What business should I start if I'm good at SEO?

Is SEO still a strong business advantage today?

Yes. Being good at SEO is still a serious business advantage today, but it works much better when search brings you customers rather than readers.

Google Search itself remains enormous. Alphabet reported more than $63 billion in Search and other advertising revenue in the final quarter of 2025, up 17% year over year. Google also says AI is pushing people to search more: AI Mode has passed one billion monthly users, its query volume has been more than doubling every quarter, and Google said overall Search queries recently reached an all-time high.

The problem sits one step later in the funnel. Ranking no longer guarantees the click it once did.

Pew Research tracked the browsing behavior of 900 U.S. adults and found that people clicked a traditional Google result on 15% of searches without an AI summary, against only 8% when an AI summary appeared. Clicking a source inside the AI summary happened just 1% of the time.

More recent data suggests the pressure is still increasing. Ahrefs analyzed Google Search Console data from more than 400,000 websites and estimates that an AI Overview now cuts the click-through rate of the number-one organic result by roughly 58% on affected informational searches.

France provided an unusually clean test when AI Overviews arrived there in 2026. Ahrefs followed 963 domains before and after the rollout. Websites with more than 20% of their queries exposed to AI Overviews saw median CTR fall 23.1% in nine days, while minimally exposed sites actually gained 2.3%. Rankings generally remained intact; clicks disappeared.

SEO still creates leverage. We simply need to attach that leverage to searches where somebody still has something to do after getting the answer.

What kind of SEO traffic is losing value fastest?

Pure informational SEO is losing value fastest because Google and AI assistants can increasingly satisfy those searches without sending anyone to a website.

Pew found that AI summaries appeared on 18% of the Google searches it observed, but exposure varied dramatically by query type. Only 8% of one- or two-word searches produced an AI summary. For searches containing ten words or more, the figure reached 53%. Searches beginning with question words such as “what,” “why,” “who” and “when” triggered one 60% of the time.

Ahrefs' much larger keyword datasets show the same pattern. AI Overviews appear overwhelmingly on informational searches, while question-style queries are among the most exposed.

That hits an old SEO playbook directly. Publishers spent years expanding into thousands of long-tail questions because these searches were relatively easy to rank for and collectively produced huge amounts of traffic. AI is unusually good at absorbing exactly those questions.

The French rollout makes the distinction even clearer. Health websites in Ahrefs' sample had the highest AI Overview exposure at 22.2% and lost 20.4% of their CTR. A medical question often ends once the searcher has received a short explanation.

Compare that with “emergency plumber near me,” “warehouse racking installer Manchester,” “best payroll software for restaurants” or “immigration lawyer Miami.” Google can summarize the options, but somebody still needs to contact, compare, book or buy from a company.

That is where we should be looking for businesses.

Search intent	Current SEO opportunity	Why
Simple informational question	Weakening quickly	AI can often finish the job inside the search result
Complex research	Mixed	Users may still need sources, data or expert analysis
Product comparison	Still valuable but under pressure	A purchase still follows, although AI can compress research
Local service	Strong	The searcher still needs a real provider
High-value B2B service	Strong	Evaluation, pricing and contact happen outside the answer
Tool, calculator or database	Strong	The user needs functionality rather than a paragraph

Should I start an SEO agency if I'm already good at SEO?

A specialized SEO agency is still one of the easiest businesses to turn SEO skills into cash, especially when we focus on one industry instead of selling generic SEO to everybody.

The entry cost is tiny. We can sell expertise before hiring employees, building software or buying inventory.

The latest industry pricing studies also show that businesses continue paying meaningful amounts for SEO. A 2026 GoodFirms survey of more than 300 agencies found the most common monthly retainer range was $1,500 to $5,000. Older Ahrefs research across 439 providers found agencies averaging about $3,200 per month, while experienced providers generally charged more.

Eight clients paying $2,500 each would produce $20,000 of monthly revenue. Fifteen would produce $37,500. Getting there can be much faster than waiting for a new website to accumulate authority and rankings.

I would still avoid building another generalist agency.

