How much should a business spend on SEO? The arithmetic, not a price list

Spend on SEO what a realistic number of extra customers is worth to you, spread over the months it takes to arrive, and not a pound before you know whether the searches exist. The arithmetic takes ten minutes with three numbers you already have: what a new customer is worth, the share of enquiries that become customers, and how many extra customers a month would change the business. The price list comes after that, not before.

Start from the customer, not from the quote

Most businesses ask "how much does SEO cost" and get a number that has nothing to do with them. The better question is what SEO has to produce to be worth paying for. A worked example, with round numbers chosen for the sum rather than taken from any client: a business where a new customer is worth £900, where one enquiry in three becomes a customer, and where four extra customers a month would matter. Four customers need twelve enquiries a month. Twelve enquiries a month from search is a modest target for most local services and an ambitious one for a niche product. Four customers a month at £900 is £3,600 of work; a retainer that costs a fraction of that and produces those enquiries within six months pays back many times over, and a retainer that costs half of it and produces nothing traceable by month six should stop.

The point of the sum is not the answer but the shape. The value of a customer decides how much a month can be justified. The close rate decides how many enquiries that needs. The number of searches that exist in your area decides whether those enquiries are available at all, and that is the number the audit checks before anyone quotes.

Worked example two: an online shop

The same arithmetic with a shop's numbers, again chosen for the sum rather than taken from a client. Suppose the average order is £80 and the margin on it is £30, so each order is worth £30 to the business, not £80. Suppose one visitor in fifty buys, which is a common enough rate for a small shop. Then a hundred extra orders a month, worth £3,000 of margin, need five thousand extra visitors a month from search. Five thousand visitors is a great deal for a small shop and a modest number for a category page that ranks well for a search with volume, which is why shop SEO lives or dies on the category pages and the feed rather than on product copy. The retainer is justified if those visitors are available, and the query volumes for the categories tell you whether they are before anyone quotes.

The shape of the sum is what matters: margin, not revenue; conversion rate, not visitors; and whether the searches exist at the volume the maths needs. Any budget that skips one of the three is a guess.

The four ways SEO is priced

What a fee actually buys

Whatever the model, the fee buys a competent person's hours, and the honest way to judge a price is to ask how many. A retainer at the low end of the UK market buys roughly eight to twelve hours a month of one experienced person, front-loaded into month one: enough to fix a small site's technical problems, build the profile, rewrite a page a month from the query data, keep the profile and reviews moving, and read Search Console. It is not enough to write ten pages a month, chase links in a national market and run advertising, and any company promising all of that at that price is either not doing it or doing it with a machine.

A price far below that level is buying software output or offshore text by the thousand words. A price far above it should come with a team and a scope that needs one. In between, judge the price by what the month contains, listed, and by whether the person who will do it is the person you spoke to.

When to spend nothing

• Nobody searches for the service. Some businesses are bought on recommendation and never typed into Google. The query volumes for your service and area are checkable before you spend; if they are near zero, SEO cannot create them.
• You need enquiries this month. SEO shows in months. A business with an empty diary needs advertising now and SEO underneath it, and a company that takes three months of retainer for a problem it cannot solve on that timescale is not being straight with you.
• The site cannot be edited. A builder that locks titles and pages cannot be optimised. Replace the site first, then spend on SEO.
• You cannot give fifteen minutes a month. The pages come from a conversation about the work. Without it they are guessed, and guessed pages do not rank.

Signs you are paying too much, or too little

Too much: the report is long and the numbers in it are not from Search Console; the month's work is not listed; the same page has been "optimised" three times; the fee rose because the market is competitive rather than because the scope grew. Too little: no technical work has been done in six months; the pages read as if a machine wrote them; nobody has asked you for a photograph or a review; the profile still has the wrong hours.

Our prices are on the pricing page rather than in this guide, on purpose: a guide about how to think about a budget should not end in a sales pitch. The calculator on that page runs the arithmetic above with your own numbers, and the free audit checks whether the searches exist before anyone quotes.

How to judge the spend at month six

By month four, impressions for the named searches should be rising in Search Console. By month six, there should be enquiries you can trace to the work, and the company should have told you in writing whether it is on track. By month twelve, the enquiries should cover the cost several times over. If the month-six judgement is that it is not working, the right response is to change the plan or stop, not to add three more months on hope.