How Do I Measure SEO ROI?

Use (value of organic conversions − cost of SEO) ÷ cost of SEO × 100. The formula is trivial; the inputs are where every credible measurement is won or lost. Value conversions at gross margin rather than revenue, count internal time in the cost, and measure over quarters — SEO returns arrive months after the spend.

The formula, and why it is the easy part

SEO ROI % = (Value of organic conversions − Cost of SEO) ÷ Cost of SEO × 100

A campaign costing $30,000 that produced $75,000 of margin returned 150%. Nobody argues about the arithmetic. The arguments are always about the three numbers going in, and each one has a standard way of being quietly inflated.

The rest of this page is about getting those three right: what counts as a cost, what an organic conversion is actually worth, and over what window you are allowed to judge.

Step 1 — Count the full cost, not the invoice

Most SEO ROI figures are flattering because the cost side is understated. Include everything that would stop if you stopped doing SEO:

Cost	Commonly forgotten?
Agency retainer or freelancer fees	No
Content production — writing, editing, design	Sometimes
Tools (rank tracking, crawlers, backlink data)	Often
Link acquisition and outreach costs	Often
Internal time — your hours, a developer's, a founder's	Almost always
Technical work shipped for SEO reasons	Almost always

Internal time is the big one. If you spend eight hours a week on SEO, that is a real cost at your loaded hourly rate, and leaving it out can turn a mediocre return into an apparently excellent one. Price it honestly even though nobody invoices you for it.

One nuance on content: a blog post is closer to a capital expense than a monthly cost. It keeps producing for years. Some teams amortise content spend over 24 months rather than charging it all to the month it was published — reasonable, as long as you say which convention you used and stick to it. How to set a link building budget covers the same discipline on the link side.

Step 2 — Value the conversions properly

This is where the number is really made. Three cases, in ascending order of difficulty.

Ecommerce. You have transaction revenue attributed to organic sessions in your analytics. Use gross margin, not revenue. A 200% ROI calculated on revenue can be a loss once cost of goods is subtracted, and reporting revenue as if it were return is the single most common way SEO reporting misleads.

Lead generation. Work backwards from the pipeline:

Value per organic lead = close rate × average deal value × gross margin

If organic produces 100 leads a month, 8% close, average deal is $6,000, and margin is 70%, that is 100 × 0.08 × 6000 × 0.7 = $33,600 per month. The number is only as good as your close rate, so segment it — organic leads frequently close at a different rate to paid ones, and using a blended figure hides that.

SaaS or subscription. Use gross-margin LTV, not first-month revenue, and be conservative about it. Discount for churn, and if your LTV estimate is based on a small or young cohort, say so in the report rather than presenting it as fact.

For anything where you genuinely cannot connect organic to money — an early-stage site, a pure content play — measure a leading indicator instead and be explicit that it is a proxy, not a return. Qualified organic sessions to money pages is a reasonable one.

Step 3 — Fix the window before you look

SEO returns lag the spend by months. Judging a campaign on a calendar month compares this month's cost against last quarter's results, which is not a measurement of anything.

Two practices fix this:

Measure over rolling quarters, not months. Monthly reporting is for activity and leading indicators. ROI is a quarterly or annual question. How long do backlinks take to work sets out the realistic delay on the link side specifically.

Cohort by publish date. Instead of asking "what did SEO return in Q3," ask "what have the twelve pages published in Q1 returned since." This is far more honest, and far more useful — it tells you which kind of page is worth commissioning again.

The four attribution problems, and what to do about each

Every SEO ROI number is an estimate. Being upfront about why makes it more credible, not less.

1. Last-click undercounts organic badly. A visitor finds you via a blog post, leaves, and returns via a branded search or a direct visit three weeks later. Last-click attribution credits the second touch and gives organic nothing. Check assisted conversions or a data-driven attribution model before concluding organic is underperforming — the gap is often large.

2. Branded search steals the credit. As SEO works, more people search your brand name. Those sessions look like cheap wins but were often created by the content that introduced you. Segment branded from non-branded and treat branded growth as an output of the programme rather than baseline traffic.

3. Zero-click answers return nothing measurable. An AI Overview or featured snippet that answers the question fully produces no session and no data. This genuinely reduces measurable return without necessarily reducing business value, and it is getting more common — see is Google search traffic declining.

