How to Measure and Prove SEO ROI When Clicks Are Declining: The Complete Attribution & Reporting Guide

Introduction: Traffic Is Down. That Doesn’t Mean SEO Stopped Working.

If you’ve sat in a meeting recently where someone pointed at a falling organic traffic line and asked what exactly the SEO budget is paying for, you’re not alone. It’s the conversation we’re having most often with clients right now, and it’s a fair question.

Here’s the uncomfortable truth: the way most businesses measure SEO was built for a search engine that no longer exists. Rankings and sessions made sense when a page one position reliably turned into a click. Today, a growing share of searches are answered on the results page itself, inside an AI Overview, or in a conversation with ChatGPT, Claude or Perplexity that never touches your website at all.

That doesn’t mean SEO has stopped creating value. It means a lot of that value now happens before the click, or instead of it, and your reporting hasn’t caught up.

This guide is the hub for our Measurement & ROI pillar. It’s written for marketing managers, business owners and anyone who has to defend an SEO or content budget to a CEO, a board or a finance team. We’ll cover:

• Why traffic is falling, and why that isn’t automatically bad news
• The new scoreboard: the metrics that matter beyond rankings and sessions
• How to set up GA4 and Search Console so they measure what matters
• Attribution models, and why SEO is systematically undervalued by most of them
• How to track “dark traffic” and AI-driven brand impressions
• How to calculate SEO ROI with a formula you can actually defend
• How to build reports that non-marketers understand, and set realistic KPIs and budgets for the next 12 months

When we covered this topic in our BrightonSEO 2026 recap on falling traffic, measurement was one of the loudest themes of the whole conference. This guide goes further, turning those ideas into a practical framework you can put in place this quarter.

Why Traffic Is Falling, and Why That Isn’t Automatically Bad News

Before you can prove SEO’s value, you need to be able to explain what’s happening to the traffic line. Otherwise, every report turns into a defence of a number that’s being pulled down by forces outside your control.

The results page now answers the question for you

Zero-click search isn’t new, but AI has accelerated it considerably. SparkToro’s 2024 zero-click search study, built on clickstream data from Datos, found that roughly 58.5% of US Google searches and 59.7% of EU searches ended without a click, and that fewer than 400 in every 1,000 searches sent a click to the open web.

AI Overviews have pushed that further. Pew Research Center tracked the real browsing behaviour of 900 US adults in March 2025 and found that users clicked a traditional search result on 8% of visits when an AI summary appeared, compared with 15% when it didn’t. Only 1% of visits involved a click on a link inside the summary itself.

Ahrefs’ analysis of 300,000 keywords reached a similar conclusion from a different angle: an AI Overview on the page was associated with a 34.5% lower click-through rate for the top-ranking result, compared with similar informational keywords that didn’t trigger one.

So why might this be fine?

Because fewer clicks and less value are not the same thing. Three things are happening at once:

• The clicks you still get are better qualified. Google’s own guidance on AI features states that clicks from results pages with AI Overviews are “higher quality”, meaning users are more likely to spend more time on the site. The people who used to click, skim and bounce are now getting their quick answer on the results page.
• AI referrals convert strongly. A Semrush study of AI search traffic found the average visitor arriving from an AI search source such as ChatGPT was 4.4 times as valuable as the average traditional organic visitor, based on conversion rate. Nick Beck highlighted this figure at BrightonSEO 2026, and it landed with the room. Semrush is clear that its projections are extrapolations, so treat this as directional rather than a guaranteed uplift for your site.
• Influence is happening without a visit. When your brand is cited in an AI Overview or recommended by ChatGPT, that shapes a buying decision even if the person searches your brand name directly a week later, or simply picks up the phone.

At BrightonSEO 2026, Laura McInley put it neatly in her session “No Clicks? No Problem!”Declining traffic is often a reporting problem, not a performance problem. The rest of this guide is about fixing the reporting.

Paul’s take: “This is the stat I bring up in almost every client conversation now. It reframes the whole anxiety about ‘traffic is down’ into ‘who is actually visiting, and what are they doing when they get here?’ Once we explain the conversion rate difference, the traffic number stops being the scary headline.”

