Answer in brief
Most of what is sold as SEO to a small business is either eligibility work that takes an afternoon or content work that takes a year. Knowing which is which is most of the skill, and it decides what is worth paying for.
Two kinds of work sold under one name
SEO for a small business is sold as a single service and is really two. One is eligibility work — making sure a page can be found, read and indexed at all — and it is largely finite. The other is answering questions better than the alternatives, and it is largely continuous.
Confusing them is expensive in both directions. Paying monthly for eligibility work that took an afternoon is waste; expecting a one-off audit to produce continuous content is disappointment.
The useful discipline is to do the finite work first, confirm it, and then decide separately whether the continuous work is something you resource internally or buy.
What follows separates the two, names what you can check yourself for nothing, and is direct about what nobody can promise.
Eligibility: the part that is finite
A page that cannot be crawled, is marked not to be indexed, or is unreachable by any link, cannot rank however well written it is. This is the floor, and it is where a surprising number of problems live.
Google's Search Essentials sets out what has to be true for a page to be eligible at all. Reading it is an hour and confirming your own site against it is an afternoon.
The most common failures are mundane. A rule written to keep a staging site private that travelled into production, a page reachable only from a menu that does not render, a duplicate address competing with itself.
Once these are correct they mostly stay correct, which is exactly why they should not be billed as ongoing work. Recheck after a redesign, not every month.
Words: use the ones people actually type
The single most useful non-technical act is to write down the exact phrases customers use when they describe their problem to you, and compare them with the phrases on your pages.
Businesses routinely appear for a category adjacent to their offer rather than for the offer itself, because the site uses industry vocabulary and the customer uses ordinary vocabulary.
This is not about density. Writing to reach a percentage produces text that reads badly and answers nothing; using the reader's words is about recognition, and recognition is what makes someone stay.
Collect them for a month from calls, emails and messages. The list will be longer than any keyword tool's suggestions and better qualified, because every phrase on it came from someone deciding whether to buy.
One page, one question
A page that tries to answer four questions answers none of them well and competes with your own other pages for the same queries.
Give each page one question and answer it in the first paragraph, then earn the rest. A reader who has to hunt for the answer remembers the hunt, and a page that delays the answer to increase time on page is optimising the wrong number.
Google's guidance on creating helpful, people-first content makes this the explicit standard: material should be produced for people, and the test is whether a reader leaves having learned what they came for.
Where two of your pages answer the same question, merge them. Competing with yourself is a common and quiet way to underperform.
Speed and usability are reader problems first
A page whose useful content arrives after several seconds of decoration loses people before any question about ranking arises. That is reason enough to fix it.
MDN's performance documentation sets out the vocabulary for what is actually being measured, which is worth having before anyone quotes you for optimisation work.
The largest wins are usually unglamorous: oversized images, third-party scripts nobody can name a purpose for, and fonts that block the first paint.
Check usability in the same pass. The W3C's WCAG 2.2 quick reference is the practical list, and a page whose controls cannot be reached by keyboard is losing readers in a way no ranking report will show.
Links between your own pages
Internal links are the part of link work that is entirely within your control, and they are usually the part left undone.
Every page should be reachable by a link from another page, and the important ones should be reachable in a click or two from your most-visited page.
Use descriptive link text. "Read more" tells a reader and a crawler nothing; the question the destination answers tells both.
This also improves the reading experience directly, which is the more reliable reason to do it. A reader who finds the next relevant page without searching again stays with you.
If you serve a place, say where
For a business with a physical service area, being specific about the place is one of the higher-return things available, and it is content work rather than technical work.
Name the areas you actually serve on the page rather than assuming it is obvious. A reader deciding whether you cover them will leave if they cannot tell.
Keep the same name, address and phone number wherever you appear. Inconsistency across listings is a small, tedious problem that is worth an afternoon to fix once.
None of this needs a monthly retainer. It needs doing once properly and reviewing when something changes.
What nobody can promise
No supplier controls a search position. Results vary by person, location and device, so a single reported position is an average of measurements that were never the same measurement.
A guarantee of a first-page ranking is therefore either about a phrase nobody searches, or it is not a guarantee. An offer containing one has told you how it intends to report on itself.
Be similarly careful with promised traffic figures and conversion rates. Both depend on your offer and your market, and a number quoted before anyone has read your pages is a starting position rather than a forecast.
