Beyond Rankings: How Small Businesses Build Search Authority Across Search and AI

A Ranking Is Useful Evidence, Not the Business Outcome

For most of my career in digital work, rankings have been one of the clearest signs that something has changed. If an important service page moves from page four of Google to the first page for a commercially relevant search, that is worth paying attention to because the business has become easier to find at a moment when people are actively looking for what it does. I would want to know that happened, and I would want to understand what changed around it.

The ranking cannot tell me what happens after the page is found. A company can appear prominently and still have weak positioning, a confusing website or content that answers a search query without giving somebody much reason to trust the firm behind it. Another business may attract less traffic overall while reaching people whose needs are much closer to the work it actually wants to win.

I have become more careful about treating visibility as a business result. Rankings, impressions and clicks tell us useful things about discovery, but they do not tell us whether somebody understood the business, whether the page gave them enough confidence to continue or whether an eventual enquiry was commercially relevant. Search visibility and business performance are connected, but they are not interchangeable.

The environment has also widened. Google’s newer AI-assisted Search experiences still rely on many of the same underlying requirements that mattered before, including accessible pages, crawlable information and sound technical implementation, while people can now encounter the same company through traditional search, Maps, AI-assisted answers, review sites, social platforms or a recommendation from somewhere else entirely. I still care where an important page ranks; I simply want that ranking interpreted as one useful observation inside a much larger discovery environment.

Search Authority Is Not a Google Score

I use the term search authority carefully because the digital industry has spent years turning complicated conditions into scores that look more precise than the underlying evidence allows. When I use the phrase here, I am not referring to a published Google metric or to a universal authority score provided by OpenAI or Microsoft. Third-party tools can model links, visibility, competitive position and other signals, but those models are still models.

When I refer to search authority at Fervid, I mean something broader and more practical. I am describing the condition that develops when enough of a company’s public digital presence helps people and relevant discovery systems understand what the company knows, what it does, where its expertise applies and why particular claims deserve consideration. Useful first-party information contributes to that condition, as do technical accessibility, identifiable expertise, credible external references, consistent business information and a website that remains useful after somebody arrives.

The same caution applies to E-E-A-T. Experience, expertise, authoritativeness and trustworthiness can be useful lenses for evaluating content, but Google has been clear that E-E-A-T is not a single ranking factor or score. I would rather examine the public reality around the company: who is writing expert material, whether important claims have support, whether the site demonstrates real understanding and whether external sources help a prospective client verify parts of the story.

Those questions require judgement, which is exactly why they are useful. They point toward things the company can genuinely improve and give us a better basis for deciding what deserves attention next, without pretending that one score can explain how a business will be treated across every search and AI system.

Small Businesses Can Compete Through Specificity, Not Publishing Volume

One assumption I have seen survive through several generations of digital marketing is that a smaller company has to publish at something close to the pace of a much larger competitor if it wants to compete in search. That logic is understandable when the larger organisation has hundreds of pages, a full content team and enough resources to keep publishing while the smaller firm is still trying to decide who has time to review the next article. I do not think most small businesses should accept that as the contest they have to win.

What smaller companies often have is much closer access to the work itself. The owner, adviser, lawyer, broker, contractor or specialist may spend every week dealing with the situations the website is supposed to explain, including the cases that do not fit neatly into the standard answer. They see which questions keep returning, where customers misunderstand the process and which apparently minor details change the recommendation once the work begins.

That is something I see in real client work. On the site for Tim Louis & Company, somebody dealing with a denied long-term disability claim needs a different kind of explanation from a person trying to understand probate or a workplace problem. With Bill Craven at Craven Financial Planning, business-owner retirement, farm continuity and insurance questions sit inside an Ontario planning context. Those are not generic content buckets; the useful material comes from the distinctions inside the work and from knowing which parts of those distinctions a reader actually needs.

Keyword research still belongs in the process because it can show demand, language and existing coverage. It cannot tell us everything the organisation knows, and an editorial programme built only from exported queries can miss the knowledge clients hear in consultations every week. The better conversation often begins when somebody says, ‘We explain this all the time,’ because that is where ordinary business experience can turn into information that competitors have not represented very well.

Not all of that knowledge belongs online. Some of it may be confidential, regulated, dependent on individual circumstances or simply too narrow to help the people the company is trying to reach. The point is not to publish everything; it is to identify the parts that can genuinely help somebody understand a problem or make a better-informed decision.

Build Information Worth Retrieving

A company may have valuable knowledge inside it without automatically having material that deserves to become a page. Once we understand what the people closest to the work actually know, another question follows: which parts of that knowledge will genuinely help somebody understand a problem or make a better decision? A technically polished page can still add very little if the substance underneath it is ordinary.