4. Your tools disagree with each other. Search Console and your analytics platform will report different numbers for the same period, always. Pick one as the source of truth for the ROI calculation and note which. The reasons they diverge are in why Search Console and Google Analytics don't match.

Four numbers that are not ROI

Reports fall apart under scrutiny because they present these as return.

Rankings. A position is not money. Plenty of ranking improvements happen on terms nobody buys from.
Traffic volume. Sessions without conversion data tell you about reach, not return.
"Traffic value" from an SEO tool. This is an estimate of what your organic traffic would cost in Google Ads at list CPC. It is a rough scale indicator and it is not revenue. Presenting it as ROI is the fastest way to lose a client's trust when they check it.
Domain authority or domain rating. Third-party scores invented by tool vendors. Useful for comparing prospects, meaningless as a business outcome.

The honest reporting structure — leading indicators monthly, outcomes quarterly — is laid out in how to report link building results to a client.

A worked example

A B2B services firm spends over twelve months:

Item	Cost
Freelance writer, 24 posts	$14,400
Tools	$1,800
Link building (outreach time + exchanges)	$3,600
Internal time, 5 hrs/week at $60 loaded	$15,600
Total	$35,400

In the following twelve months, non-branded organic produces 340 enquiries. Historic close rate on that source is 6%, average project is $9,500, gross margin 55%.

340 × 0.06 × 9500 × 0.55 = $106,590
ROI = (106,590 − 35,400) ÷ 35,400 × 100 = 201%

Two things make this credible rather than promotional. The cost line includes internal time, which is the largest single item. And the return is measured over the year following the spend, not alongside it.

Comparing SEO against paid

The comparison that actually matters to a finance team is against the alternative channel.

Paid search buys certainty and stops the day you stop paying. SEO buys an asset that compounds and takes two to three quarters to show up. A fair comparison prices that difference: what would it cost in Google Ads to buy the non-branded organic sessions you earned this quarter, at your actual account CPCs rather than list rates, and how does that compare to your SEO cost over the period that produced them?

That framing is more defensible than a raw ROI percentage, because it answers the question a CFO is really asking. The wider case, including when SEO genuinely is not worth it, is in is SEO a waste of money.

The one part of the cost stack you can compress without compressing results is link acquisition. Backlinkster replaces paid placements with one-for-one in-content swaps between real site owners, verified live by code — which takes the most expensive and least predictable line item out of the calculation. Five swaps a month free; plans from $19.

Frequently asked questions

What is the SEO ROI formula? (Value of organic conversions − Cost of SEO) ÷ Cost of SEO × 100. Value conversions at gross margin rather than revenue, and include internal time in the cost or the result will be flattering rather than accurate.

What is a good SEO ROI? It depends entirely on margin and deal size, so there is no universal figure. The more useful test is comparative: does SEO return more per dollar than your next best channel over the same period, accounting for the fact that SEO's returns arrive later and persist longer.

How long before SEO ROI is measurable? Typically six to twelve months for a new programme, because rankings, then traffic, then conversions each lag the work. Anything measured in the first quarter is a leading indicator, not a return.

Should I use revenue or profit to calculate SEO ROI? Gross margin. Revenue-based ROI overstates the return by whatever your cost of goods is, and it is the most common reason an SEO report does not survive a finance review.

How do I measure SEO ROI without ecommerce tracking? Work backwards from leads: close rate × average deal value × gross margin gives a value per organic lead. If you cannot connect organic to leads at all, report a clearly labelled proxy such as qualified sessions to money pages, and do not call it ROI.

Does traffic value from Ahrefs or Semrush count as ROI? No. It estimates what your organic traffic would cost at Google Ads list prices. It is a scale indicator, not revenue, and presenting it as return will not survive scrutiny.

How do I account for zero-click AI answers in SEO ROI? You largely cannot measure them, so note the gap explicitly rather than ignoring it. Track impressions and branded search volume as partial proxies for visibility that produces no session.

The bottom line

The formula was never the hard part. A defensible SEO ROI number comes from three unglamorous decisions: put internal time in the cost, value conversions at margin rather than revenue, and measure the quarter that followed the spend rather than the one containing it. Do those and the number will be lower than the ones in agency case studies — and it will survive the meeting where someone checks it.