The CTR Curve: Modeling Expected Click-Through Rate by Position

What a CTR Curve Is

A CTR curve is a model that maps search ranking position to an expected click-through rate. It answers a deceptively simple question: of all the people who see a particular position in the search results, what fraction click it? When those fractions are plotted from position 1 downward, they form a steeply descending curve — high at the top, falling quickly, and flattening into a long, thin tail toward the bottom of the page and beyond.

The shape is not linear. The drop from position 1 to position 2 is far larger than the drop from position 9 to position 10. This is the defining characteristic of search click behavior, and it is why ranking improvements near the top of page one are disproportionately valuable. Moving from position 4 to position 3 might add a few percentage points of CTR; moving from position 2 to position 1 can nearly double it.

Aggregate CTR curves are published regularly by firms that have access to large volumes of impression and click data. Advanced Web Ranking maintains a continuously updated organic CTR tool drawn from thousands of sites and millions of keywords; Backlinko analyzed four million Google search results; First Page Sage publishes an annual benchmark model; and SISTRIX has long reported CTR segmented by SERP layout. These studies disagree on the exact numbers — sometimes substantially — but they agree on the shape.

Understanding that shape, and learning to build a version of it calibrated to a specific site, is one of the most practical analytical exercises in search optimization. It turns a vague sense that a page "should get more traffic" into a measurable expectation against which actual performance can be compared.

The Power-Law Decay Shape

CTR by position follows what is best described as a power-law decay. In a power-law distribution, the value at each rank is a fraction of the value at the previous rank, so the absolute drops shrink as you descend even though the proportional drops remain large. Applied to search, this means the first result captures a dominant share of attention, the second captures a meaningful but much smaller share, and each subsequent position captures progressively less.

SISTRIX's analysis, based on roughly 80 million keywords, illustrates the curve clearly. On a SERP with no special features, position 1 earns approximately 34.2% of clicks, position 2 around 17.1%, and position 3 around 11.4%. Across all SERP layouts combined, SISTRIX reports an average position-1 CTR of about 28.5%, because the presence of ads, snippets, and other features pulls clicks away from the organic results. The gap between positions 1 and 2 is larger than the gap between positions 2 and 10 — the hallmark of power-law decay.

The curve is also remarkably consistent in form across data sources even when the absolute values differ. First Page Sage's 2026 benchmark models position 1 as high as 39.8% on clean SERPs and reports that the top three organic results together take roughly 68.7% of clicks, with the first result alone receiving more clicks than results three through ten combined. SISTRIX's pure-organic figure sits lower at 34.2%, and seoClarity's dataset of more than 750 billion impressions — the largest of the commonly cited studies — places desktop position 1 nearer 8.17% because it measures a far broader and more feature-heavy population of SERPs. The numbers span a wide range, but each describes the same descending power-law profile. The general benchmarks are explored in more depth in the page on CTR by Google search position.

A Modeled CTR Curve

The table below presents a modeled composite curve for positions 1 through 10, drawn from the pure-organic and aggregate figures reported across the major studies. It is intended as an illustrative baseline, not a precise prediction for any individual site. Two columns are shown: a pure-organic curve (SERPs with no ads, snippets, or AI Overviews) and an all-layouts average that reflects the lower CTRs typical of feature-rich results pages.

The final column is the most useful for intuition: it expresses each position's CTR as a share of position 1's. By position 5, a result earns under one-fifth of the clicks the top result earns. By the bottom of page one, that share has fallen below 10%. This is why the page on zero-click searches and the displacement caused by SERP features matter so much — anything that pushes the organic block down the page moves results onto the steep part of the curve where small positional losses translate into large traffic losses.

These figures should be treated as a starting point, not as ground truth. The most reliable curve for any given site is one built from that site's own data, which is the subject of the next section.

Building Your Own Curve from Search Console

Generic CTR curves are useful for orientation, but they describe an average of many sites with different brands, audiences, and SERP layouts. A site-specific expected-CTR curve, derived from Google Search Console, is far more accurate for diagnostic work because it controls for those variables automatically. The procedure is straightforward.

Step 1: Export query-level performance data

In Search Console, open the Performance report and export query-level data with four metrics enabled: query, clicks, impressions, and average position. A longer date range produces more stable averages; three to six months is typical. Position-level granularity matters, so export at the query level rather than relying on the page-level summary.

Step 2: Bucket rows by position

Round each row's average position to the nearest whole number and group the rows into position buckets — all queries averaging position 1, all averaging position 2, and so on. Buckets beyond position 10 can be collapsed, since page-two traffic is negligible: only about 0.63% of searchers click any page-two result, and roughly 75% never scroll past page one.

Step 3: Compute click-weighted CTR per bucket

For each position bucket, divide total clicks by total impressions to get a click-weighted CTR. A simple average of per-query CTRs would over-weight low-impression queries and distort the curve; weighting by impressions produces a curve that reflects where the traffic actually is. The resulting table of position-to-CTR values is the site's own expected-CTR curve.

Once built, the curve becomes a reusable yardstick. Every page can be measured against the expected CTR for its position, which is where the real diagnostic value emerges.

Expected vs. Actual CTR as a Diagnostic

The single most useful application of a CTR curve is the expected-versus-actual comparison. For any page, look up its average position, read the expected CTR off the curve, and compare it to the page's actual CTR. The difference isolates the snippet variable from the ranking variable — it tells you how well a result converts impressions into clicks, independent of how high it ranks.