AI has made audits, keyword clustering, briefs, metadata and ordinary SEO content dramatically cheaper to produce. Clients can also shop among thousands of providers selling roughly the same deliverables. The weakest agencies are competing on a service that is becoming easier to imitate.

A specialist can reuse knowledge. If we work only with dental practices, industrial manufacturers or cybersecurity companies, the second client is easier than the first. By client ten, we already understand the keywords, competitors, conversion paths, link opportunities, typical economics and sales objections of the market.

Why does local SEO look especially attractive now?

Local SEO looks unusually attractive now because AI is changing how people discover businesses without removing the need to choose an actual business.

BrightLocal's 2026 consumer research found that 84% of consumers had searched online for a local business during the previous three months. Google remained the most common starting point, although three quarters of consumers used more than one channel during their latest local search.

AI has entered that journey remarkably fast. BrightLocal found that 45% of consumers had used an AI tool for local business recommendations during the previous year, up from just 6% in the prior survey.

The interesting part comes next.

Only 18% of people who used AI for local discovery said they were ready to contact the recommended company immediately. Most went elsewhere to verify the recommendation. Another BrightLocal survey found 88% fact-checked reviews mentioned by AI tools.

A recent BrightLocal study went beyond surveys by analyzing 200,000 actual local AI searches. Business websites represented 93% of all unique domains used as sources and 42% of citations. AI platforms changed which companies they recommended from run to run, but individual businesses still reappeared more than half the time.

A plumber now needs to be understandable to Google Search, Google Maps, ChatGPT and other recommendation systems. Reviews, business information, the company website, local authority and clear service pages feed several discovery channels at once.

Which SEO niches are actually worth targeting?

The best SEO niches today have expensive customers, clear buying intent and enough fragmented suppliers that a new search asset can realistically compete. Raw search volume comes after that.

WordStream analyzed more than 13,000 U.S. search advertising campaigns running through 2026. The average advertiser was paying $66.69 for a lead, but the variation between industries was much more interesting.

Legal leads averaged $131.63. Real estate reached $102.51. Business services averaged $93.69, home improvement $90.92 and industrial or commercial services $75.19.

Advertisers have been bidding on many of these searches for years, so those prices give us a rough market test. Companies keep paying $80, $100 or $130 for enquiries because at least some enquiries turn into customers worth substantially more.

The most obvious niche is not automatically the best one. “Lawyer” has fantastic customer economics and ferocious competition. A narrower legal service in selected locations may produce a much more realistic opportunity.

The same applies elsewhere.

Instead of “home improvement,” we might find better economics in foundation repair, commercial roofing, water-damage restoration or a specific type of insulation.

Instead of “B2B consulting,” we could focus on ISO certification, cybersecurity compliance, industrial safety or ERP implementation.

This is also why search volume on its own is a poor selection metric. A keyword receiving 100,000 searches from people who want a simple explanation can be less valuable than one receiving 500 searches from companies looking for a specialist where one contract is worth $30,000.

Current paid-search prices make the difference tangible. Companies pay roughly $132 per legal lead, $103 in real estate, $94 in business services and $91 in home improvement. Average cost per click is also close to $10 in legal services and above $8 in home improvement.

We still need to inspect margins, close rates, competition and customer lifetime value. An expensive click can simply indicate a brutal market. The sweet spot often looks boring: valuable customers, clear intent and mediocre competitors.

What we want	Why it helps
High customer value	A small number of organic leads can support a real business
Strong hire, quote or buy intent	Traffic sits close to revenue
Fragmented competitors	We are less likely to fight one dominant brand
Repeatable geography or subcategories	Successful SEO pages can expand systematically
Weak existing websites	SEO skill creates a larger relative advantage
Trackable leads or sales	We can tell which rankings actually make money

Should I build a lead-generation business instead of selling SEO services?

Yes. If we can tolerate slower initial cash flow, an SEO-driven lead-generation business is more attractive than indefinitely selling SEO because we keep ownership of the demand we create.

An agency gets paid to increase somebody else's visibility.

A lead-generation company builds the website, rankings, brand, call history, conversion data and customer pipeline itself. We can sell enquiries to several businesses, negotiate exclusivity by territory, charge for appointments or eventually operate the service behind the leads.

