Nearly 60% of U.S. Google searches ended without a click to any external website in 2024. On mobile, that number topped 75%. And by early 2026, zero-click searches hit 68% overall. Those numbers should haunt anyone building a traffic forecast off ranking improvements alone.
The old math was simple: fix the technical issue, climb a few positions, multiply by the keyword's search volume and an estimated CTR. Ship the spreadsheet. But that math assumed clicks followed rankings in a predictable ratio. They don't anymore.
The ceiling moved while you weren't looking
AI Overviews now appear in about 7.6% of Google searches. That sounds small until you see the damage: Ahrefs data shows AI Overviews are associated with a 58% lower average CTR for the top-ranking page. On queries where AI Overviews showed, organic CTR fell to 0.61% in one September 2025 data set. Zero point six one percent. For informational queries where your content answers the question directly, Google may already be serving your answer without sending anyone to your site.
Gartner-based coverage projects traditional search volume could decline 25% by 2026. Some websites reported 20% to 40% traffic drops in 2025, tied largely to organic search erosion. This isn't theoretical risk. It's the operating environment.
So when someone on the team says "fixing our canonical tags and internal linking should get us X thousand more visits," the right response is: maybe, but the X depends on which queries you're fixing for, and whether those SERPs still send clicks at all.
A forecasting framework that accounts for reality
Here's the approach that survives scrutiny in a planning meeting. Five steps, no magic.
Step 1: Pull a 90-day baseline. For every URL or URL group you're fixing, grab the last 90 days of clicks and impressions from Search Console. Note the trend direction. Flat? Declining? Seasonal? This is your starting point, not some third-party tool's modeled estimate.
Step 2: Find comparable past fixes on your own site. Did you fix crawl depth issues six months ago? Restructure a subdirectory? Add schema to product pages? Look at the before/after lift on those changes. Your own historical data is the most honest predictor of what similar fixes will produce. Third-party traffic estimators are directional at best; internal baselines are where the signal lives.
Step 3: Build three scenarios, not one number. Conservative, expected, aggressive. If past internal linking improvements lifted affected pages 8% to 15%, your conservative scenario uses 8%, expected uses 12%, aggressive uses 15%. Single-point estimates create false precision. Ranges communicate uncertainty honestly, which is what your CFO actually needs.
Step 4: Apply a zero-click and AI Overview discount. This is the layer most teams skip. For each query cluster, check whether AI Overviews or featured snippets dominate the SERP. If they do, discount your CTR assumption by 40% to 60% compared to a clean blue-link SERP. Informational queries get the heaviest discount. Comparison queries, use-case pages, and opinionated content tend to retain stronger click-through because the searcher needs to evaluate, not just get an answer.
Step 5: Translate clicks into pipeline, not just visits. For B2B SaaS, a traffic number alone isn't decision-useful. Take your incremental click estimate, multiply by your historical organic conversion rate to demo/trial/PQL, then multiply by average deal value and win rate. Now you have an estimated pipeline impact range. That's what gets budget approved.
Where the upside concentrates
Not all fixes are equally exposed to click suppression. Technical fixes (crawl issues, site speed, architecture) and bottom-of-funnel pages (comparisons, pricing, use cases) tend to produce faster, more measurable gains than broad top-of-funnel content plays. They also target queries where the searcher still needs to click through because a snippet can't substitute for the full evaluation.
Internal linking and schema improvements fall into the same category: high-leverage for discoverability, lower risk of the traffic getting absorbed by an AI answer box. When you're prioritizing your fix backlog, weight these higher.
One nuance worth noting: some teams have found that even as click volume drops, the remaining organic traffic converts better. The casual browsers get their answer in the SERP. The visitors who do click through have stronger intent. So your forecast should include a quality adjustment, not just a volume number.
The hypothesis to test
If we fix [specific technical/architectural issue] on [URL group], then organic clicks to those pages will increase [conservative]% to [aggressive]% within 60 days, because comparable past fixes on this site produced lifts in that range and these queries have low AI Overview exposure.
Success = incremental organic clicks within the scenario range. Guardrails = monitor impression share (are rankings actually moving?) and conversion rate (is traffic quality holding?). Stop-loss = if impressions rise but clicks don't after 45 days, the SERP environment has shifted and your CTR assumptions need revisiting.
Close the loop
That 68% zero-click number isn't going down. The teams that forecast well in this environment won't be the ones with the best SEO tools. They'll be the ones honest enough to discount their own optimism, anchor in first-party data, and connect every traffic estimate to a pipeline number that finance can actually use. The spreadsheet that says "it depends" with three scenarios will always beat the one that promises 50,000 visits with false confidence.