Diagnosing whether your own drop is AI-driven, seasonal, technical or a genuine problem is its own job. Our scheduled cluster guide, Why Your Traffic Is Dropping (And How to Tell If It’s a Problem), will walk through it step by step, so make sure you follow us on LinkedIn so you don’t miss that later post.

The New Scoreboard: Metrics That Matter Beyond Rankings and Sessions

If traffic can no longer be the headline, something else has to be. The answer isn’t to throw traffic out. It’s to put it in its proper place, as one input alongside visibility, brand demand and, above all, commercial outcomes.

The framework we find most useful comes from Laura McInley’s BrightonSEO 2026 session. She groups SEO measurement into three layers: Presence, Preference and Performance. We’ve adapted it slightly to reflect how our clients’ reporting actually works.

The order matters. Presence metrics move first and tell you whether the work is landing. Performance metrics move last and tell you whether it’s paying. Reporting only the last layer means months of silence before anything shows up. Reporting only the first means nobody in finance ever believes you.

Search Console’s Generative AI performance report

One of the most useful new Presence metrics comes straight from Google. Search Console’s Generative AI performance report shows how often links to your site appeared in Google’s generative AI features, currently AI Overviews and AI Mode, and which pages earned those impressions. You can break the data down by page, country, device and date, and export it.

Google says it rolled these insights out to all websites worldwide as of 31 August 2026, although the same documentation notes that access depends on a site having enough impressions in generative AI features, so smaller sites may not see it straight away.

It’s important to be clear about what it measures. The report shows impressions only, with no clicks or click-through rate. That makes it a Presence metric, not a Performance one, and that’s exactly where it earns its place.

Clustering the data to show what supporting content is worth

The real value comes when you stop reading the report page by page and start grouping it. By tagging every URL with its topic cluster and content type (service page, pillar guide, blog or news article), you can show:

• Which topic clusters are earning visibility in AI answers, and whether that’s growing month on month
• How much of that visibility comes from supporting blog and news content. These articles rarely get last-click credit for a lead, but if Google is surfacing them in AI Overviews and AI Mode, they’re building brand awareness with the people you want to reach, long before they’re ready to enquire
• Whether AI visibility for a cluster is followed by growth in branded search and enquiries for the related service

A simple monthly process:

1. Export the page table from the Generative AI performance report.
2. Tag each URL with its cluster and content type, using the same pillar tags as your GA4 content groups.
3. Total impressions by cluster and content type, and track the trend month on month.
4. Report each cluster’s AI impressions alongside its branded search trend and key events, so Presence can be read against Preference and Performance.

This is a metric we’re now building into the Presence layer of our own client reporting. It finally gives blog and news content a measurable job, rather than leaving it as a cost that’s hard to justify on a last-click report.

Paul’s take: “Blog and news content has always been the hardest thing to justify on a last-click report. Being able to show a client that their supporting articles are the pages Google is surfacing in AI Overviews changes that conversation completely. It gives that content a clear role: building awareness with buyers before they’re ready to get in touch.”

Branded search is now much easier to measure

One of the most useful additions to the new scoreboard is branded search demand, because it captures influence that happens away from your website. When your brand is cited in AI answers, covered in the press or recommended in a Reddit thread, a share of those people will later search for you by name.

Until recently, separating branded from non-branded queries in Search Console meant maintaining your own regex filters. In early 2026 Google made a branded queries filter available to all eligible Search Console properties. It automatically classifies queries containing your brand name, common misspellings and brand-related products or services. There are limits worth knowing: it isn’t available for sub-properties or low-impression sites, it can’t be customised, and it isn’t applied retroactively, so older comparisons still need your own filters.

Speak the language of the people signing off the budget

Brittany Deller’s BrightonSEO session, “How to Make Measurement Impossible to Ignore”, made a point every SEO should keep front of mind: executives respond to revenue, cost and risk, not crawl stats. Every metric on your scoreboard should be able to answer the question “so what does that mean for the business?” If it can’t, it belongs in the appendix, not the summary.

Paul’s take: “What I liked about Laura’s approach is that it doesn’t throw out traffic reporting altogether – it puts it in its proper place. Traffic is one input into a bigger picture, not the whole picture. Clients generally respond really well to this once it’s explained properly, because it actually gives them more confidence, not less.”

