has since left.

    Move it into a subdirectory on the main domain. Migrate with redirects and expect a dip. It is one of the few structural moves in this list with a compounding effect.

  • Event and webinar pages are never retired

    An afternoon to set the pattern

    Every webinar gets a registration page that stays live forever, so a domain accumulates hundreds of pages advertising things that happened three years ago. They rank for the speaker's name and offer a visitor nothing.

    Convert each past event into a real page with the recording, a transcript and a summary, or redirect it to the topic page. The transcript is often better content than anything the marketing team wrote that quarter.

  • Case studies are anonymised into uselessness

    A week of asking

    "A leading financial services organisation" is a claim nobody can check, so it does the work of neither proof nor content. Legal review usually removed the name, and nobody went back to ask whether it had to.

    Go back and ask. Many customers will agree to be named if asked properly and given approval over the text. Where they genuinely cannot be, publish the mechanism and the numbers with the method, which is still checkable.

Structured data that applies here

The types below fit this industry specifically. Most of them earn no rich result on their own, which is worth knowing before anyone sells the work on that basis. What they do is describe the entity precisely, which matters for how search engines and answer engines resolve who you are.

  • Organization

    The home page, referenced by every other node

    One node with a stable identifier carrying the identifiers a knowledge graph is built from. In B2B the useful additions are the legal name, the registration identifier and the sameAs profiles on the review platforms, which is how an engine connects your site to your reviews.

    Organization.jsonld
    {
  "@context": "https://schema.org",
  "@type": "Organization",
  "@id": "https://[YOUR-DOMAIN]/#organization",
  "name": "[COMPANY NAME]",
  "legalName": "[REGISTERED LEGAL NAME]",
  "url": "https://[YOUR-DOMAIN]",
  "logo": "https://[YOUR-DOMAIN]/[LOGO.PNG]",
  "foundingDate": "[YYYY]",
  "numberOfEmployees": {
    "@type": "QuantitativeValue",
    "value": "[NUMBER]"
  },
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "[STREET]",
    "addressLocality": "[CITY]",
    "addressRegion": "[STATE OR REGION]",
    "postalCode": "[POSTCODE]",
    "addressCountry": "[COUNTRY CODE]"
  },
  "contactPoint": [
    {
      "@type": "ContactPoint",
      "contactType": "sales",
      "email": "[SALES EMAIL]",
      "areaServed": "[REGIONS YOU ACTUALLY SELL IN]"
    },
    {
      "@type": "ContactPoint",
      "contactType": "technical support",
      "email": "[SUPPORT EMAIL]"
    }
  ],
  "sameAs": [
    "[LINKEDIN COMPANY URL]",
    "[G2 OR CLUTCH PROFILE URL]",
    "[CRUNCHBASE URL]"
  ]
}
  • Service

    Each capability or offering page, where you sell a service rather than a product

    Use it where the thing being sold is genuinely a service with a defined scope and audience. Keep areaServed honest: listing every country you would take a call from is the commonest way this node stops being true.

    Service.jsonld
    {
  "@context": "https://schema.org",
  "@type": "Service",
  "name": "[SERVICE NAME]",
  "url": "https://[YOUR-DOMAIN]/services/[SERVICE-SLUG]",
  "description": "[TWO SENTENCES ON WHAT IT IS AND WHO IT IS FOR]",
  "provider": { "@id": "https://[YOUR-DOMAIN]/#organization" },
  "serviceType": "[CATEGORY IN THE WORDS BUYERS USE]",
  "areaServed": {
    "@type": "AdministrativeArea",
    "name": "[REGION YOU ACTUALLY DELIVER IN]"
  },
  "audience": {
    "@type": "BusinessAudience",
    "name": "[WHO THIS IS FOR]",
    "numberOfEmployees": {
      "@type": "QuantitativeValue",
      "minValue": "[MIN]",
      "maxValue": "[MAX]"
    }
  },
  "offers": {
    "@type": "Offer",
    "priceCurrency": "[CURRENCY]",
    "priceSpecification": {
      "@type": "PriceSpecification",
      "minPrice": "[MIN]",
      "maxPrice": "[MAX]",
      "description": "[WHAT DRIVES THE PRICE]"
    }
  }
}
  • Article with a named author

    Problem-level and evaluation content

    The author node is the point of this one. B2B content is routinely published under a company byline, which throws away the only expertise signal available. Point the author at a real person with a real profile page and real credentials in the buyer's function.