Three diagnostic patterns emerge:

• Actual well below expected (under-performing). The page ranks but is not earning its share of clicks. This usually points to a weak title tag, an uncompelling meta description, or a mismatch between the snippet and the searcher's intent. A page ranking third with a 4% CTR against an 8-11% benchmark is leaving clicks on the table that a better snippet could recover.
• Actual well above expected (over-performing). The page earns more clicks than its position predicts. Its title and snippet are doing something right — strong messaging, a compelling promise, structured-data enhancements, or brand recognition. These pages are templates worth studying and replicating elsewhere.
• Actual near expected (on-curve). The snippet is performing normally for its position. Here the lever is ranking improvement rather than snippet optimization; CTR is already where the curve predicts.

This framing reorders an optimization backlog efficiently. The fastest wins come from the under-performing group: pages that already rank well but convert poorly, where rewriting a title tag or meta description can lift CTR without any change in position. The deeper, slower work of moving up the rankings is reserved for pages that are already on-curve. Practical tactics for closing the gap are covered in the guide to improving organic CTR.

It is worth noting that benchmarks vary widely by sector. A 3% CTR at position 2 might be normal in one vertical and alarming in another, which is why the page on organic CTR by industry is a useful companion when interpreting a site's gaps. A site-specific curve sidesteps much of this variance by construction, since it is calibrated to that site's own queries.

How Position-Normalization Plausibly Underlies NavBoost

The expected-versus-actual logic is not only a practitioner's diagnostic; it plausibly mirrors how Google itself interprets click data. NavBoost, the click-based re-ranking system that Google VP of Search Pandu Nayak described under oath as one of the company's "most important" ranking signals, evaluates how users interact with results. A naive system that simply counted clicks would be hopelessly biased, because higher positions receive more clicks by virtue of position alone. Any click-based ranking system must therefore separate the clicks a result earns from the clicks its position would generate regardless of quality.

This is conceptually the same problem that an expected-CTR curve solves. To judge whether a result is genuinely satisfying searchers, the system needs a position-normalized baseline — an expectation of how many clicks a result in that slot ought to receive — against which the result's actual click behavior can be compared. A result that earns more clicks than expected for its position is a candidate for promotion; one that earns fewer is a candidate for demotion.

The 2024 Google API leak, disclosed publicly by Rand Fishkin and analyzed technically by Mike King, exposed click fields consistent with this kind of relative evaluation. The documented types — goodClicks (satisfied clicks), badClicks (pogo-sticking back to the SERP), and lastLongestClicks (the final, longest-dwell click in a session, the strongest positive signal) — describe quality-weighted clicks rather than raw counts. Combined with the squashing function that compresses extreme volumes, the architecture points toward a system that asks not "how many clicks did this result get?" but "did this result earn more or fewer clicks than its position warrants?"

Read this way, the practitioner's CTR curve and Google's internal click evaluation are two views of the same underlying idea. When a page beats its expected CTR, it is sending the kind of relative signal that a position-aware re-ranking system is built to reward. The relationship between click behavior and rankings, and the limits of what the evidence supports, is examined in detail on the page covering whether CTR affects SEO rankings.

Keeping Curves Current

A CTR curve is a snapshot of searcher behavior at a point in time, and that behavior is shifting. The most significant recent change is the spread of AI Overviews and other generative SERP features, which intercept clicks before they reach the organic results and have pushed the top of the curve downward.

GrowthSRC's 2025 study of 200,000 keywords quantified the shift. Position 1 CTR fell from 28% to 19% year over year — a 32% decline — and position 2 dropped from 20.83% to 12.60%, a 39% decline. Across positions 1 through 5 the average decline was about 17.92%. Counterintuitively, positions 6 through 10 rose roughly 30.63%, which analysts attribute to users scrolling past AI Overviews to reach traditional organic results. The net effect is a curve that is both lower at the top and somewhat flatter overall.

The practical lesson is that a CTR curve has a shelf life. A curve built in 2023 overstates the clicks available at the top of the page in 2026, and using a stale curve as the "expected" baseline will make on-curve pages look like under-performers. Rebuilding the curve from fresh Search Console data every quarter, and segmenting it by whether AI Overviews appear for a query, keeps the diagnostic honest. The broader impact of these features is detailed in the analysis of how AI Overviews changed CTR.

The collapse of zero-click context also reframes what a "good" CTR even means. With 58.5% of US searches ending without a click according to Semrush's 2025 data, and 83% zero-click when AI Overviews are present, the pool of available clicks is shrinking. A curve that does not account for that shrinking pool will systematically overestimate expected CTR, and the gap analysis built on top of it will be off in the same direction.

Frequently Asked Questions

Further Reading

• CTR by Google Search Position — the underlying benchmark data behind any modeled curve, with figures broken out by study and SERP layout.
• How to Improve Your Organic CTR — practical tactics for closing the gap when a page's actual CTR falls below its expected value.
• Does CTR Affect SEO Rankings? — the evidence on whether click signals influence rankings, and the limits of that evidence.
• Organic CTR by Industry — sector benchmarks that contextualize whether a given CTR gap is normal or alarming.
• What is NavBoost? — the foundational overview of Google's click-based re-ranking system that position-normalized click analysis plausibly feeds.
• How AI Overviews Changed CTR — why the curve has shifted downward and why it now requires regular rebuilding.