Sources

This page lists the studies and statistics referenced on other EQwise pages. EQwise only cites statistics that can be traced back to their source: each study includes the title, publisher, date, and link. A statistic from another company is an order of magnitude and not a prediction for your site, because the research methodology, country, and time period vary from study to study.

Speed and Revenue

Vodafone: A 31% improvement in LCP increased sales by 8%

web.dev (Google), 17 March 2021

In an A/B test on Vodafone Italy’s website, a 31% faster LCPLCP LCP, which stands for Largest Contentful Paint, measures the time it takes for the largest visible element of a page to appear on the screen. (the time it takes for the majority of the page to become visible) resulted in an 8% increase in sales. This is just one company in one country, so it’s an example and not a standard.

Featured in: WordPress website development, Internet marketing FAQs, SEO and GEO audit, Calculate revenue based on load time

How Swappie increased mobile revenue by 42% by focusing on Core Web Vitals

web.dev (Google), 15 September 2021

After three months of work, the LCP was 55% lower and revenue from mobile visitors was 42% higher. At the same time, the CLS (-91%) and the FID (-90%) also improved, so the revenue gain isn’t solely due to the LCP. This is a pre- and post-measurement study, not a test with a control group.

Featured in: WordPress website development, Internet marketing FAQs, SEO and GEO audit, Calculate revenue based on load time

The Need for Mobile Speed

DoubleClick by Google, 2016 (data from March 2016)

Of all mobile visits, 53% are abandoned if a mobile site takes longer than 3 seconds to load. The source counts visits, not visitors. The data comes from a sample of 3,700 mobile sites that shared their benchmark data from Google Analytics, and the report focuses on publishers. It is more than ten years old.

Featured in: SEO and GEO, AI marketing, Internet marketing FAQs

State of Online Retail Performance, Spring 2017

Akamai, Spring 2017

A 100-millisecond delay in load time cost up to 7% in conversion. With a 1-second delay, conversion dropped by 21.8% on desktop, 20.5% on mobile, and 17.8% on tablet. The data comes from webshops that are Akamai clients (approximately 10 billion visits in one month). Akamai is a vendor and has a vested interest in the topic. The PDF file linked from this page is hosted by a third party, not Akamai.

Featured in: SEO and GEO, AI marketing, Internet marketing FAQs

Page Speed Report

Unbounce, 2019 (fieldwork October and November 2018)

In a survey of 750 consumers (via Google Surveys), 22% said they would close the tab and 14% would visit a competitor if load time was 3 seconds. Nearly 70% said speed affects their willingness to buy. These are self-reported answers, not measurements. Unbounce also surveyed 395 marketers.

Featured in: SEO and GEO, AI marketing, Internet marketing FAQs

Site Speed is (Still) Impacting Your Conversion Rate

Portent, 20 April 2022

Research on over 100 million page views across 20 B2B and B2C sites and more than 27,000 landing pages, over 30 days. For B2B sites, conversion at 1 second load time was 3 times higher than at 5 seconds and 5 times higher than at 10 seconds. For webshops, it was 2.5 times higher (1 vs 5 seconds). On average, conversion decreased by approximately 0.3 percentage pointsPercentage Point The difference between two percentages, calculated by simple subtraction. per extra second. Portent is a marketing agency and the sites belong to its own clients.

Featured in: SEO and GEO, AI marketing, Internet marketing FAQs

Largest Contentful Paint (LCP)

web.dev (Google), continuously updated (read 5 October 2026)

Google considers an LCP of 2.5 seconds or less to be good, measured at the 75th percentile of page visits, separately for mobile and desktop.

Mentioned on: SEO and GEO, AI marketing, Internet marketing FAQs, Calculate load time revenue, LCP, Increase conversion

The 2025 Web Almanac: Performance

HTTP Archive, 15 January 2026

On mobile, 62% of pages have a good LCP; on desktop, 74%. Measured with real user data from the Chrome UX Report.

Featured in: Calculate revenue based on load time

Web Vitals

web.dev (Google), updated 31 October 2024 (accessed 17 October 2026)

LCP is one of the Core Web Vitals.

Featured in: LCP

Why does speed matter?

web.dev (Google), updated 1 November 2023 (accessed 17 October 2026)

Visitors drop off if a page is slow. The source describes this in words, without a specific figure that applies to every site.