    Article with a named author.jsonld
    {
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "[THE PAGE TITLE]",
  "url": "https://[YOUR-DOMAIN]/[PATH]",
  "datePublished": "[YYYY-MM-DD]",
  "dateModified": "[YYYY-MM-DD]",
  "author": {
    "@type": "Person",
    "name": "[AUTHOR NAME]",
    "url": "https://[YOUR-DOMAIN]/people/[AUTHOR-SLUG]",
    "jobTitle": "[ROLE]",
    "worksFor": { "@id": "https://[YOUR-DOMAIN]/#organization" },
    "sameAs": ["[LINKEDIN PROFILE URL]"]
  },
  "publisher": { "@id": "https://[YOUR-DOMAIN]/#organization" },
  "about": "[THE TOPIC, NOT YOUR PRODUCT]",
  "audience": {
    "@type": "BusinessAudience",
    "name": "[THE ROLE THIS IS WRITTEN FOR]"
  }
}
  • FAQPage

    The trust page, the pricing page and the implementation page

    One node per page and only for questions genuinely visible on it. These three pages are where the questions are real, short and asked in exactly this form, which is not true of the FAQ block somebody added to the home page.

    FAQPage.jsonld
    {
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "Where is our data stored?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "[REGIONS, AND WHETHER THE CUSTOMER CAN CHOOSE]"
      }
    },
    {
      "@type": "Question",
      "name": "What certifications do you hold?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "[CERTIFICATIONS, WITH DATES AND HOW TO REQUEST THE REPORT]"
      }
    },
    {
      "@type": "Question",
      "name": "How long does implementation take?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "[A REAL RANGE, AND WHAT MAKES IT THE LONG END]"
      }
    }
  ]
}
  • Review and testimonial markup, used carefully

    Customer stories, and only where a real rating exists

    Emit a Review node only where an identifiable customer gave a rating you can evidence. Never generate an aggregateRating from testimonials with no scores behind them. The reviews rule of October 2024 and the search guidelines point the same way here.

    Review and testimonial markup, used carefully.jsonld
    {
  "@context": "https://schema.org",
  "@type": "Review",
  "itemReviewed": { "@id": "https://[YOUR-DOMAIN]/#organization" },
  "author": {
    "@type": "Person",
    "name": "[REVIEWER NAME]",
    "jobTitle": "[ROLE]",
    "worksFor": {
      "@type": "Organization",
      "name": "[CUSTOMER COMPANY]"
    }
  },
  "datePublished": "[YYYY-MM-DD]",
  "reviewBody": "[THE QUOTE, AS APPROVED BY THE CUSTOMER]",
  "reviewRating": {
    "@type": "Rating",
    "ratingValue": "[ONLY IF A REAL RATING EXISTS]",
    "bestRating": "5"
  }
}
  • BreadcrumbList

    Solution, capability and resource pages

    Dull and worth having, because B2B sites develop a real hierarchy across solutions, industries and resources, and the URL structure usually stopped reflecting it two redesigns ago. It also makes the segment visible in the result.

    BreadcrumbList.jsonld
    {
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    {
      "@type": "ListItem",
      "position": 1,
      "name": "Home",
      "item": "https://[YOUR-DOMAIN]/"
    },
    {
      "@type": "ListItem",
      "position": 2,
      "name": "[SECTION, E.G. Solutions]",
      "item": "https://[YOUR-DOMAIN]/[SECTION-SLUG]"
    },
    {
      "@type": "ListItem",
      "position": 3,
      "name": "[PAGE NAME]",
      "item": "https://[YOUR-DOMAIN]/[SECTION-SLUG]/[PAGE-SLUG]"
    }
  ]
}

What it costs

The bands below are editorial estimates rather than quotes. They are driven less by company size than by how many committee roles the market has: a two-person buying group and a ten-person one need very different amounts of writing.