For a deeper look at each of these metrics, see our cluster guide, Beyond Rankings: The New SEO Metrics That Actually Matter. For the thinking behind the shift from rankings to influence, our earlier post From Rankings to Influence: Reframing SEO Success in 2026 is a useful companion read.

Attribution Models, and Why SEO Is Systematically Undervalued

Attribution is where most SEO value quietly disappears. It’s rarely anyone’s fault. It’s just that the default ways of assigning credit were designed around the last thing someone clicked, and SEO’s biggest contribution usually happens much earlier.

How the main models treat organic search

That last point matters. According to Google’s own Analytics documentation, GA4 now offers three models in its attribution reports: data-driven, paid and organic last click, and Google paid channels last click. First click, linear, time decay and position-based models were retired in November 2023. Data-driven attribution is a genuine improvement, but it can’t give credit for an AI Overview someone read, a ChatGPT recommendation that didn’t pass a referrer, or a blog post they read on their phone before converting on a laptop without signing in.

At BrightonSEO 2026, James Yorke’s session “SEO Has a Measurement Crisis” made the case bluntly: you can’t look at SEO traffic in isolation any more, especially through a last-click lens. The practical response is to report organic search alongside overall demand, and to add evidence from outside your analytics platform.

Paul’s take: “James’s session stuck with me because he wasn’t dramatic about it. He simply said that last-click attribution is increasingly meaningless because so much of the value SEO creates now happens before the click, or instead of the click. That’s not a reason to panic. It’s a reason to change what we measure.”

Add self-reported attribution

The simplest fix for invisible influence is to ask. A required “How did you hear about us?” field on enquiry and checkout forms, with a free-text option, regularly surfaces answers your analytics can’t: “I asked ChatGPT for the best agencies in Auckland”, “saw you mentioned in an article”, “Googled it and read your guide”. It’s imperfect, because people misremember, but set against your analytics data, it gives a far more honest picture than either source alone.

Paul’s take: “When we add a free-text ‘how did you hear about us?’ field for clients, the answers are often the most revealing data in the whole report. We regularly see enquiries that GA4 files under Direct where the person tells us they found the business through an AI recommendation or a guide they read weeks earlier.”

Setting up GA4 and Search Console to measure what matters

None of this works if the foundations are wrong. The essentials:

1. Define your key events properly. In GA4, any event can be marked as a key event, meaning an action that’s particularly important to your business. Form submissions, calls, bookings, quote requests and purchases should all be key events. Newsletter sign-ups and content downloads can be too, but report them separately from revenue-generating actions.
2. Give AI traffic its own line. GA4’s default channel group now includes an AI Assistant channel for visits from sources such as ChatGPT, Gemini, DeepSeek, Copilot and Grok. Check that it’s populating in your property. Note that it explicitly excludes Google’s AI Overviews and AI Mode, which are reported within organic search.
3. Connect Search Console to GA4 so query and landing page data sit alongside behaviour and key events.
4. Group pages by topic or pillar. Content groups in GA4 and page filters in Search Console let you report on a whole topic cluster rather than page by page, which is where the real trends show up.
5. Pass lead source into your CRM. Capture the landing page, source and medium in hidden form fields so every lead in your CRM carries its origin through to closed revenue.
6. Annotate everything. Algorithm updates, site migrations, campaign launches and AI Overview expansions all explain changes in the data. Without notes, you’ll be guessing six months later.

Our upcoming cluster guides, Setting Up GA4 and Search Console to Measure What Matters and Attribution Models Explained: Last Click, First Click, and Multi-Touch, will cover each of these in more depth.

Tracking “Dark Traffic” and AI-Driven Brand Impressions

“Dark traffic” is the catch-all term for visits and influence your analytics can’t properly attribute. It isn’t new, but AI search has added a lot more of it. Some AI platforms don’t pass referral data consistently, some users copy a brand name out of an answer and search for it separately, and some never visit at all but still choose you later.

You won’t capture all of it. What you can do is build a set of signals that, taken together, show whether your AI visibility is growing and whether it’s turning into demand.