Mentioned on: LCP, Increase conversion

To the topics

Mobile

Desktop vs Mobile Market Share Netherlands

StatCounter Global Stats, September 2026

In September 2026, more than half of page views in the Netherlands came from a phone: 66% according to the page, while another period showed 56%. It counts page views across more than a million websites, not visitors, and the figure varies by period.

Featured in: WordPress website development, Internet marketing FAQs

Mobile-first indexing best practices

Google Search Central, continuously updated

Google uses the mobile version of a site for indexing and rankingPosition Your position is the place your website holds in the search results for a specific keyword. Position 1 is the top organic result.. This is Google’s own documentation and contains no figures on the effect on your position.

Featured in: WordPress website development

WebKit Features for Safari 27.1

WebKit blog (Apple), 23 October 2026 (accessed 7 November 2026)

Apple writes that the page width in Safari 27.1 on the iPhone Duo can range from 382 to 951 CSS pixels, and that websites and web apps on the iPhone Duo work just like everywhere else.

Mentioned in: Foldable phone: what changes for your website?

Viewport, Resolution, Diagonal Screen Size, and DPI for the Most Popular Smartphones

DeviceAtlas, updated 5 November 2026 (accessed 7 November 2026)

Traffic measurement of the screen width as used by the browser, with approximately 412 × 960 pixels folded and 750 × 832 pixels unfolded for the Galaxy Z Fold7.

Mentioned in: Foldable phone: what changes for your website?

Viewport Segments explainer (WICG visual-viewport)

WICG, explanatory document on GitHub (accessed 7 November 2026)

Explanatory document about Viewport Segments, with an example that falls back to standard responsive layout via the resize event when the device is folded.

Mentioned in: Foldable phone: what changes for your website?

Global Foldable Shipments to Hit 100 Million Cumulative Milestone by End of 2026

Counterpoint Research, 7 September 2026 (accessed 7 November 2026)

Counterpoint cites foldable phones as approximately 2% of the total smartphone market and predicts a 25% share of the foldable market for Apple in 2026, written before the presentation of the iPhone Duo.

Mentioned in: Foldable phone: what changes for your website?

IDC press release on foldable phones (prUS54017625)

IDC, press release dated 9 December 2025 (accessed 7 November 2026)

IDC estimates global sales of foldable phones in 2025 at 20.6 million units, 10% more than a year earlier.

Mentioned in: Foldable phone: what changes for your website?

iPhone Duo – Apple

Apple.com (accessed 7 November 2026)

Apple product page showing the US price of the iPhone Duo starting from $1,999 for 256 GB.

Mentioned in: Foldable phone: what changes for your website?

Device Posture API

W3C, Candidate Recommendation Draft dated 26 October 2026 (accessed 7 November 2026)

Specification of the Device Posture API, which states regarding accessibility that content should not be designed for a single device orientation.

Mentioned in: Foldable phone: what changes for your website?

Analytics dimensions and metrics – Analytics Help

Google Analytics Help (accessed 7 November 2026)

Overview of dimensions in Google Analytics 4, including Device model and Screen resolution; a separate dimension for foldable phones is not included.

Mentioned in: Foldable phone: what changes for your website?

To the topics

Search Engines and AI Answers

What 107,000 pages reveal about Core Web Vitals and AI search

Search Engine Land (Dan Taylor, SALT.agency), 13 January 2026

In an analysis of 107,352 pages visible in Google AI Overviews, there is no strong positive correlation between Core Web Vitals and AI visibility. Very slow pages perform worse, but faster than the baseline doesn’t yield extra benefits. It shows correlation, not causation.

Featured in: WordPress website development

State of SEO 2026

Search Engine Journal, 10 September 2025

In a survey of 371 SEO professionals in 52 countries, 60% reported more organic traffic and 34% more leads or conversions. These are self-reported experiences from professionals, not a measurement of SEO’s effect.

Featured in: WordPress website development, Internet marketing FAQs

2024 Organic Search Traffic Benchmarks

Conductor, 16 July 2024 (updated)

Organic search accounted for an average of 33% of website traffic, measured across more than 800 large corporate websites in 26 sub-sectors, from June 2023 to May 2024. Conductor itself provides SEO software. The figure applies to large global sites, not small Dutch sites.