Lean

$2,000 to $5,000
  • The trust, pricing and implementation pages, which is the work that shortens cycles already running
  • Two comparison pages, written honestly
  • Ungating the two highest-value assets and republishing them as pages
  • Review platform profiles claimed and maintained

Who it suits

A company selling to a small committee, or one with a strong founder-led channel already working

Where it stops

It does not fund the problem-level content that reaches a practitioner before a category exists, which is where the compounding actually happens.

Funded

$5,000 to $12,000
  • Coverage across every committee role rather than the three easiest
  • Named customer stories with checkable detail, which take time rather than money
  • Documentation opened and treated as content
  • Reporting that segments by role and states the attribution gap rather than hiding it

Who it suits

A company with a defined category, a real sales team and a cycle measured in months

Where it stops

It cannot create a category. If the buyer does not know the problem has a name, that is a demand creation programme and search is downstream of it.

Enterprise

$12,000 to $20,000 and above
  • The full role map maintained as the product and the market change
  • Segment depth by industry, size and region, with a real answer in each
  • The analyst relationship run properly, which is not content work but sits beside it
  • Attribution modelling that a revenue leader will actually defend in a board meeting

Who it suits

A multi-product company with several buying committees and a long cycle

Where it stops

None of it shortens a procurement process that takes two quarters by policy. Content removes questions from a cycle; it does not remove the cycle.

How to do it with no budget

More of this is free than people expect, because the highest-value material already exists inside the company. It is in sales calls, in support tickets and in the documents somebody sends manually three times a week.

  1. 1

    Read the last twenty lost deals and write down each killing objection

    A day

    Your own CRM

    This produces a better content plan than any keyword tool, because every item on it is a real reason somebody did not buy.

  2. 2

    Ask sales which PDF they send most often, and publish it as a page

    Half a day per asset

    A conversation, then your CMS

    If it is being sent manually every week, it is answering a real question and it is currently invisible to everyone who did not ask.

  3. 3

    Claim the review platform profiles in your category

    2 hours each

    The platforms themselves

    Free, and these pages frequently rank above yours for your own category. Fix the description and the screenshots at minimum.

  4. 4

    Read what your competitors' comparison pages say about you

    2 hours

    A search for your own brand plus vs

    Usually inaccurate and usually ranking. Correcting the record on your own page is cheaper than complaining about theirs.

  5. 5

    Write the security page from the questionnaire you already fill in

    A day

    The last three security reviews you completed

    The answers are already written and already approved. Publishing them removes a fortnight from cycles that have not started yet.

  6. 6

    Segment Search Console by committee role

    2 hours

    Google Search Console

    Filter for pricing, security, implementation and comparison language separately. The gap between those segments is the finding.

  7. 7

    Ask three customers whether they will be named

    An hour, plus waiting

    Email

    Anonymised case studies are usually anonymous because nobody asked twice. Offer approval over the text and the answer changes.

The tool stack

The tooling here is unusual in that rank tracking matters less than in most industries and answer-engine monitoring matters more. The queries are few and specific, so precision on a short list beats coverage of a long one.

  • Track a short list of very high-value queries precisely

    AccuRankerSEO APIRead the review

    The list that matters here is dozens of queries, not thousands, and each one is attached to a real objection. Daily accuracy on forty terms is worth more than weekly estimates on four thousand.

    Free routeSearch Console position data, which is averaged and lagged but free

  • See which comparison pages competitors already rank for

    AhrefsSEO APIRead the review

    Use it to find who is publishing about you, not to build a keyword list by volume. The useful export here is competitor pages containing your brand name.

    Free routeSearching your own brand plus vs, which finds the important ones anyway

  • Check whether you appear in AI-generated answers

    ProfoundSEO APIRead the review

    Increasingly the shortlist is assembled by an assistant summarising review platforms and forums. Worth monitoring specifically, and worth knowing which sources the summary is drawing on.

