Picture yourself in a plumber's cramped back office, laptop open, showing the owner why the company never comes up when someone across town searches for an emergency callout. You have sold nothing yet. You have simply made a problem visible, and you can see them leaning in. That moment is where most SEO clients are won.
Now think about your current job. Maybe you already do marketing for someone else, and you watch agencies bill your boss for work you could do. Maybe you just understand search better than most people you know. Either way, your salary moves slowly while the cost of everything moves fast.
Here is why the timing matters. AI answers have changed how search works, and plenty of old-school SEO folks are struggling. Local businesses still need to be found, and the practitioners who learn the new rules now become the people business owners trust. The ones who wait are learning while the early movers collect referrals.
The numbers: getting to that first meeting costs you somewhere between nothing and $500, because the tools most people begin with are free until you have a few clients paying. The low end of this work runs about $500 a month, and it takes 2 to 6 months before that properly covers your time, since technical fixes and fresh pages need until months 3 to 4 to start registering.
Tonight, pick one local business you already use, search the phrase a customer would actually type and write down where it ranks.
Any honest guide to this work has to start with that change, because it decides what you can sell and what you should stop promising.
The Pew Research Center analysed the real browsing behaviour of US adults in March 2025. Its findings are the clearest public data you will find on what AI summaries have done to clicks.
Three conclusions follow, and they should shape everything you sell.
The service business survives. It has been repriced, and the practitioners in trouble are the ones still selling volume of informational content by the article.
So look at the clients you would want. Do their customers search to learn something, or to book someone? Your answer tells you how much of the change above lands on them.
Your clients want more of the enquiries that turn into revenue. They used rankings as a stand-in because rankings were easy to measure.
That distinction now matters commercially, because the stand-in has drifted away from the outcome. The engagements that last are the ones you scope around the outcome from the start.
Technical work is the easiest for you to sell, because it produces findings a client can see. It is also the easiest to over-sell, because a technically perfect site with nothing worth reading still earns nothing.
Which of these three are you strongest at today? Lead with that one in your first conversations, and be candid with the client about the other two.
So you know what the job consists of. What it is worth to the person paying is a separate question, and the answer has little to do with the hours you put in.
Three structures dominate, and the differences between them matter more to you than the rate.
One local retainer at the lowest figure here, 500 dollars a month, could pay for something your household can feel once costs are out: the kids' football fees and new boots, covered by fixing a plumber's map listing and service pages. That client also tends to stay, because they can trace the phone calls back to you.
The most reliable way into a new client is a paid audit. A free proposal rarely gets you there.
A free proposal is guesswork you produce before you have access to anything, and it competes with every other free proposal on price and promises. A paid audit is a small, defined transaction that settles three things at once: that the client pays, that you produce work to a standard, and what is actually wrong.
Price it low enough to be an easy yes and high enough to be taken seriously. Deliver it as findings a non-technical owner can act on, ordered by expected impact rather than by category, with an honest note of what you cannot know without more data.
The audit often sells the retainer without any sales conversation, because the findings make the case for you. It also protects you: if an audit shows a site cannot really be fixed within the client's budget, you can both stop before a year of disappointment.
What would you price your first audit at so a busy owner says yes without calling their accountant? Pick the number now, and write it at the top of your outreach notes.
Local is the segment least hurt by the change above, and it is the easiest door for you to walk through as a new practitioner.
A search for a service in a place ends in a map pack and a set of local results, because the user needs somebody nearby who can actually do the job. An answer engine can describe what a plumber does; it cannot send one to your door.
The work is concrete and mostly unglamorous. Claiming and completing the business profile properly. Keeping name, address and phone number consistent across the places that list businesses. Choosing categories correctly, which matters more than most owners realise. Asking for genuine reviews systematically instead of now and then. Writing location and service pages that say something specific, instead of the same paragraph with the town name swapped.
The commercial appeal for you is clear. Clients see outcomes within weeks rather than quarters, your competition is other small local firms rather than national brands, and the businesses buying it can trace a phone call to you in a way a national e-commerce client often cannot.
You can do all of this well and still be asked what it bought. That conversation goes better when you open it yourself, early, before a client opens it for you.
SEO has an honesty problem in reporting, and getting ahead of it protects your relationship with the client.
If a client asked you tomorrow what last month's work bought them, could you answer in two sentences with their own enquiry numbers? If not, agree those numbers before you sign.
Getting the first clients
That settles the craft side. None of it earns you anything until someone agrees to pay, and the first yes takes the longest.