The strongest material often sits close to a decision. A prospective client may need to understand whether a service applies to their situation, why two apparently similar options lead to different outcomes or what happens when the standard process stops working. Someone who does the work regularly may know which facts really change the answer and which details only look important at first.

I would not confuse that with manufacturing originality. A company does not need a contrarian position on every subject, and sometimes the better contribution is simply a clearer explanation of something established. A familiar topic can still deserve a strong article when the existing material is vague, overly technical or disconnected from the decision somebody actually has to make.

Polished writing is now very easy to produce, which has made the support underneath the writing more important to me. A sentence can sound authoritative long before the material beneath it deserves that confidence, and fluency does not tell us whether a claim is current, whether a cited source actually supports the conclusion or whether an exception changes the answer. The page should preserve enough of those boundaries that the information can be used sensibly.

Depth also has a limit. Pulling every related question, exception and secondary keyword into one page can make the material harder to use, so architecture becomes part of the editorial judgement. Some information belongs on a service page, a narrower question may deserve its own article, and supporting evidence may work better in a reference section than inside the main explanation.

Make Expertise Easier to Connect and Verify

Good information can still sit inside a weak public context. An article may explain a difficult subject well while leaving the reader unsure who wrote it, why that person has a reasonable basis for the judgement or how the expertise connects to the company offering the service. The individual pieces can be credible while the relationships between them remain poorly built.

That does not require a résumé on every page. It does require enough connection that the reader is not forced to invent the relationship.

If somebody is reading about a complicated legal, financial, technical or business problem, a useful biography and sensible authorship should make the connection between the person, the experience and the subject reasonably clear. Company claims deserve similar judgement. Some facts are best established by the company itself, while an award, designation or membership may be easier to assess when an independent source can confirm it. Ordinary company copy does not need to become citation-heavy, but significant claims should have support when a reasonable reader would expect to see it.

I am also wary of one strong page sitting inside a site that tells a different story everywhere else. Websites accumulate, services change and older sections can fall out of step with current expertise, which means a legitimate claim can look less settled than it really is. A reader moving from article to biography to service page should not have to reconstruct the business for themselves. Structured data and other machine-readable information can help describe those relationships, but the technical layer comes after the visible one for me; I would rather repair the public information first and then make the machine-readable representation match it.

Technical SEO Still Carries More Weight Than the Hype Suggests

I have worked through enough changes in search to become cautious whenever a new development is supposed to make the fundamentals obsolete. Mobile, machine learning, featured answers and now generative search have all changed the environment in meaningful ways, yet when I get into a real website I still find familiar problems deciding whether good information has a reasonable chance of being found and used.

A service page can contain excellent material and still be buried behind poor navigation. An article can point to the wrong canonical URL, compete with an older version of itself or sit several levels deep with almost no useful internal links. I have also seen pages that look completely normal in a browser while the underlying implementation prevents important information from being handled the way the team assumes it is.

The audit score is not the strategy. Two warnings in the same report can have entirely different consequences for the company.

A missing metadata field and an important page accidentally excluded from indexing may both be labelled as technical problems, but I would never give them equal urgency simply because software placed them in the same list. Real sites also carry history: old templates, plugins, redirects, tracking scripts, forms, CRM connections and campaign pages accumulate over years. An apparently simple change may therefore affect paid advertising, a sales workflow or a measurement event somebody has relied on for a long time, which means correcting the technical issue is only part of the job.

Structured data and emerging conventions deserve the same restraint. Schema should describe something meaningful that already exists, and Google’s position on llms.txt is a useful reminder that a new convention does not automatically become a universal requirement.

For most small businesses, the technical foundation remains practical. Important pages should be reachable, the right versions should be indexable, internal links should make sense, forms should work and measurement should be reliable enough to distinguish meaningful activity from noise. The sophistication lies less in producing a longer checklist than in understanding which technical problem is actually getting in the way of the business.

For a Local Business, Authority Extends Beyond the Website

A company that depends on a geographic market has a different problem from one that can serve almost anybody from anywhere. Someone searching for a lawyer in Vancouver, a financial planner in Chatham or a contractor in a particular neighbourhood is bringing location into the decision as well as the service. The company has to make sense in the subject and in the place.

Google’s framework is relatively plain: local results are influenced by relevance, distance and prominence. I like having distance in that model because it makes one limit obvious; geography is not something a company can rewrite its way around.

The site still gives the organisation the greatest control over its local story. A service page can explain the work, an About page can establish who is responsible for it, and genuinely local material can show how provincial rules, market conditions or regional practices change an answer. That is very different from producing thin city pages whose strongest local connection is the place name in the heading. The Google Business Profile sits beside that website, not outside the strategy, because somebody may encounter the address, services, hours, photos and reviews before they ever click through.

