Making the case

SEO ROI Is Unknowable in Advance. The Break-Even Isn't.

Nobody can honestly forecast your SEO return. But anyone can calculate, before signing, how many extra customers a year the engagement must produce to break even — and that number settles the decision.

Ask an agency what return you can expect and you will get one of two answers. The vague one — “it depends on your market” — is at least honest. The specific one is a number from someone else’s case study, produced by a different business with a different site, a different category and a different sales process.

Neither helps. But the impossibility of forecasting the return has quietly excused everyone from doing the arithmetic that is available, which is the cost side and the break-even. That calculation takes ten minutes, requires no forecast, and settles most of these decisions.

The only number that matters before you sign

How many additional customers a year does this engagement need to produce in order to pay for itself?

That is it. Three inputs, all of which you already know:

  1. Annual cost of the engagement. The retainer times twelve, plus content and development costs quoted separately.
  2. Gross profit on one customer. Not revenue — what is left after delivering.
  3. Break-even = 1 ÷ 2.

Worked through at the directory’s most common rate band ($100–$149/hr, published by 46.6% of the 1,143 agencies that publish a rate), a twenty-hour engagement runs around $2,500 a month, or $30,000 a year:

Gross profit per customer Extra customers needed per year
$200 150
$1,000 30
$5,000 6
$20,000 2

Now look at that against your current volume. If you close 400 customers a year at $1,000 profit each, thirty more is a 7.5% lift — plausible. If you close 40 a year at $200 profit each, 150 more is not a marketing plan, it is a different company.

This is the calculation that should decide whether to proceed, and it is the one most buyers skip in favour of asking the agency for a forecast.

Why the arithmetic favours some businesses so heavily

The break-even table makes something visible that “is SEO worth it” arguments usually obscure: the channel’s viability is mostly a property of your unit economics, not of the channel.

High-value, low-volume businesses — professional services, B2B, considered purchases, most trades — need a small number of additional customers to justify substantial spend. For them SEO is often the cheapest acquisition channel available, because the same engagement cost is amortised over very few, very valuable wins.

Low-value, high-volume businesses need search volume that many markets simply do not contain. A local business selling a $30 product with $10 of margin needs three thousand extra transactions a year to clear a $30,000 engagement, from a local search market that may not have three thousand relevant searches in it.

Neither of those is a statement about whether SEO works. It is arithmetic, and it is knowable before you spend anything.

The cost side is knowable — use it

There is an asymmetry worth exploiting here. The return is genuinely uncertain. The cost is not, and it is more variable than most buyers realise.

Our directory shows a wide spread in what agencies publish: 25% of firms that publish a rate are under $100/hr, 46.6% are at $100–$149, and 26.5% are at $150–$199. The same twenty-hour engagement can therefore be a $30,000 or a $45,000 annual commitment depending only on who you hire — which moves your break-even by half.

Two consequences:

  • Scope in hours and compare on hours. Two proposals at the same monthly price frequently represent very different amounts of work. The break-even you calculated is only valid against a known hour count.
  • Rate band is a lever on the break-even, not just a quality signal. If the arithmetic is marginal at $150/hr and comfortable at $100/hr, that is a real finding — provided you have established what the lower rate implies about who does the work. (What SEO costs covers what published rates do and do not tell you.)

Measuring it once you start

Break-even tells you whether to begin. Measurement tells you whether it worked, and it collapses without one thing: a baseline recorded before work starts.

Capture, in writing, on day one:

  • Organic sessions, split by landing page, for the pages the engagement targets.
  • Conversions from organic — forms, calls, transactions — however imperfectly you can currently track them.
  • Current positions and impressions for the target terms.
  • Your current close rate and average customer value.

Then measure the same things monthly against that record. Attribution will not be clean; organic search is upstream of a lot of behaviour that later gets credited elsewhere. It does not need to be clean. It needs to be consistent, and it needs to have started before the work did. A baseline captured in month six is not a baseline.

The trap in judging it too early

An SEO engagement’s cash profile is front-loaded cost and back-loaded return. Cost begins in month one; return, realistically, begins around month six (the checkpoints worth watching before then are leading indicators, not revenue).

That means a twelve-month cash view of a compounding channel will frequently show a loss at exactly the point the channel is starting to work. The engagements that get cancelled at month eleven are often the ones that were about to pay.

The correction is to judge it on two horizons at once: leading indicators monthly (pages indexed, impressions, new terms appearing), and cumulative break-even annually. The first tells you whether the work is happening. The second tells you whether it was worth it. Neither question is answerable with the other one’s evidence.

The short version

Do not ask an agency to forecast your return; nobody can. Calculate your break-even in customers per year, compare it against your current volume, and let that decide. Then scope in hours so the number stays valid, record a baseline before anyone starts, and judge on leading indicators monthly and cumulative return annually.

If the break-even looks achievable, the next question is who to hire — start with the SEO company rankings, or local and ecommerce if the work is specialised, and run each shortlist through the vetting checklist.

Common questions

What is a good ROI for SEO?

Any agency quoting you a specific multiple is quoting a number from a case study about a different business. The useful figure is your own break-even — annual cost divided by the profit you make on one customer — because it tells you how many extra customers a year the engagement must produce. If that number looks obviously achievable given your current volume, the case is strong. If it requires doubling the business, it is not.

Is SEO worth it for a small business?

It depends almost entirely on the value of one customer. A business with a $4,000 average project value needs a handful of extra jobs a year to justify a $2,000 monthly retainer. A business selling a $30 product needs a volume that most local search markets cannot supply. Same channel, same cost, completely different answer — and the arithmetic tells you which case you are in before you spend anything.

How do I measure SEO ROI?

Track organic sessions to the pages the work targets, the conversions from those sessions, and the closed value of those conversions — measured against a baseline you record before work starts. The baseline is the part people skip, and without it every later number is arguable. Attribution will never be perfect; a consistent, imperfect measure beats an accurate one you started six months late.

How long before SEO pays for itself?

Longer than the results start appearing, because the early months are cost with no return against them. If an engagement begins producing at month six and runs at a level that covers its monthly cost from month nine, cumulative break-even typically lands somewhere in the second year. Judging a compounding channel on a twelve-month cash view will usually reject it just before it starts working.

Where to go next

The research behind this post lives in the rankings. These are the lists it draws on.