Please note, for speed, I have used AI to generate this post based on research I have performed for my brightonSEO April 2026 talk.
One of the hardest questions in SEO is also one of the most reasonable:
How do I measure the ROI of my SEO investment?
The frustrating answer is that you usually cannot measure SEO ROI perfectly.
The more useful answer is that you can measure it credibly.
That distinction matters more now than it used to. Search is increasingly answer-led. A lot of SEO influence now happens before the click, without the click, or across messier journeys that do not fit neatly into last-click analytics. Research in the files I used for my BrightonSEO talk points to a growing share of discovery happening inside AI summaries and assistants, while direct AI referrals are real but still usually small enough that they rarely tell the full commercial story on their own.
So if you are still trying to answer SEO ROI with nothing but “organic sessions went up” or “organic revenue went down,” you are probably undercounting the value of the work.
Here is the better approach.
The core principle
You place a value on SEO by linking demand signals to commercial outcomes, not by pretending the demand signal itself is revenue.
The logic is:
SEO visibility → demand signal moves → commercial outcome moves
That is the model.
Not:
SEO visibility = revenue
That shortcut is where a lot of bad SEO ROI conversations start.
Why the old SEO ROI answer is getting weaker
The old model was much easier to explain.
You ranked.
People clicked.
They visited.
They converted.
That model is not completely dead. It is just no longer complete. In modern search, users can see your brand in SERP features, AI Overviews, AI assistants, comparison pages, local packs, or product surfaces, and only convert later through branded search, direct, paid, email, or another route. That makes classic last-click organic reporting a weaker proxy for actual SEO value. My talk frames this as the shift from deterministic SEO to probabilistic SEO: rankings and visibility still matter, but they now increase the likelihood of influence rather than guaranteeing one neat, proportional outcome.
The research I reviewed supports the same idea. Direct AI referral traffic can be measured when a click happens, but influence without a click is increasingly part of the story, especially as AI summaries reduce external clicking. That is why a lot of modern SEO ROI work has to be based on triangulation rather than fantasy-level certainty.
What SEO ROI should mean now
At its simplest, ROI still means:
SEO ROI = (value generated from SEO – SEO cost) / SEO cost
That part has not changed.
What has changed is how you estimate the value generated from SEO.
For some businesses, that value is easy to measure:
- ecommerce revenue from organic sessions
- demo requests from organic landing pages
- booked calls from local SEO
- pipeline from SEO-sourced leads
But for a lot of businesses, SEO also shapes demand upstream. It gets the brand seen, shortlisted, remembered, and returned to later. If you ignore that, you understate the commercial return.
So the better question is not just:
How much revenue came from last-click organic?
It is:
How much downstream value does SEO appear to have created or influenced?
Step 1: Pick the demand signals SEO is likely shaping
These are not the value themselves. They are the signals that suggest demand may be strengthening.
Common examples:
- branded search growth
- direct return visits
- repeat visits
- assisted conversions
- self-reported discovery
- AI / SERP visibility signals
- local visibility signals
- category-level search demand shifts
These work because they sit in the middle of the journey. They are often the first signs that visibility is starting to become preference.
If you saw my BrightonSEO talk, this is basically the middle layer of the framework:
- Presence = did we show up?
- Preference = did people lean toward us?
- Performance = did value follow?
Most ROI conversations skip straight from Presence to Performance and then wonder why the explanation feels weak.
Step 2: Connect the signal to a business outcome
This is the key step.
Do not stop at:
- “brand search went up”
- “direct visits increased”
- “assisted conversions were higher”
Ask:
- when branded search rises, do leads rise too?
- when direct return visits rise, does conversion rate rise?
- when assisted conversions rise, does closed revenue rise?
- when SEO visibility improves in a category, do qualified enquiries rise in that part of the business?
You are not looking for a fantasy of perfect proof.
You are looking for a credible relationship.
Step 3: Put a value on the downstream outcome
This is where the money comes in.
Example
Let’s say you see:
- 100 extra branded visits
- branded conversion rate = 8%
- average order value = £150
Then the estimate is:
- 100 × 8% = 8 conversions
- 8 × £150 = £1,200 estimated value
That does not mean branded search itself is abstractly “worth” £1,200.
It means the behavioural shift appears to be associated with about £1,200 of downstream commercial value.
That is a much more honest statement.
Three practical ways to estimate SEO-shaped ROI
Method 1: Use branded search as a demand proxy
This is one of the simplest ways to value SEO influence.