Google can also draw local information from the company, its website, third-party sources and user contributions, so an outdated description elsewhere can continue creating confusion after the site itself has been corrected. Reviews are especially useful because they are not written in the company’s marketing language; over time they can reveal what customers actually remember about the experience. Professional associations, local media, chambers, community involvement and directories people genuinely use can add further context, but I still want to know what each source actually establishes.

I trust a smaller number of relevant local references more than a manufactured footprint assembled for a report. The public record does not need identical wording everywhere, but it should make sense when somebody moves through it.

AI-Assisted Discovery Adds New Surfaces, Not New Certainty

I have become increasingly interested in AI-assisted discovery because it changes where a company can be encountered and how much information somebody may receive before reaching the website. A person can ask a detailed question, receive a synthesized answer, inspect one or two sources and arrive at a company with considerably more context than they would have had after scanning a conventional results page. That deserves serious attention without requiring us to pretend every AI-assisted system behaves the same way.

The available research already suggests meaningful differences between systems. A 2026 study in Findings of ACL compared traditional Google results with generative search systems from Google, OpenAI and Perplexity and found variation in source diversity, retrieval behaviour and stability across repeated runs. I would not use one study to predict every future system, but it is enough to make me skeptical of universal advice about ‘ranking in AI.’

A conventional ranking and an AI citation are different observations. If a page appears fifth in a search result, I can see its position inside an ordered interface; if the same page appears as a cited source in a generated response, I know it was surfaced in that experience without knowing whether it will appear for another user or another version of the question. Microsoft makes a similar boundary explicit in its AI reporting by warning that citation counts are not rankings or authority scores.

OpenAI exposes a different part of the environment. Allowing OAI-SearchBot makes a public site eligible for inclusion in ChatGPT Search, but OpenAI says placement is not guaranteed. Google, meanwhile, continues to connect AI Overviews and AI Mode to its existing Search infrastructure and says publishers do not need special AI markup or a separate writing style simply to participate.

Those differences are enough for me to resist advice that begins with ‘optimize for AI’ as though the phrase describes one technical environment. We can make important information accessible, improve the source, clarify expertise, support consequential claims and observe provider-specific referrals or citations when they are available. What we cannot do is select the source on behalf of the external system, and I would rather preserve that uncertainty than replace it with a theory simply because the theory makes the work easier to sell.

The same caution applies to AI-visibility scores. A score can help organize repeated observations, but I want to know what was tested, which systems and prompts were included, whether geography changed and how repeatable the result appears before deciding what the number means. A higher score may reflect a meaningful change, or it may reflect a different prompt set, a provider update or a temporary retrieval pattern.

I expect AI-assisted discovery to become more measurable and more interactive as provider reporting and browser agents develop. There will be legitimate technical work around those changes, but an emerging protocol should still compete for priority against problems we can already see. If the service pages are weak, the evidence is thin or the forms are broken, preparation for a future agent interaction should not automatically outrank fixing the website people are using today.

Measure the Chain, Not One Number

I have always liked numbers, but I do not trust a number simply because a platform can place it on a dashboard. A ranking, an impression, a citation, a session and an enquiry are measurements of different events, and trouble begins when one of them is quietly asked to explain the others. A ranking can improve while traffic barely changes, and traffic can rise without bringing the business better opportunities.

The journey is not especially difficult to describe, even if it is difficult to attribute perfectly. A company becomes visible somewhere, somebody encounters it, some people reach the website, some engage with what they find and a smaller number eventually do something with commercial meaning. The route, the timing and the parts we can observe will differ from one business to another.

Provider data is becoming more useful, but the observations still need to retain their original meaning. Google now provides dedicated Search Console reporting for impressions and page visibility inside its generative AI Search experiences, Microsoft can show citation activity across supported AI experiences, and ChatGPT referral data can tell us that somebody actually reached the website through ChatGPT. I would use all three when they are relevant without blending them into a single invented authority score.

The website becomes an important meeting point because several discovery routes eventually arrive there. Analytics can show whether somebody reached an important service page, opened a contact path or completed an action we have decided deserves to be measured. A useful event has to earn its place in the reporting, because a video play, a phone click and a qualified consultation do not carry the same commercial weight simply because software can count all three.

The same care is needed after an enquiry arrives. Ten form submissions can look better than six until the client tells us that eight of the ten were poor fits while four of the earlier six became genuine opportunities. Some of the most useful reporting information therefore comes from owners, advisers and sales teams, and I would not disregard that operational knowledge simply because it arrives in a conversation rather than a dashboard.

Attribution becomes even less tidy when several substantial changes happen together. If traffic rises after a rebuild that also included new content, technical repairs and stronger internal linking, the timing is useful without proving that one intervention caused the whole result. I can describe what changed, examine which pages and queries moved and look for evidence that strengthens one explanation over another, but I do not want to assign the result to whichever part of the work is easiest to promote.

