Is SEO Worth It for a Small Business in 2026?

Is SEO worth it? Search the question and you will find page after page saying yes, written by people who sell search marketing. We sell SEO too, so treat the rest of this with appropriate suspicion and check the arithmetic yourself. The arithmetic is the point. It is also the thing nobody publishes.

It comes down to three numbers

Not benefits lists. Three numbers about your business, and you can gather them in twenty minutes.

1. What a customer is worth to you. Not the first invoice. The first invoice plus whatever they spend afterwards, minus what it costs you to serve them.
2. How many people search for what you sell, in the places you serve. A real monthly figure, not a feeling. Any keyword tool will give you a usable estimate for free.
3. How long you can fund something before it has to pay for itself. This one is about your bank balance and it decides more than the other two.

The break even sum

Here is the whole calculation. Monthly fee, divided by customer value, gives you the customers per month you need. Divide that by the rate at which enquiries become customers, and you have the enquiries you need. Divide that by the rate at which visitors become enquiries, and you have the visitors you need.

Two worked examples with the same fee.

Both of those are small businesses. The fee is identical. The answer is opposite, and no amount of discussion about brand awareness changes that, because the roofer needs fifty of the right visitors and the coffee shop needs a different channel entirely.

When the answer is honestly no

This is the section that does not appear on the pages ranking for this question. If several of these describe you, search is not your first move and anybody telling you otherwise is selling.

• Nobody searches for it yet. If you are creating a category, your problem is demand creation and search captures demand that already exists. Advertising and social are the right tools for that job.
• You need customers inside ninety days to survive. There is no version of this that reliably delivers inside a quarter. Fund something that can.
• Your realistic budget is under a few hundred a month. At that level you are buying a report, not a programme. Buy a one off audit and do the recommendations yourself.
• Your whole market is one map pack. A single location business whose customers search and then look at the map should spend on the profile, the reviews and the photos first. That is cheaper, faster, and often the entire job.
• Your website cannot take an enquiry. If the form is broken, the phone goes unanswered, or the pricing scares people off, more visitors will only produce a more expensive version of the same problem.
• You cannot give it any time. Someone has to approve content, answer questions about your trade and sign off technical changes. An engagement with no access to anybody who knows the business produces generic pages.

When it is clearly worth it

Three patterns pay back reliably. An established site that has never been optimised, where authority already exists and is simply being wasted. A business with high customer value and measurable search demand, where a handful of the right visitors covers the fee. And a business already buying clicks successfully, where the ad account proves which phrases convert and search can take over the cheapest of them permanently.

That third pattern is the strongest single argument for funding it, and almost nobody makes it: if you already know a phrase produces customers because you have been paying for it, building the page that ranks for that phrase is the lowest risk version of this work available.

The 2026 complication

The honest answer to this question got worse in the last two years, and the pages telling you it is worth it mostly have not updated their arithmetic.

Google now answers many questions directly on the results page. Pew Research Center measured in July 2025 that the share of searches producing a click on a normal result dropped from roughly 15 per cent to roughly 8 per cent when an AI summary was present, and that about one per cent of people clicked a link inside the summary. Ahrefs, looking at 300,000 keywords in April 2025, found the top position lost about 34.5 per cent of its clicks when a summary appeared above it.

What that means for your sum: the number of visitors a given ranking produces is lower than it was, so the ranking you need is higher than it was, so the timeline and the budget are both worse than a 2022 case study implies. It does not mean the channel stopped working. Informational questions lost the most, and commercial searches by people ready to buy lost the least. But if a proposal quotes you a traffic projection built on old click rates, it is projecting a number that no longer exists.

What a real budget looks like

The published figures contradict each other and the contradiction is the useful part. Ahrefs surveyed 439 providers in 2024: average monthly retainer around 2,900 US dollars, agencies around 3,200, freelancers around 1,350. Clutch directory data puts the average around 3,200 a month. SE Ranking surveyed 260 providers in December 2024 and found 64 per cent charging under 1,000 a month.

Both pictures are accurate and they describe different products. Under roughly a thousand a month you are buying one person doing part of the work, which can be exactly right if you know which part you need. Between one and three thousand you are buying a small amount of all three kinds of work, and the risk is that it is spread too thin to move anything. Above that you are buying enough capacity to attack content, technical and authority at the same time.

The number that matters is not the fee. It is the fee divided by your customer value, which is why the roofer and the coffee shop get different answers to the same price.

The value that never shows up as traffic

Three real benefits that do not appear in a sessions graph, offered without exaggeration.

The pages keep working. Advertising stops the day the card declines. A page that ranks keeps answering the question next year, which is the actual difference between the two channels.

It makes your other channels cheaper. The content written for search is the content your sales team sends, your ads land on and your emails link to. Much of it would have been commissioned anyway.

It is increasingly what AI answers quote. The pages getting cited in AI summaries are overwhelmingly pages that already rank well, so the work that earns a ranking is currently also the work that earns a mention. Nobody can promise a share of those citations, and any proposal pricing that separately as a new service is charging twice for one job.

How to buy it so you can stop

If the arithmetic closes, protect yourself with structure rather than optimism. Buy a diagnostic first, priced and scoped on its own, so you learn what is actually wrong before committing to a year. Then insist that the engagement names what ships each month, not just the areas it covers.

Set a review at month four on the leading indicators: are the pages published and indexed, are impressions and query counts up, are the audit items closed. Revenue at month four tells you nothing. Those three tell you plenty. And make sure you hold the accounts: your Search Console, your analytics, your Google Business Profile, your website. If leaving means losing access to your own data, the contract has made the decision for you.