Marketing Statistics Library 2026: Sourced Numbers by Topic

Twelve statistics pages, every number sourced and copyable — the citable figures for AI search, email, paid media, SEO, conversion, attribution and the rest of the growth stack, curated from the largest published datasets and our own annual reports. Each stat ships with its source and year attached, plus a one-click copy button so the attribution travels with the number.

The channel numbers

The money side of the stack — what advertising costs, returns, and where budgets are moving:

• Paid media statistics — CPC/CPM medians, auction inflation, the creative-variance findings
• Video advertising statistics — YouTube and short-form pricing and behavior
• Retail media statistics — the fastest-growing major channel, quantified
• Email marketing statistics — the $36–42 ROI story, placement rates, complaint ceilings
• SMS marketing statistics — 6–8x email click-through, priced accordingly

The visibility numbers

Where discovery is moving and what it takes to be found:

• AI search statistics — assistant adoption, zero-click growth, citation behavior
• SEO statistics — organic click curves, AI Overviews impact, Core Web Vitals research

The AI-search page distills our State of AI Search report; if the numbers convince you the shift is real, the GEO definition and our AI Visibility Checker are the practical next steps.

The conversion & measurement numbers

What happens after the click, and whether you can trust what your dashboard says about it:

• Ecommerce conversion statistics — CVR medians, cart abandonment, the speed-revenue research
• Landing page conversion statistics — page-level distributions and testing win rates
• Marketing attribution statistics — signal loss, server-side recovery rates, the MMM resurgence
• AI in marketing statistics — adoption by team size and use case
• MarTech statistics — stack sizes, utilization rates, consolidation

Why a stats library instead of one giant roundup

Because context is what makes a number usable. A "70% cart abandonment" stat means something different to a furniture brand than to a grocery subscription, and pages scoped by topic can carry that nuance — each one explains what its numbers include, what they exclude, and where the honest uncertainty lies. The industry benchmark library then localizes the same figures by vertical, and the glossary pins down the definitions underneath them.

There's also a distribution reason, worth being transparent about since it's the same play we recommend to clients: answer engines cite pages that state figures cleanly with sources attached. Publishing sourced statistics is among the most reliable ways to earn citations from ChatGPT, Perplexity and AI Overviews — the mechanics are in our guide to getting cited, and it's a core motion of our AI search optimization practice.

Using the numbers well

Three habits separate teams that use statistics from teams that decorate slides with them. Anchor decisions to your own data and use market stats as the prior — the statistics pages tell you what's normal, your analytics tell you what's true. Check the year before citing; a 2019 number in a 2026 deck erodes trust with anyone who knows the field. And when a stat surprises you, read its source note before repeating it — the definition usually explains the surprise.

When a number sends you hunting for the deeper how, the guides collection covers execution, and our free tools let you measure most of these metrics on your own properties in minutes.

The dozen numbers that anchor everything

If you only carry twelve figures out of the whole library, carry these. Google Search: $4.66 median CPC and 6.42% CTR (WordStream/LocalIQ). Meta: $14–15 blended CPM (Revealbot/Varos). Auction inflation: roughly 10% a year. Email: $36–42 returned per dollar (Litmus/DMA UK), ~83% global inbox placement (Validity), 0.3% spam-complaint ceiling at Gmail and Yahoo with 0.1% the stated target, and 25–30% of ecommerce revenue attributed to the channel (Klaviyo). Web: 0.1 seconds of mobile speed worth +8.4% retail conversions (Deloitte and Google), mobile bounce probability up 32% between one-second and three-second loads (Google/SOASTA), ~70% cart abandonment, 2–3% ecommerce conversion. And the meta-number that contextualizes all of them: a 2–4x spread between average and top-quartile accounts on the same channel.

Each appears on its topic page with the full study context, alongside the interactive comparators that show where your own numbers land against them.

The provenance problem in marketing statistics

A surprising share of widely-cited marketing stats are orphans — numbers whose original study died years ago while the figure kept circulating. The famous "it takes 7 touches to make a sale" has no traceable modern source; various "attention span" statistics trace to a misread infographic; email ROI numbers get quoted without the year or methodology that made them meaningful. This library's rule is simple: every number names its source and year inline, directional estimates are labeled directional, and when the honest answer is a range, we publish the range. That policy costs us some impressive-sounding precision and buys the thing that matters — you can repeat any figure here in front of the most skeptical person in your company and survive the follow-up question. When a number you need is missing from the library, that's usually because no defensible source exists for it yet; we'd rather leave a gap than invent a decimal point, and the gaps close as the underlying studies publish.