Sometimes measurement tells us where to investigate before it tells us why something happened. If visibility improves while engagement deteriorates, the audience or landing experience may deserve attention; if engagement is strong while meaningful actions remain weak, the constraint may have moved further down the experience. I want the reporting to follow the business rather than make the business fit the reporting.

A Practical Starting Point for a Small Business

If I were looking at a small-business website for the first time, I would not begin by counting missing articles or listing every keyword the company does not rank for. I would want to understand what the website is supposed to accomplish, which work matters most commercially, who the customer is and what a good enquiry actually looks like. Those questions often tell me more about priority than the first audit score I see.

I listen carefully to the owner’s or marketing team’s view without assuming the first explanation is necessarily the right one. A company may believe it needs more traffic when the pages already receiving traffic are doing a poor job of moving people forward, while another may think rankings are weak when its most valuable services are barely represented on the site. I have also seen dated-looking websites continue to produce strong enquiries, which is a useful reminder that visible imperfection is not automatically the most important problem.

Then I try to look at the public version of the company with fresh eyes. That sounds simple, but prior knowledge can hide an extraordinary amount of confusion.

Can somebody tell what the company does, who it helps and which services matter most, and can they see the people behind the expertise? In work with clients such as Tim Louis & Company or Bill Craven at Craven Financial Planning, that question quickly becomes specific. A legal site needs to represent the real differences between disability, employment and estate problems, while a financial-planning site has to carry the realities of retirement, business ownership, farm families and Ontario planning. Calling all of that ‘content’ hides the judgement involved.

Search research, technical evidence, analytics, local visibility and AI-assisted discovery can all sharpen that picture, but none should define it in advance. Keyword data can show demand, a crawl can expose indexing problems and analytics can show where people arrive or leave; meanwhile, a form can be technically functional while sending enquiries somewhere nobody monitors, and an event can fire perfectly while measuring an action nobody cares about. Newer AI opportunities belong in the review too, but they still have to compete for priority against defects we can already see and explain.

By that point, one company may genuinely need more content while another needs fewer pages and a better representation of the services it already offers. I would rather arrive at that conclusion after looking across the business, the site, the search environment and the available evidence than begin with a package of tactics that assumes the diagnosis in advance. Having more capabilities available should give us more possible ways to solve the problem; it should not create pressure to use every capability simply because we have it.

The recommendation then needs to remain open to challenge. I want the client to understand what appears to be limiting the current situation, what supports that judgement, what we intend to change and what we would expect to observe if the diagnosis is reasonably accurate. Implementation can expose something we missed, and better information should be allowed to change the plan.

Authority Accumulates When the Business Keeps Earning It

I have watched search change enough times that I am cautious about building a long-term strategy around whatever happens to look newest at the moment. Some changes genuinely reshape discovery, and generative search is doing that now, but the companies that remain credible through those shifts usually have more holding the digital presence together than one successful tactic. They know what they do, useful knowledge exists around the work and there are identifiable people who can stand behind it.

I would not confuse that condition with protection from search volatility. A competitor can publish something better, Google can change the way results are presented, an AI system can surface another source tomorrow and a local result can change because the searcher is standing somewhere else. What can accumulate is the quality of the public evidence around the company, not control over the systems evaluating it.

That accumulation is usually less dramatic than people expect. One page gets rebuilt properly, an old description is corrected, a customer question becomes an article worth keeping and a technical problem that has been interfering with discovery is finally removed. Enough of those changes can alter the way the company is understood without any one of them deserving to be called a transformation.

The work also has to be maintained because the company itself keeps changing. Services evolve, people develop new expertise and older information can become less accurate without anybody deliberately publishing something false. Articles need another look when the underlying guidance changes, biographies should move with responsibilities and service pages should not preserve work the company no longer wants.

I apply the same restraint to Fervid’s own systems. A framework, tool or proprietary capability earns its place when it helps us see the problem more clearly or perform the work better, and building something sophisticated does not make it necessary in every engagement. Earlier in my career there was always another optimization available, another page that could be built or another metric that could be improved; I am more interested now in whether the next piece of work deserves to exist.

I still care when an important page moves upward, when traffic improves, when a source is cited or when better enquiries arrive. I simply do not want one favourable number to substitute for understanding what is changing across the company. A small firm may never have the publishing resources or name recognition of a national competitor, but it can become unusually clear about the work it does and much easier for the right people to evaluate.

I want Fervid held to the same standard. Our work should remain open to inspection, claims should stay within the available support and our methods should be allowed to change when better information gives us a reason to change them. I do not think authority is something a company finishes building; the digital presence should be allowed to change with the business rather than becoming a permanent record of what the company used to be.