Does SEO pay for itself, and when

SEO occupies an awkward position in a budget. Its costs are immediate and fairly predictable. Its returns are delayed, uncertain in size, and compound in a way that makes a twelve-month view misleading in both directions — too pessimistic if you stop there, too optimistic if you assume the curve continues.

The useful discipline is to model it properly before committing rather than after, including the scenario where it does not work.

The chain, in order of certainty

Every SEO return model is the same three-step chain, and the steps differ enormously in how confident you can be about them.

Notice the direction: you are most confident about the last step and least confident about the first, which is the one that gets multiplied through everything. That is why SEO forecasts have such wide ranges, and why a supplier presenting a single confident traffic number should be treated with more suspicion than one presenting a range.

Payback period, not annual return

Annual return is the wrong headline for SEO because it hides the shape. A channel returning nothing for five months and then a great deal has a respectable annual figure and a cash-flow profile that can still sink a small business.

Payback period — how long until the cumulative return covers the cumulative spend — is the number that actually governs the decision, because it tells you how long you have to fund it from something else. If that period is longer than your runway, the channel is wrong for you right now regardless of how good the eventual return is.

What makes the compounding real

The strongest argument for SEO is not its rate of return but its persistence. An article published in month two that reaches a good position is still earning in year three without further spending. Paid advertising has no equivalent property — traffic stops the day the budget does, at exactly the rate it started.

This has two practical consequences for the model. First, a three-year view is more honest than a one-year view, because a one-year view charges all the cost against a fraction of the benefit. Second, the maintenance cost of existing content belongs in the model — a page earning in year three probably needed updating in year two, and that is not free.

When the answer is no

There are situations where the model reliably says do not, and recognising them saves considerably more money than optimising a campaign that should not exist.

• Nobody is searching. If the volume for what you do is genuinely tiny, ranking first for it changes very little, and no amount of execution creates demand that is not there.
• The competition is structurally out of reach. If the first page is occupied by organisations with content teams and a decade of authority, the honest cost of competing may exceed what the customers are worth.
• Your runway is shorter than the payback period. Correct channel, wrong time.
• Your conversion rate is very low. Multiplying more traffic by a poor conversion rate produces a poor result at a higher cost. Fix the site first — it is cheaper and it improves every other channel simultaneously.

The last one is the most common and the most actionable. SEO amplifies whatever your site already does with a visitor. If that is not much, amplification is an expensive way to find out.