The economics become especially interesting in categories where paid acquisition is expensive.

Take a home-service niche where Google Ads leads cost around $90 on average. If an owned organic asset reliably produces 80 qualified enquiries a month, we have created access to a stream of demand for which advertisers might otherwise spend roughly $7,000 just to generate a similar number of paid leads. The exact commercial value depends on lead quality, close rate and job size, but now we are discussing a customer-acquisition asset rather than a traffic website.

The model requires more work than ranking a site and collecting rent.

We need call tracking. We need to understand what constitutes a qualified lead. Providers need to answer quickly. Duplicate enquiries create arguments. Some locations will rank and others will fail. We eventually need relationships with several operators rather than one SEO client.

I would also avoid the old version of “rank and rent,” where hundreds of thin city pages funnel users to essentially the same provider. Google's current spam policies explicitly call out doorway pages aimed at similar city or regional searches, along with scaled content created mainly to manipulate rankings.

A modern lead-generation site should feel like a real customer-facing business. Useful price information, genuinely different location pages, verified providers, real photographs, customer feedback, calculators and original market data all make the asset stronger.

Can affiliate SEO still make real money now?

Affiliate SEO still makes money, but a generic review site would be one of my least attractive choices for someone starting from zero today.

The economics have become uncomfortable from both sides.

AI summaries intercept a growing share of informational and comparison clicks. At the same time, the affiliate takes only a small percentage of the transaction once a visitor finally buys.

Amazon's current standard Associates rates show how thin that can be. Home improvement, furniture, pets, tools, sports and many other large categories generally pay 3%. Apparel and several electronics categories pay 4%. Books and kitchen products pay 4.5%.

A $100 order at a 3% commission generates $3.

That math can still produce a large business at scale, but it demands a lot of purchase volume.

NerdWallet gives us a better indication of what search pressure looks like near the top of the market. The company is far more sophisticated than a normal affiliate publisher, with a major brand, proprietary tools and hundreds of millions of dollars in annual revenue. Yet in its second-quarter 2026 results, NerdWallet said consumer credit-card revenue fell partly because pressure on organic search traffic had continued for multiple quarters.

Consumer revenue still grew 8% overall because personal loans and deposit accounts performed better. That makes the example more useful: the company itself is healthy, while one search-heavy area is clearly struggling.

Affiliate SEO still makes sense when we bring something difficult to copy, such as hands-on testing, proprietary data, a strong audience or a real community.

Is an SEO directory or marketplace a better idea?

An SEO-first directory can become a very good business when the directory contains information people cannot get from another generic search result.

Directories naturally match how people search.

“Wedding photographers in Bristol.”

“CNC machining companies in Ohio.”

“French-speaking accountants in Dubai.”

“Pilates studios with reformer classes near me.”

One structured database can create hundreds or thousands of useful landing pages around locations, categories, prices and attributes.

AI has made the bad version incredibly easy to build. We can generate thousands of city pages in a weekend. Everybody else can too.

The valuable layer is the database underneath those pages.

Suppose we know each provider's real prices, availability, response time, service area, certifications, inventory or customer reviews. Now every SEO page can answer questions from information we actually own.

This also fits Google's current direction. Its 2026 guidance for generative search specifically encourages valuable, unique and non-commodity information. Its spam policies, meanwhile, warn against scaled pages that add little original value.

Should I build an SEO-first SaaS business?

SEO-first SaaS has more upside than an agency or lead-generation site, but I would usually build it after learning a market rather than inventing software because SaaS sounds attractive.

Recurring revenue changes the value of organic traffic.

If 10,000 highly relevant monthly visitors reach a free tool and 2% eventually become $40-per-month customers, that cohort represents 200 paying users and $8,000 of monthly recurring revenue. An informational publisher would need dramatically more traffic to create the same economic value from advertising.

Software also gives us a better defense against answer engines.

Google can summarize “how to calculate a mortgage payment.” A searcher who wants to model fifteen scenarios still benefits from an actual mortgage calculator.

The same idea applies to compliance checkers, generators, pricing tools, templates, reporting dashboards, benchmarks, converters and searchable databases.