What a supplier can promise is specific work: a check, a list of findings, a set of changes, a rhythm. Those are the things worth buying and the things you can verify were delivered.
What the small paid check actually contains
For most small businesses the honest first purchase is narrow: where am I visible today, who else answers this, and what should I do first.
Search Snapshot is priced at $70 and delivered in 1-2 working days. It covers a visibility check on your top queries, three competitors side by side, and one page of findings with no deck, and it carries one clarification round. Implementation, content and ad spend are excluded. Here the practical constraint is equally important: Nobody controls a search position, and a supplier promising one is describing a metric you did not agree.
Note the exclusions carefully. The findings are the deliverable; acting on them is a separate decision you make afterwards, which is a more honest structure than bundling both into a monthly fee.
If the outcome is that the argument about what to fix is shorter, the purchase worked, and no analytics platform will record that.
When the answer is the offer, not the optimisation: Nobody controls a search position, and a supplier…
Sometimes the pages are eligible, the words are right, and enquiries still do not arrive. At that point the constraint is usually what the page says rather than whether it can be found.
Audit & Positioning is priced at $100 over 2-3 working days and covers audience framing, an offer rewrite and a channel map, with one round of revisions on the positioning document. Ad spend and media buying are excluded. For this case, the decision criterion is specific: Answering a real question in the words people use outperforms optimising a page nobody needed.
This is worth considering before more search work, because a rewrite improves the response of every channel at once and costs nothing on any platform.
Diagnosing which of the two you have is exactly what the visibility check is for, and it is cheaper than guessing.
When continuous work is genuinely the answer
Where a market is competitive and content is the mechanism, the work is continuous and buying it as a rhythm is more honest than buying it as repeated projects.
The Monthly Growth Retainer is $330 per month on a monthly cycle with 30 days notice to stop, covering monthly planning, a reporting rhythm and channel optimisation. Channels and volume are agreed each cycle; ad spend and third-party tool subscriptions are excluded. For the reader, the relevant outcome is concrete: Most of what is sold as SEO to a small business is either eligibility work that takes an afternoon or content work that takes…
Agreeing scope per cycle is what keeps a retainer from becoming a subscription to last quarter's decision, and the stated notice period is what keeps it easy to leave.
Note that content itself is not included in these packages. If writing capacity is your constraint rather than direction, say so when asking for a quote, because it changes what you should be buying.
A one-afternoon version
Confirm nothing important is blocked from crawling or indexing. Write down ten phrases customers actually used and search for three of them. Check whether each of your key pages answers one clear question in its first paragraph.
Open the site on a phone, note what loads first, and complete a form using only the keyboard. Fix whatever that reveals.
Make sure every important page is linked from another one with descriptive text, and merge any two pages competing for the same question.
That is most of small-business search work, it costs an afternoon, and it tells you whether the paid version has anything left to find.
Practical checklist
- Confirm your important pages are not blocked from crawling or indexing.
- Write down the exact words customers use, and check whether you appear for them.
- Give each page one clear question to answer and answer it in the first paragraph.
- Check the page loads and is usable on a phone and by keyboard.
- Make sure every page can be reached by a link from another page.
- Ignore any offer that promises a position by a date.
Questions and answers
What is the first thing to check in SEO?
Eligibility. A page blocked from crawling, marked noindex, or unreachable by any link cannot rank no matter what else is done to it. Google's Search Essentials sets out these basics, and confirming them takes an afternoon and costs nothing.
How much does an SEO check cost?
Search Snapshot is $70 over 1-2 working days and covers a visibility check on your top queries, three competitors side by side, and one page of findings with no deck, carrying one clarification round. Implementation, content and ad spend are excluded, so the fixes are a separate decision.
How long does SEO take to work?
Longer than paid placement and it varies by how competitive the query is and how much content already exists. Any supplier giving a firm timeline without knowing both is guessing. What is predictable is that a page answering a real question keeps working after publication.
Can anyone guarantee a first-page position?
No. Nobody controls a ranking, results vary by person, place and device, and a guarantee is either about a phrase nobody searches or is not a guarantee at all. An offer that includes one is telling you something about how it will report on itself.
Is keyword density still a thing?
Not as a target to hit. Writing to reach a percentage produces text that reads badly and answers nothing. Use the words people actually use because they are the words your reader recognises, and let the frequency be whatever a natural answer requires.