Start where you have context. An industry you have worked in, a town you know, a type of business you understand. Knowing how a client's customers search is worth more than general SEO skill, and it is the thing an out-of-town agency cannot copy.
Lead with a finding. A short message naming one specific, verifiable problem on a prospect's site beats any pitch. It shows competence in the first sentence, and it obviously came from you, by hand.
Take the work that is visible. Local businesses, professional practices, trades. They can see results, they refer readily to others in the same trade, and they are underserved because most agencies chase larger accounts.
Publish what you find. Case studies with real numbers, including the parts that did not work, are the most effective marketing in this field, and almost nobody does them.
Which trade or town do you already know better than any agency an hour away? That is where your first three audits should go.
Common mistakes
Selling rankings. They have drifted away from traffic. Sell enquiries and report rankings as supporting evidence.
Publishing volume in answer to the AI summary problem. More articles restating general knowledge means more of exactly what gets absorbed without a click.
Buying links. Still sold everywhere, still against every search engine's guidelines, and now with a worse return than it had when it worked.
Promising a timeline you cannot control. Results depend on the client acting on your recommendations. Say so in writing, and track what you recommended against what was done.
Ignoring what the site already ranks for. Existing positions that are slipping are usually cheaper to recover than new ones are to win, and clients feel a recovery straight away.
Taking clients whose site cannot be fixed. A business with no distinct offering, no capacity to publish and a hostile developer will end as a refund and a bad review. Walk away early.
Waiting feels safe and costs you more than you notice. Every month, the plumbers, dentists and cafes near you either hire someone or give up on search. The first clients go to whoever shows up with a clear problem on a screen. Tonight, run that search for one business and write a three line note about what you found.
Who should skip this
If you need results in weeks, choose paid acquisition instead. SEO compounds and it starts slowly, and the honest sales conversation says so.
If you are unwilling to keep learning, choose a steadier field. This discipline has just been reshaped by generative answers, it will move again, and practitioners who learned one playbook are the ones struggling right now.
If you are uncomfortable delivering bad news, do not take retainers. Months where the numbers fall for reasons outside your control are certain, and your ability to explain them accurately is what keeps the client.
If you hope to sell links or automate content at volume, you have picked the part of this business that is both least defensible and most exposed to the change described at the top of this page.
Where this goes next
Here are three judgements about direction, offered as reasoning you can weigh for yourself.
Optimising to be a cited source becomes its own discipline. If most informational queries resolve without a click, the value of appearing inside the answer rises even though it sends little traffic, because it shapes the buyer before they shop. Expect this to be sold as a separate service, and expect its measurement to stay poor for some time.
Local and transactional work gains share. Budgets will gather in the segments where a website still has to be visited, and practitioners now serving informational publishers will move toward businesses that take bookings.
Brand becomes the lasting asset. Branded search is the one query type an answer engine has no reason to intercept, and building a name people search for directly is more and more the point of the content, where it used to be a side effect.
Brand and local search are where the lasting work is heading, and the practitioners who build those skills now will be the obvious hires as budgets shift. Business owners stay loyal to whoever helped them first. Each month of delay hands that loyalty in your town to somebody else.
The technical audit in practice
Most technical audits sold to small businesses are a tool export with a logo on it. The ones worth paying for answer a small number of questions properly, and those are the ones you want to be selling.
Can search engines reach everything that matters, and nothing that does not? Check what is actually indexed against what should be. The two most common findings point in opposite directions: important pages excluded by a directive nobody remembers adding, and thousands of near-duplicate URLs from filters, sorts and session parameters diluting the site into noise.
Does the site render? A page that needs JavaScript to produce its main content may look very different to a crawler than it does to you. Fetch the page the way a crawler would and compare it with what a visitor sees. This one check explains a large share of otherwise baffling invisibility.
Is there one canonical version of each page? Sites routinely serve the same content on several addresses at once: with and without www, over both protocols, with and without a trailing slash, with tracking parameters attached. Consolidating that is often the cheapest ranking improvement you can offer.
Is it fast on the device people actually use? Test real mobile performance, measured on field data where the site has enough traffic to produce it. A desktop connection in your office tells you very little.
Does the structured data describe what the page really is? Correct markup helps a page be understood and can earn richer results. Incorrect markup, describing things the page does not contain, becomes a liability.
Is the internal linking doing any work? Most sites bury their most commercially important pages several clicks from anywhere, and link to them with anchor text that says "click here". Reorganising internal links costs nothing, needs nobody's permission, and is consistently underused.