Google's latest AI guidance is relevant here. The company continues to say that normal SEO practices remain the foundation for appearing in AI Overviews and AI Mode. Google has also rolled out separate Search Console reporting for generative-AI visibility worldwide, which confirms that AI search is being folded into ordinary organic discovery rather than becoming a completely separate discipline.

The difficult part of SaaS remains unchanged: people have to need the product enough to keep paying. Running a niche agency or lead-generation business first gives us a much better chance of discovering that problem from real customers.

Does ecommerce make sense if SEO is my main skill?

SEO can create a major advantage in ecommerce when the product catalog naturally generates valuable long-tail searches, but ecommerce is a poor choice when SEO is the only reason we like the business.

Selling products adds inventory, suppliers, returns, logistics, customer support and working capital. Ranking well does not solve any of those problems.

The more interesting opportunities appear in catalogs where the structure of the inventory itself creates search demand.

Replacement parts work this way.

So do industrial components, specialist tools, collectibles, compatible accessories, professional equipment and products differentiated by dimensions, model numbers or use cases.

A store with 5,000 properly structured replacement components can naturally create useful pages for thousands of searches. Product attributes become SEO assets because people genuinely search by compatibility and specification.

Compare that with launching another generic beauty brand.

We might rank some informational articles, but our SEO skill gives us less of a structural advantage because the real competition happens through brand, creative advertising, product, influencers, retention and merchandising.

What kind of SEO content still holds up against AI?

The most defensible SEO content now contains data, experience or functionality that an AI cannot reconstruct cheaply from everybody else's articles.

AI has made generic explanation extremely abundant.

A page called “10 Benefits of Solar Panels” can be rewritten endlessly. Searchers increasingly receive that information before visiting a website anyway.

Now imagine that we collect the actual quoted price of 4,000 solar installations and let users compare cost per watt by location, roof type and system size.

We suddenly own information.

The same principle works in almost every vertical.

An accountant-fee article becomes more useful when it contains hundreds of real quotes by city.

A CRM comparison becomes stronger when we track implementation times, migration costs and customer churn.

A home-renovation site becomes harder to copy when completed projects continuously feed a local price calculator.

A local marketplace gets better when it knows which providers are genuinely available this week.

Current search data reinforces this approach. Ahrefs' 2026 organic CTR study found the weakest click-through rates in answer-heavy categories such as health, science, pets and home-related information, while AI Overviews appeared heavily on question queries. Meanwhile, Google's own recent guidance tells publishers to provide unique, non-commodity content for generative Search.

How much should AI search change the SEO business I build?

AI search should push us toward higher-intent searches and harder-to-copy assets, but abandoning Google SEO for a completely separate “GEO” business would be premature.

Google's own behavior makes that fairly clear.

AI Overviews now have more than 2.5 billion monthly active users, while AI Mode has passed one billion. Google has rebuilt parts of Search around conversational answers and agents, so treating AI search as a temporary feature no longer makes sense.

At the same time, the underlying web index still matters. Google's current documentation says pages generally need to be indexed and eligible for normal Search before they can appear as supporting links in AI Overviews or AI Mode. It explicitly says the same foundational SEO practices continue to apply and that no special AI markup is required.

Google has even created dedicated Search Console reporting for generative-AI impressions.

Local AI discovery tells a similar story from another angle. In BrightLocal's 200,000-query study, business websites made up 93% of the unique domains used as AI sources.

So our skill set is broadening.

Technical SEO, site architecture, strong pages, brand authority, reviews, structured product or local information and original data can feed traditional rankings and AI recommendations at the same time.

How dangerous is it to build a company around Google traffic?

Depending on Google for nearly every new customer is still a serious business risk, even if we are excellent at SEO.

The latest AI changes make that risk easier to see because rankings can stay stable while traffic falls.

During the recent French AI Overview rollout studied by Ahrefs, impressions often remained healthy while CTR dropped. One extreme website gained 183% in impressions while clicks fell 1.9%. The page had not disappeared from Google. Google simply needed to send fewer people to it.

Algorithms create another layer of risk. SERP layouts change. Maps expand. Shopping modules appear. AI answers occupy more space. A marketplace can enter the search results above us.