    Free routeAsking the assistants your buyers use, on a schedule, and keeping the answers

  • Publish documentation and pricing outside the login wall

    SanityHeadless CMSRead the review

    The requirement is that documentation and marketing share a domain and a search index. The failure mode is docs on a separate host that accumulates its own reputation and passes none of it on.

    Free routeA static site generator, which is what most documentation already runs on

  • Join site behaviour to opportunities rather than to sessions

    GA4 Data APISEO APIRead the review

    Pull the data out rather than reading dashboards, and join it to the CRM on the account rather than the person. One person is not the buyer, so a person-level report answers the wrong question.

    Free routeIncluded with the analytics you already run

  • Model what a committee-length cycle has to return

    Marketing Funnel CalculatorFree toolOpen the tool

    Useful mainly to make the cycle length explicit before anybody signs off a budget. Put your own close rate and cycle length in it rather than a benchmark from a report about somebody else's market.

    Free routeIt is free

Take it from here

Everything below is yours to take. Fill the [BRACKETS] and it is ready to use. Start with the lost-deal audit, because it produces the content plan that every other item on this page assumes you already have.

Checklist

The pass over closed-lost opportunities that turns real objections into a content plan, replacing a keyword export sorted by volume.

LOST-DEAL CONTENT AUDIT
Company: [COMPANY NAME]
Run by: [NAME]     Date: [YYYY-MM-DD]     Period covered: [FROM] to [TO]

WHY THIS REPLACES A KEYWORD EXPORT
A keyword tool tells you what many people search. It cannot tell you what
ended a deal worth six figures. This does, and every line it produces is
attached to money that did not arrive.

PULL THE DEALS
[ ] Last 20 closed-lost opportunities above [DEAL SIZE]
[ ] Include the ones lost to "no decision", which are usually the most
    informative and are usually skipped

FOR EACH DEAL, ONE BLOCK

  Deal: [ACCOUNT]              Value: [AMOUNT]     Closed: [YYYY-MM]
  Lost to: [COMPETITOR / NO DECISION / BUILT IT / BUDGET]

  Who was involved, by role:
    [ ] Practitioner       [NAME OR "unknown"]
    [ ] Champion           [NAME]
    [ ] Evaluator          [NAME]
    [ ] Security or IT     [NAME OR "never reached"]
    [ ] Finance/procurement[NAME OR "never reached"]
    [ ] Future owner       [NAME OR "never identified"]

  The objection that actually killed it:
    [ONE SENTENCE, IN THE CUSTOMER'S WORDS IF YOU HAVE THEM]

  Which role raised it: [ROLE]
  At what stage: [STAGE]
  Do we have a public page answering it? [YES / NO / PARTLY]
  If yes, did anyone at the account read it? [YES / NO / UNKNOWN]

ROLL IT UP

  Objection                          Times raised    Public page exists
  [OBJECTION]                            [N]              [Y/N]
  [OBJECTION]                            [N]              [Y/N]
  [OBJECTION]                            [N]              [Y/N]
  [OBJECTION]                            [N]              [Y/N]

  Roles we never reached in a lost deal: [LIST]
  <- This list is the content plan. Start at the top.

THE QUERY TRANSLATION
For each of the top five objections, write the search somebody would run.

  Objection: [OBJECTION]
  They would search: "[QUERY]"
  Who ranks for it today: [WHO]
  Asset that answers it: [PAGE TYPE]
  Owner: [NAME]      Due: [YYYY-MM-DD]

THE RULE
Do not check search volume until this table is complete. Checking it
first is how the plan ends up aimed at people without budget.

What to publish

The test for everything below is whether it survives being forwarded. B2B content is read on a screen the author never sees, pasted into a document by somebody building a case for a person who will never visit your site.

  • The trust page

    Once, then reviewed whenever a certification or a sub-processor changes

    Security, certifications, sub-processors, data residency and a named contact, in public. It answers the fastest veto in the process and almost every competitor keeps it behind an NDA.