What you look for
- brand search impressions up
- brand search clicks up
- branded organic or direct sessions up
- branded traffic converts at a strong rate
- average lead value or order value known
Example
- extra 500 branded visits per month
- branded conversion rate = 6%
- average lead value = £200
Value estimate:
- 500 × 6% = 30 conversions
- 30 × £200 = £6,000
That gives you a line like:
One estimate of SEO-shaped demand value is the additional branded demand multiplied by its historical conversion value.
This is especially useful in a world where people often see a brand in AI answers or answer-led search experiences, then come back later via a branded query. Several of the case studies and pattern summaries in the research show AI visibility rising alongside branded search growth, direct interest, and downstream leads or sales.
Method 2: Use assisted conversions
This is often a better approach than pretending SEO either deserves 100% of the value or 0%.
Example
- SEO assisted 40 conversions
- average conversion value = £300
- attribution weighting = 50%
Estimated influenced value:
- 40 × £300 = £12,000 gross influenced value
- £12,000 × 50% = £6,000 weighted influenced value
This is more defendable because it acknowledges the role SEO played without claiming it closed the whole deal alone.
Method 3: Use incremental behaviour change
This is useful when SEO visibility rises and a particular behaviour rises after it.
Example
- AI / SERP presence improves
- direct return visits rise by 800
- returning visitor conversion rate = 4%
- average revenue per conversion = £120
Value estimate:
- 800 × 4% = 32 conversions
- 32 × £120 = £3,840
Again, not perfect proof.
But it is useful business evidence.
A useful formula
A clean version of the model is:
Demand-shaped value = incremental demand signal × downstream conversion rate × average commercial value
You can apply that to:
- branded visits
- direct return visits
- assisted journeys
- repeat visitors
- calls
- bookings
- demo requests
- pipeline entries
How to be conservative without becoming useless
If you want the estimate to be more defensible, make it more conservative.
You can do that by:
- using lower-end conversion rates
- using gross profit instead of revenue
- applying a weighting factor like 30–70%
- clearly labelling it as estimated influenced value
Example
Estimated influenced value = incremental branded visits × branded CVR × average profit × attribution weighting
That is a more cautious number, but often a more credible one in a stakeholder conversation.
What not to do
Avoid these mistakes:
1. Do not assign random monetary values to impressions alone
Impressions can matter. Visibility can matter. But a visibility metric is not the same thing as revenue.
2. Do not claim every rise in brand search is 100% SEO
Brand search can be shaped by SEO, PR, paid media, offline activity, product launches, and general market movement.
3. Do not mix weak signals with hard commercial claims
If your evidence is weak, say it is weak.
4. Do not act like correlation equals proof
A good SEO ROI model can still be useful without pretending it proves every pound perfectly.
The stronger move is:
- show the signal
- show the behavioural change
- show the commercial value attached to that behaviour
- be explicit that it is an estimate
How to talk about this internally
This bit matters.
Do not say:
We proved SEO caused exactly £X.
Say:
We estimate SEO-shaped demand created or influenced around £X of downstream value, based on changes in branded demand, return behaviour, assisted conversions, and their historical conversion value.
That sounds:
- more honest
- more mature
- more believable
- less like you are overclaiming
And in my experience, that is what makes SEO ROI conversations better.
What this looks like in practice
A modern SEO ROI view should probably include three things:
1. Direct SEO return
The easy-to-attribute bit:
- organic leads
- organic sales
- organic revenue
- organic pipeline
2. Demand-shaping signals
The middle-layer clues:
- branded search
- direct return
- assisted conversions
- repeat behaviour
- self-reported discovery
- AI visibility / citation patterns
3. Downstream business outcomes
The actual commercial score:
- qualified leads
- bookings
- revenue
- pipeline
- closed sales
That is the kind of reporting structure I argued for in my BrightonSEO talk, because it is much closer to how modern search influence actually works than a single traffic line.
The uncomfortable truth
The uncomfortable truth is that SEO ROI is no longer just a clean “organic revenue divided by SEO cost” conversation for many businesses.
That formula is still useful.
It is just incomplete.
The evidence from current AI and LLM research suggests:
- direct AI referrals are measurable but often small
- no-click influence is real but not cleanly attributable
- visibility, brand demand, and commercial outcomes often move together
- the strongest current standard is not perfect proof, but pattern-based business evidence
That does not mean SEO ROI is unknowable.
It means you need a better model for valuing it.
The best short answer
You place a value on SEO shaping demand by valuing the downstream behaviours that demand produces — like branded visits, return visits, assisted conversions, and leads — rather than pretending the visibility signal itself is revenue.
Final thought
If you want to track the ROI of SEO more honestly:
- measure the direct value where you can
- estimate the influenced value where you cannot
- separate signals from outcomes
- and stop pretending the click tells the whole story
That is not weaker measurement.
It is better interpretation.