We cannot SEO our way out of platform dependence.

We can, however, convert searchers into something we own.

A SaaS company gets an account and recurring usage.

An ecommerce company gets a customer and hopefully repeat purchases.

A marketplace can develop direct traffic from buyers and suppliers.

A local service business can earn reviews, referrals and branded searches.

A lead-generation company can build relationships with providers across many regions rather than relying on one buyer.

Should I use Google Ads before investing months in SEO?

Yes. Paid search is one of the fastest ways to test whether an SEO opportunity deserves six months of work.

SEO feedback arrives slowly.

We can rank a page after months of content, links and technical work only to discover that the leads are poor, customers hate the service or providers refuse to pay enough for the enquiries.

Google Ads can answer part of that question immediately.

Suppose we are considering a lead-generation site for commercial epoxy flooring. Before building 100 location and service pages, we can buy the exact searches we hope to rank for.

We see which queries generate calls.

We hear what prospects actually want.

We learn how many enquiries are genuine.

Providers tell us which jobs make money.

We discover what a lead is worth.

Then SEO scales what has already shown signs of working.

The reverse works too. Organic Search Console data reveals strange long-tail queries that we had never considered. We can buy those searches through paid ads, test their economics quickly and decide whether they deserve dedicated organic pages.

Which SEO business gives the best mix of cash flow and upside?

A vertical lead-generation business gives us the best overall mix today, while a niche SEO agency is the easiest way to finance the early stage.

Freelancing pays first but depends heavily on our time.

A niche agency can generate meaningful revenue quickly and teaches us one market in depth.

Affiliate publishing has weaker economics now because AI can absorb the informational searches on which many affiliate sites depend.

A directory or marketplace can become much larger, although getting enough useful suppliers and customer demand takes time.

SaaS offers recurring revenue and excellent upside, but building the wrong product can waste far more time than ranking the wrong page.

Lead generation sits in a useful middle ground. We need a website, rankings, conversion tracking and provider relationships, but we do not need a large engineering team or physical inventory. We can also measure the result in calls, forms and revenue rather than pageviews.

Business model	Time to first revenue	Value of strong SEO	Long-term ownership	Main problem
SEO freelancing	Fast	Very high	Low	Revenue follows our time
Niche SEO agency	Fast	Very high	Medium	Clients own most of the asset
Affiliate publishing	Medium	High	Medium	Informational clicks are under pressure
Vertical lead generation	Medium	Very high	High	We carry ranking and lead-quality risk
Directory / marketplace	Slow	Very high	Very high	Supply and demand must both reach scale
SEO-first SaaS	Slow	High	Very high	Product and retention are harder than SEO
Specialist ecommerce	Medium	High in the right catalog	High	Operations become much more complex

What business should I actually start if I'm good at SEO?

If SEO is our strongest skill, I would start a vertically focused customer-acquisition business in a high-value service niche, use a specialized agency offer to generate early cash, and gradually move the best search demand onto assets we own.

The order is important.

We first pick one industry where customers are valuable and search behavior is clear. Commercial roofing, specialist dental work, compliance consulting, commercial cleaning, industrial maintenance or another narrow service can fit; the exact niche has to survive real keyword and unit-economic research.

Then we get close to the operators.

A few agency clients give us immediate revenue and access to the information keyword tools cannot provide. We learn which calls become sales, which services have the best margins, which locations are underserved, which enquiries businesses reject and how much they already spend acquiring customers.

That knowledge tells us where to build our own search assets.

Over time, the revenue mix can shift away from SEO retainers and toward leads or transactions generated by websites we control. The strongest locations produce more pages, better data, reviews, pricing information and conversion history. Weak locations get cut instead of being subsidized forever.

Eventually, the company may stop looking much like an SEO business.

If we control a steady flow of customers in one vertical, we could build the marketplace those customers use. We could create software for the providers. We could acquire a provider. We could launch the service ourselves.

SEO got us into the market cheaply. The market knowledge and customer relationships make the next business possible.

That is why I would choose vertical lead generation over a generic agency, an affiliate blog or another large informational website today.

The opportunity in SEO is still very real. The better question now is what valuable transaction we can own once search brings us the customer.