  • A real pricing page

    Once, reviewed each time the commercial model changes

    The query runs whether or not you answer it. Publishing the model and the variables beats leaving a review site or a forum to publish a number on your behalf.

  • The honest comparison page

    One per serious competitor, revised twice a year

    Your competitors already publish one about you. A version that concedes two things they do better is the only kind that gets believed, and it is the one a champion is willing to forward.

  • The implementation page

    Once, updated when the delivery model changes

    It answers the quietest veto: the person who will inherit the work asking whether this becomes their problem. Real durations, real staffing and the failure modes.

  • Named customer stories with checkable numbers

    One a quarter, which is realistic given approvals

    An identifiable company, an identifiable person and a number with the method attached. Everything else is a logo wall, which proves you had a customer once.

  • Problem-level content for the practitioner

    Continuous, and judged over years rather than quarters

    It reaches somebody before they know the category exists, converts to nothing measurable, and is the reason you are already trusted when the evaluation starts.

  • Business-case material built to be forwarded

    One per primary use case

    The only asset whose reader is not the audience. It gets pasted into a deck, so every number needs a source and nothing can depend on the surrounding page.

And what not to

  • Twelve industry pages that differ only by the noun. They read as filler and they are the exact shape search engines treat as scaled content.
  • Ungrounded thought leadership about the future of the category. It reaches peers and competitors, and no member of a buying committee has ever forwarded one.
  • Gated benchmark reports. The gate collects a few hundred addresses and forfeits every reader who would have cited it.
  • Anonymised case studies. "A leading provider in the sector" is not proof and is not content, and it is usually anonymous because nobody asked twice.
  • Chasing the bare category head term with a product page. The review platforms have owned it for years and a vendor page is the wrong result for that query anyway.
A vertical pipeline following one page from the champion who found it, through a Slack message, into a slide, into a business case, to a finance director who never visited the site. Each stage strips away more of the original context.
By the last stage nothing remains but a claim and a number. That is why the number needs a source printed next to it.

The expensive mistakes

Building the keyword plan by volume

Costs you A year spent ranking for queries belonging to people who will never have budget

Build it from lost deals and sales calls, then check volume last as a tiebreaker rather than a filter.

Writing for one persona

Costs you A champion who is convinced and five colleagues who are not

Map the roles on your last five closed deals and write for the four in the middle nobody covers.

Gating everything worth reading

Costs you Invisible to search, unquotable by an answer engine, and read by nobody who dislikes forms

Publish the substance and gate a tool or a session. Measure both sides before deciding it was wrong.

Refusing to publish pricing

Costs you The answer gets published anyway, by somebody who may have it wrong

Publish the model, the variables and a range. It is a commercial decision, so take it with the people who own the number.

Treating the form fill as the start of the process

Costs you A measurement system that reports on the last quarter of the buying journey

Measure coverage of the committee and cycle length, and treat the conversion as a late confirmation.

Claiming a single touch caused an eleven-month deal

Costs you A report nobody senior believes twice

State the model, state the gap, and show which role clusters an account touched before it closed.

Assuming business contact data sits outside privacy law

Costs you A lead database governed by rules the privacy notice does not describe

Check the notice and the rights process against the current position, which changed in California at the start of 2023.

What to measure

Attribution in this industry does not close, and the useful response is to stop trying and report something else. A deal touched by six people over eleven months has no single source, and any model that produces one is making a choice rather than finding a fact.

Leading indicators

Move first. They predict, they do not prove.

  • Impressions by committee role cluster

    Search Console, segmented by pricing, security, implementation and comparison language

    The most useful single view on this page. It shows which parts of the committee you reach and which you are invisible to, which no other metric here does.

  • Named accounts touching more than one role cluster

    Analytics joined to the CRM at account level, where the join exists

    Imperfect and worth reporting anyway. An account that has read the security page and the pricing page is behaving differently from one that read a blog post.