Order your deliverable by expected impact and write it in the client's language. A finding a business owner cannot act on becomes a note for a developer who has never been briefed, and it goes nowhere.
When you run your next audit, could you name the three findings that will change this client's revenue? If you can only produce the full list, the ranking work is still ahead of you.
Content that survives an answer engine
If a summary can fully answer a question, a page answering that same question will be read less. The content that keeps earning attention has qualities a generated summary simply cannot reproduce, and these are what you help your clients make.
First-hand experience. What happened when you actually did this, with the specifics. An answer engine can pull together what is generally known; it cannot report what you saw last month.
Original data. A survey of the client's customers, an analysis of their own numbers, prices they collected. There is a second benefit too: original data gets cited, and being cited builds authority.
Judgement, stated plainly. Which option is wrong for whom, and why. General sources hedge, because hedging is safe. A specific recommendation with reasoning is more useful to the reader and easier to remember.
Depth on a narrow thing. A complete treatment of one specific decision beats a shallow tour of a whole category, and the narrow piece is the one that ranks, because it matches how people search once they are close to acting.
Something to do at the end. A calculator, a template, a checklist. A page that gives the reader a result to keep has a reason to be visited that a summary cannot satisfy.
The matching instruction about what to stop producing is uncomfortable if you sell content packages: definitional articles, listicles assembled from other listicles, and "ultimate guides" that restate consensus. That material is the raw input answer engines consume, and publishing more of it means competing with a machine at the thing it does best.
Look at the last five articles on a prospect's blog. How many could a summary replace word for word? That count is a useful opening line for your next call.
Authority, without buying it
Links remain a significant signal, and the market for acquiring them is almost entirely disreputable. The honest routes are slower, and they work.
Be worth citing. Original data and genuinely useful tools attract references without you asking. This is the only method that compounds.
Digital public relations. Give journalists and industry publications something factual they can use, at the moment they are writing about it. This is a relationship business, and it sits closer to journalism than to marketing.
Fix what already points at the client. Sites that mention them without linking, links pointing at pages that have moved, and profiles that were never completed. Recovering existing references is cheaper than earning new ones, and almost nobody does it first.
Genuine partnerships. Suppliers, industry bodies, local organisations, sponsorships that actually happened. These are unglamorous, defensible and permanent.
What to avoid is simpler to state. Anything sold per link with a guaranteed placement. Networks of sites that exist to host links. Mass outreach offering payment for a mention. These carry risk, search engines increasingly discount them in practice, and they leave a footprint that a future practitioner, maybe you, will have to clean up.
Running it as a business
Up to here you have been thinking like the person doing the work. From here you are also the person who has to keep the work coming in.
Delivery is only half the job, and the other half is the one that ends most SEO practices.
Contract the scope precisely. What is included each month, what counts as extra, how many hours or deliverables, who makes the changes. The most common unprofitable retainer is one where you ended up doing the client's web development for free.
Decide who implements. Recommendations that never get implemented produce no results, and the client will still judge you on results. Either you have access and do the work, or the client commits in writing to a process for acting on it.
Cap your client count on purpose. SEO work expands to fill whatever time you have, and with too many retainers you deliver thin work to all of them and lose several at once. Fewer clients at higher prices is the version that survives.
Expect churn, and plan for it. Clients leave when they are happy, because they think the work is done, and when they are unhappy. Keep a pipeline running all the time, whether or not a slot is open.
Keep your own site working. If your own site does not rank, you carry a credibility problem into every sales conversation, and it is the first thing a prospect checks.
Document everything you change. Keep dated records of what was implemented, so when a ranking moves either way you can say why. This is also the evidence that protects you when a client's developer breaks something and the traffic falls.
How many retainers could you serve properly next to everything else in your week, honestly? Write that number down before your first proposal goes out, and treat it as a hard cap.
Three or four retainers billed on the 1st change how a month feels. If a dentist, a roofer and an accountant each pay on time and leave enough after costs, the rent goes out on schedule and you stop refreshing the banking app in the last week. Keeping a dated record of every fix is what keeps those invoices coming back.
A realistic timeline
- Months 1 to 2. Audit, fixes, and the unglamorous work of getting access to things. Little visible movement. What you deliver is the removal of obstacles, which is real and invisible.
- Months 3 to 4. Technical improvements start to register, slipping pages recover, and the first new content is indexed. Impressions usually move before clicks do.