What you can measure directly

• AI Assistant referrals in GA4. Visits from ChatGPT, Gemini, Copilot and similar platforms now have their own default channel. Track sessions, engagement and key events from this channel month on month.
• Landing pages AI platforms send people to. Filter the AI Assistant channel by landing page. It often reveals which guides, comparison pages or service pages AI tools trust enough to link to.
• AI Overview and AI Mode Impressions by page. Search Console’s Generative AI performance report, covered in the scoreboard section above, shows which pages earned impressions in these features. The same impressions are also counted within the main Performance report’s Web search type, as Google confirms in its AI features documentation. The Generative AI report doesn’t show clicks, so pair it with the main Performance report: rising impressions on long, question-style queries alongside falling click-through rate is a strong sign of which topics are being answered in AI-heavy results.

What you have to measure indirectly

• An AI visibility prompt log. Build a fixed list of 20 to 50 prompts your buyers are likely to use, run them monthly across ChatGPT, Google AI Overviews and AI Mode, Perplexity, Claude and Gemini, and record whether your brand is mentioned, cited or recommended, and how accurately it’s described. Results vary between sessions and users, so look at trends over several months rather than single answers. Dedicated AI visibility tracking tools can automate some of this, but a well-designed manual log is still a credible starting point.
• Branded search trend. Use the Search Console branded queries filter to watch branded impressions and clicks over time. When AI visibility and brand coverage grow, branded demand should follow.
• Direct traffic and direct-to-key-event rate. Unexplained growth in direct traffic that converts well is often dark AI and word-of-mouth traffic in disguise.
• Self-reported attribution. As covered above, ask people how they found you and tag AI mentions in your CRM.

Paul’s take: “The prompt log is the piece clients are most sceptical about and then find most useful. It’s not perfect data, and we’re upfront about that. But when a board member asks whether ChatGPT recommends us over a competitor, a documented monthly record is a much better answer than a shrug.”

This is where measurement and our wider AI search work meet. If you want to improve what those prompts return, not just measure it, our Complete Guide to Generative Engine Optimisation (GEO) and our guide to getting your brand mentioned by ChatGPT, Claude and Google’s AI are the places to start.

Our upcoming cluster guide, How to Track ‘Dark Traffic’ and Brand Impressions From AI Platforms, will cover the GA4 configuration and workarounds in detail.

How to Calculate SEO ROI (With a Formula You Can Defend)

Once your measurement foundations are in place, the ROI calculation itself is simple. The hard part is being honest about the inputs.

SEO ROI =


Gross profit attributable to SEO − Total SEO investment


Total SEO investment


× 100

Three rules keep it credible:

• Use gross profit, not revenue. Finance teams will discount a revenue-based ROI straight away, and they’ll be right to.
• Include every cost. Agency or in-house time, content production, development work, tools and any digital PR that supports organic visibility.
• Show a range, not a single number. Present a conservative, an expected and an upper figure based on how much of the unattributed influence you include. It’s more honest, and it’s much harder to pick apart in a meeting.

A worked example

The figures below are illustrative, for a B2B services business investing NZ$3,000 a month (NZ$36,000 a year) in SEO and content. Each lead converts to a customer at 20%, with an average first-year customer value of NZ$8,000 and a 40% gross margin.

Even the conservative scenario is positive, and that’s the story that tends to land with a finance team: “On the strictest possible reading, SEO paid for itself. On the most realistic reading, it more than doubled its money.”

Two supporting measures worth adding

• Lifetime value. If customers typically stay for three years, first-year value understates the return considerably. Show it separately, rather than folding it into the headline ROI, so the core figure stays conservative.
• Paid search equivalent. Estimate what it would cost to buy your non-branded organic clicks through Google Ads at current cost per click. It isn’t ROI, but it’s a useful sense check for stakeholders who already understand paid media budgets.

Remember that SEO compounds. A guide published this year can keep generating leads for several years at little additional cost, while paid media stops the day the budget does. A 12-month ROI figure will almost always understate the long-term return, so say so.

For a content-specific version of this calculation, watch out for our cluster guide, How to Calculate the ROI of Content Marketing.