  • Assets published that sales actually sends

    Ask sales, monthly

    A crude measure of whether the content programme is answering real objections. If nothing published this quarter has been sent by a rep, that is the finding.

  • Non-brand impressions on evaluation queries

    Search Console, excluding brand terms

    The slowest indicator here and the one that compounds. Brand demand is created by other channels, so mixing it in makes the content programme look better than it is.

Business indicators

The ones a manager acts on.

  • Pipeline influenced, with the model named on the page

    The CRM, using whatever multi-touch model you have chosen

    Report the model and its limitations in the same breath as the number. A figure presented without its method is the reason marketing reports stop being read.

  • Cycle length, split by whether the account read the trust and pricing pages

    The CRM, joined to page-level analytics

    Not proof of causation and still the most persuasive number available here, because it points at a mechanism a sales leader recognises from their own calls.

  • Win rate against named competitors

    The CRM, where competitor is recorded on the opportunity

    Worth reporting beside the comparison pages, since that is the specific thing those pages are supposed to move. Requires sales to record the competitor, which is the hard part.

  • Closed revenue from accounts with organic first touch

    The CRM, first-touch attribution

    Report it with the untracked share stated. In a cycle this long a meaningful proportion of first touches were never recorded at all, and saying so is what keeps the rest credible.

A before-and-after pair. On one side, eleven months of reading, forwarding, forum threads and a security review. On the other, the single row the CRM stored: one source, one date, one contact.
The right-hand side is not wrong so much as narrow. Reporting it as the cause is what makes a marketing report stop being believed.

The verdict

What makes this hard is not competition. It is that the buyer is plural, so a page that convinces one person can still lose the deal to a colleague who never read it and asked a different question.

The work follows from that directly. Cover the roles nobody writes for, publish the answers currently trapped in sales calls, and stop building keyword plans from a column of search volumes.

When you assess B2B SEO services, listen for what they ask about first. If the opening question is about keyword volume rather than about your last twenty lost deals, they are about to build the wrong plan very competently.

The other test is what a provider of SEO services for B2B says about attribution. Anyone promising a clean line from a search to a signature in an eleven-month cycle is selling a report rather than a result.

FAQ

B2B SEO questions

  • Should we really ignore search volume?
    Not ignore it, demote it. Build the plan from real objections and real deals, then use volume to choose between two equally valuable options. Leading with volume reliably produces a plan aimed at people without budget.
  • Is it worth publishing pricing when nobody in our market does?
    That is usually the argument for doing it. The query runs regardless and currently lands on a competitor or a forum. If a list price genuinely does not exist, publish the variables and a range instead of silence.
  • How do we justify content that produces no leads?
    By measuring the right thing. Problem-level content reaches a practitioner before a category exists and is why you are trusted later. Report it as reach and impressions, and never let it be judged on form fills.
  • Do review sites matter if we do not sell software?
    Yes, though different ones. Clutch for services, ThomasNet in industrial supply, and the trade publications your customers actually read. Find out which by asking customers rather than by checking domain metrics.
  • How long is this going to take?
    Nine to eighteen months for the compounding work, but the trust, pricing and implementation pages affect cycles already running. That is the part worth doing first, because it shows a return inside a quarter.
  • Should we ungate our best report?
    Test it with one asset before deciding. Publish the substance as a page and gate a template or a session instead, then measure both the addresses collected and the reach gained. The argument is usually settled by the numbers.
  • Are the budget figures here quotes?
    No. They are editorial estimates, and the bands are driven more by how many people sit on your buying committee than by the size of your company. Treat them as a shape rather than a price.
  • What is the one thing to do first?
    Read your last twenty lost deals and write down the objection that killed each one. That list is a better content plan than any tool will give you, and it takes a day.

Run it yourself, or have someone own it

Everything above is written to be run without us, and the free path is genuinely most of the value for a single-location business. Where these plans stall is almost never the plan. It is that nobody owns it after the first month. That is the job we do, with search as one distribution layer inside a wider system rather than the whole engagement.

See how that worksAll industries