- Months 5 to 8. If the work is right, this is where clients see it: enquiries rising, positions on commercially useful terms improving. It is also when most engagements get cancelled by clients who expected it sooner.
- Months 9 to 12 and beyond. Compounding, if the client kept publishing and implementing. By now the gap between clients who acted on recommendations and those who did not is very wide, and it comes almost entirely down to implementation, with strategy playing a small part.
So say all this before signing, and not in month three. A client who was shown the shape of the curve in advance is a client who is still with you when it turns upward.
Specialising, and why generalists struggle now
The practitioners doing well are almost all specialists, and they usually specialise by industry rather than by technique.
Industry specialisation compounds. Serving your fifth dental practice is far more efficient than serving your first, because you already know how patients search, which directories matter, what the regulator lets a practice claim, and which page structures convert. You can quote faster and with more confidence, and referrals within an industry keep coming.
Technique specialisation is fragile. Calling yourself "the link building specialist" or "the technical SEO" ties your business to a tactic that one algorithm change or policy shift can devalue. The people whose whole practice was one technique are the ones reinventing themselves right now.
Some industries make much better clients than others. Businesses with a high value per customer, a real need to be found locally, and a decision-maker who can approve spending are where the work pays. Businesses selling low-margin products into a national market against large competitors are where small practices lose money.
Regulated sectors reward specialists heavily. Legal, medical, financial and property services all have advertising rules that limit what may be claimed. If you know those limits, you are worth far more than a cheaper practitioner who writes something that creates a compliance problem, and clients in those sectors know it.
Which industry could you serve five times over without getting bored? That is probably your specialism, and the sooner you name it, the sooner the referrals start.
What to do about answer engines specifically
Since the change at the top of this page defines the current market, it deserves a practical response from you as well as an observation.
Measure the gap. Track impressions against clicks for the client's informational pages. A widening gap is the effect arriving, and showing the client that measurement before they notice it in revenue is what keeps their trust.
Move effort toward queries that need a site. Comparison, local, transactional, branded. Audit the client's existing content by intent and be honest about which pages are now unlikely to earn traffic whatever their position.
Make the site the kind of source an answer draws on. Clear structure, factual precision, plain statements of who the business is and what it does, and structured data that describes reality. Being cited brings few clicks directly, and it shapes what the user believes before they search for a supplier.
Build the branded query. The one search an answer engine has no reason to intercept is somebody looking for the business by name. Everything that makes a business known, including work that is outside conventional SEO, now feeds the channel that stays reliable.
Stop selling what no longer works. The most valuable thing you can do as an honest practitioner in this market is tell a client that the content package they bought last year should not be renewed. It costs you a line item, and it keeps the account.
The tooling in this field is expensive, and most of it is optional when you serve small businesses.
What you genuinely need. Search Console and an analytics package, both free, both connected to the client's property rather than yours. A crawler to audit sites. A way to check what a page looks like once rendered. These cover most of the work.
What is worth paying for once you have clients. A rank tracker, because clients ask and manual checking does not scale. A keyword and backlink dataset, because competitive research without one is guesswork. A local listing management tool if you serve multi-location businesses.
What usually wastes your money early on. Enterprise platforms priced for in-house teams, automated content generators, and anything sold on the promise of automating outreach. The first is capacity you do not need, the second produces the material described above as least defensible, and the third damages the relationships the work depends on.
Charge tools to the engagement. Client-specific subscriptions belong in the price, out of your margin, and clients accept this readily when you itemise it at the start.
A word of warning: tool fluency and skill are different things. If you can produce a hundred-page report from a crawler and cannot say which three findings matter, you are selling a document, and clients work that out by month three.
Working with the client's developer
More SEO engagements fail on implementation than on strategy, and implementation usually runs through somebody who does not report to you and never asked for your recommendations.
Assume good faith and a full queue. The developer has a backlog, competing priorities and no context on why a canonical tag matters commercially. Treat them as a busy colleague.
Write tickets. A recommendation phrased as a specific change, with the affected URLs, the current state, the desired state and a way to check it, gets implemented. A twenty-page PDF sits unread.
Rank ruthlessly. Three changes that matter, sent as a short list, beat forty findings. You can send the next three when those are done.
Explain the commercial consequence. "This will take about an hour and we estimate it recovers a page that produced eleven enquiries last year" moves up a queue far faster than "this is a technical SEO best practice".