Building Reports Non-Marketers Actually Understand

The best measurement setup in the world won’t protect your budget if the report lands as a wall of charts nobody reads. A good SEO report for a CEO, board or finance team does three things: it tells them whether it’s working, why, and what happens next. Everything else is supporting detail.

The one-page executive view

The example figures are illustrative. Build this in Data Studio or your reporting tool of choice, keep it to one page, and put the detailed channel and page-level data behind it for anyone who wants to dig in.

Reporting cadence

• Monthly: the one-page view, with a short written commentary. Presence and preference metrics are the early warning system.
• Quarterly: a fuller review of performance against targets, the ROI range, and any reallocation of effort or budget.
• Annually: a strategic review of ROI, compounding value and next year’s investment case.

Setting Realistic KPIs and Budgets for the Next 12 Months

Most SEO disappointment comes from targets that were never realistic in the first place. The fix is to separate leading indicators from lagging ones, and to agree up front which you’ll judge the work on, and when.

These timeframes are a general guide rather than a promise. Competitive markets, site history and how much you invest all change the curve, which is exactly why the agreement matters more than the numbers.

On budget, start from the outcome rather than the channel. Work backwards from the revenue target, your average customer value, close rate and lead-to-customer journey, to how many qualified organic leads you need. Then compare what it would realistically cost to generate that demand through SEO, paid search and other channels. That turns the SEO budget from a line item someone has to defend into an investment case someone can evaluate.

Our upcoming cluster guides, Building an SEO Reporting Dashboard Your CEO Will Actually Understand, How to Present SEO Results to Non-Marketers, Setting Realistic SEO KPIs and Targets for the Next 12 Months, and Marketing Budget Planning: How Much Should You Really Spend on SEO?, will go deeper on each of these.

Five Measurement Mistakes That Undersell Your SEO

1. Reporting traffic as the headline. It invites the wrong conversation. Lead with outcomes and use traffic as context.
2. Relying on last-click attribution. It hands the credit for SEO-started journeys to whatever channel happened to close them.
3. Ignoring branded demand. Growth in people searching for you by name is one of the clearest signs your wider visibility is working.
4. Treating AI visibility as unmeasurable. It’s imperfect, not impossible. Search Console’s Generative AI performance report, a consistent prompt log and the GA4 AI Assistant channel give you a defensible trend.
5. Presenting a single ROI figure. A range is more honest and much harder to dismiss.

Supporting Guides in This Pillar

This hub page is the centre of our Measurement & ROI pillar. Each of the guides below goes deeper on one part of the framework, and we’ll link them here as they’re published:

• Why Your Traffic Is Dropping (And How to Tell If It’s a Problem)
• Beyond Rankings: The New SEO Metrics That Actually Matter
• How to Track ‘Dark Traffic’ and Brand Impressions From AI Platforms
• Setting Up GA4 and Search Console to Measure What Matters
• Building an SEO Reporting Dashboard Your CEO Will Actually Understand
• Attribution Models Explained: Last Click, First Click, and Multi-Touch
• How to Calculate the ROI of Content Marketing
• How to Present SEO Results to Non-Marketers (Templates Included)
• Setting Realistic SEO KPIs and Targets for the Next 12 Months
• Marketing Budget Planning: How Much Should You Really Spend on SEO?

Measurement sits on top of everything else in this series. If you haven’t read them yet, our Complete Technical SEO Checklist for 2027, Audience-First Content Playbook, GEO guide, and Modern Digital PR Playbook cover the work this guide helps you measure. And make sure you follow us on LinkedIn so you never miss a post.

Prove What Your SEO Is Actually Worth

Falling clicks don’t have to mean a falling budget. With the right metrics, honest attribution and reporting that speaks the language of the business, you can show exactly what organic search and AI visibility are contributing, even when the traffic line is heading the wrong way.

Book a free SEO ROI & Reporting Audit with Digital Hothouse. We’ll review your GA4 and Search Console setup, check how leads are being attributed from search and AI platforms, and show you where your current reporting is underselling your SEO. You’ll come away with a clear picture of what your organic search is really worth, and a reporting framework your leadership team will actually read.

We’d love to work with you. Let’s chat.