Check after deployment, and say thank you. Changes get half-implemented, reverted by later releases, or applied to a template that affects more than intended. Check, and tell the developer what their work produced, because a developer who sees results becomes your ally.
Where a client has no developer, decide openly whether you are taking on implementation. If you are, price it. Doing it unpriced because it feels easier than asking is the most common way a retainer turns unprofitable.
Who actually makes changes to the site at your next client? Find out in the first meeting, because that person decides whether your work shows up.
Reporting that keeps clients
The report is the only contact many clients have with your work, and most reports are built to show effort when they should be helping someone make a decision.
Lead with the outcome the client cares about. Enquiries, calls, bookings, revenue where you can see it. Traffic second. Rankings third, and in context.
Explain what changed and why. Three sentences on what was done, what moved, and what it means. A client who cannot summarise their own report to a business partner will not renew it.
Say what did not work. Something did not, every month. Reporting it costs you nothing in a good month and buys enormous credibility in a bad one.
State the next three actions and what you need from them. Most reports stop at description. A report that ends with a decision to make, or a task handed back to the client, produces the implementation that produces results.
Keep it short. A one-page summary with the detail available underneath. Length reads as padding, especially to owner-managers who will spend four minutes on it.
Realistic income, and how practices actually scale
The economics of this business are set by how many clients one person can serve properly, and that number is smaller than most new practitioners assume.
As a solo practitioner delivering genuine work, you can usually hold somewhere between five and ten retained clients depending on the depth of each engagement, plus occasional project work. At small business retainer levels that is a solid professional income, and it is your ceiling until something structural changes.
The three routes past that ceiling are the same three open to any service business, and each feels different.
Raise the price, which requires specialising and having results to show. This is the cheapest route and the one most practitioners put off longest.
Hire, which turns you from a practitioner into a manager and brings the problem every agency has: quality control on work you no longer do yourself. Many people who make this move discover they preferred the old job.
Productise, which means selling a defined, repeatable deliverable at a fixed price instead of an open-ended retainer. Audits, local visibility setups, migration support. This scales better than hourly work, and it suits you if you do not want staff.
The failure to avoid is taking more retainers than you can serve at the price you set, which brings thin delivery, poor results and cancellations all at once. Capacity is the limit, and pretending otherwise is what turns a good practice into a bad one.
Of those three routes, which one fits the life you actually want: higher prices, a team, or a product? Choose before you hit the ceiling, so the ceiling does not choose for you.
Sustained income near 15,000 a month, the top of this guide's income range, belongs to a practice with a full roster and a specialism, held for a long time. If you ever reach a sustained share of that, the commute and the alarm clock become optional, and the hours you work are the ones you choose. Getting there runs through capacity, one retainer you can serve well at a time.
Picture a few local businesses paying you every month because their phones ring more since you started. That income covers the rent before the month has even begun. Each new retainer adds a little more breathing room, until the job you have now becomes a choice.
Getting started this month
Week one. Pick the industry or place you already understand. Set up Search Console and analytics on your own site, because you cannot credibly sell what you have never configured. Learn one crawler properly.
Week two. Audit three real businesses in your chosen niche without being asked. Find one specific, verifiable problem in each. This is your training and your outreach list at the same time.
Week three. Contact those three with the finding, and hold the offer back. Propose a paid audit at a price low enough to be an easy decision. Expect two of the three to ignore you.
Week four. Deliver the first audit properly, in the client's language, ordered by impact. At the end, ask whether they would like help implementing it. That is the retainer conversation, happening at the moment your credibility is highest.
Then repeat, and write up what you find. The practitioners with the strongest pipelines are the ones with public evidence of their work, and that evidence builds one client at a time.
What makes this durable
Two things have kept this a viable service business through every earlier upheaval, and they are your reasons to expect it to survive this one.
Businesses still need to be found by people actively looking for what they sell, and almost none of them have anyone in-house who understands how that happens. That gap has existed for twenty years and shows no sign of closing, because search is a technical subject that changes all the time and business owners are busy running businesses.
And the work rewards judgement more than access to a tool. Anyone can buy a crawler. Knowing which three of its four hundred findings will change this specific client's revenue is the part worth paying for, and you learn it by doing the work on real sites with real stakes.
The practitioners in difficulty are those who sold a repeatable tactic where they should have sold judgement. That was true when directory submissions stopped working, when article spinning stopped working, when guest post networks stopped working, and it is true now that publishing volume of general information has stopped working. The specific tactic changes every few years. The underlying job has stayed the same, and it is waiting for you.