Featured in: Internet marketing plan, Internet marketing FAQs

Understanding page experience in Google Search results

Google Search Central, continuously updated (read 5 October 2026)

Google: “Core Web Vitals are used by our ranking systems.” At the same time: “Google Search always seeks to show the most relevant content, even if the page experience is sub-par.” Speed counts, but relevance takes precedence.

Featured in: SEO and GEO, AI marketing, Internet marketing FAQs

Why Language Models Hallucinate

Kalai et al., arXiv:2509.04664, 4 September 2025

A language model can provide plausible-sounding but incorrect answers.

Featured in: Agentic AI

AI in society: experiences and opinions, chapter 2: Experience with AI

CBS, 25 February 2026 (2025 measurement)

Over half have used chatbots at some point (57 per cent). This refers to a measurement taken over 2025.

Mentioned on: AI assistant, GEO, More visitors to your website

AI Features and Your Website

Google Search Central, updated 10 December 2025 (accessed 17 October 2026)

AI OverviewsAI Overviews AI Overviews are the summaries that Google provides using AI for search queries, appearing above the regular links. show links to supporting pages. Google only displays them if, according to Google, they add value to the regular search results.

Mentioned on: AI Overviews, More visitors to your website

Understand JavaScript SEO Basics

Google Search Central, updated 4 March 2026 (accessed 17 October 2026)

Search engines also read a page’s code. Text that only appears via JavaScript is sometimes seen later, or not at all by some bots.

Featured in: Source code

Get started with Google Business Profile

Google Business Profile Help, continuously updated (accessed 17 October 2026)

A business profile lets you manage how your business appears in Google Maps and Google Search free of charge.

Featured in: Google Business Profile

Tips to improve your local ranking on Google

Google Business Profile Help, continuously updated (accessed 17 October 2026)

Businesses with complete and accurate data are more likely to appear in local search results. More reviews and positive ratings can help local rankings.

Mentioned in: Local SEO, More visitors to your website, Multilingual website: one or more domains?

Spam Policies for Google Web Search

Google Search Central, updated 28 August 2026 (accessed 17 October 2026)

Google defines buying or excessive exchanging of links to influence rankings as link spam, and excessive repetition of keywordsKeyword A keyword is the word or phrase that someone types into a search engine to find something, such as 'plumber Zaandam' or 'best savings account'. (keyword stuffing) as spam. Multiple pages targeting regions or cities that lead visitors to a single page are called doorway abuse.

Mentioned on: Link profile, Local SEO, Keyword density, More visitors to your website

Site Moves and Migrations

Google Search Central, updated 20 August 2026 (accessed 17 October 2026)

During a migrationMigration A migration is the process of moving a website to a new system, a new server, or a new address., visibility may fluctuate temporarily, and errors can prevent the new site from being fully indexed. Google recommends permanent 301 redirects from old to new addresses.

Mentioned in: Migration, Multilingual website: one or more domains?

Google users are less likely to click on links when an AI summary appears in the results

Pew Research Center, 22 July 2025

For searches with an AI summary, users click less on regular results. The measurement was conducted in the United States.

Featured in: Position

How Google determines the ranking of results

Google Search, continuously updated (accessed 17 October 2026)

Google uses information such as location, past search history, and search settings to tailor results.

Featured in: Semantic search

Creating Helpful, Reliable, People-First Content

Google Search Central, updated 5 October 2026 (accessed 17 October 2026)

Google asks whether someone has learned enough after reading to achieve their goal.

Mentioned in: Search intent, Multilingual website: one or more domains?

To the topics

Domains, languages and migrations

Managing Multi-Regional and Multilingual Sites

Google Search Central, updated 10 December 2025 (accessed 6 November 2026)

Google compares country domains, subdomains, subdirectories and address parameters side by side, without preference. For subdirectories, advantages include being easy to set up and low maintenance, while disadvantages include a single server location and sites being harder to separate. Extensions that are not country codes are considered generic.

Mentioned in: Multilingual website: one or more domains?

Change of Address tool

Search Console Help, undated (accessed 6 November 2026)

The tool works per domain or subdomain, also separately for www. Do not move multiple sites to one location simultaneously, but one by one. Signals are retained for 180 days and you must keep the old domain registered for at least a year.

Mentioned in: Multilingual website: one or more domains?

Add a website or platform property to Search Console

Search Console Help, undated (accessed 6 November 2026)

A domain property applies to one root domain with its subdomains. For multiple domains, you add a property for each domain.

Mentioned in: Multilingual website: one or more domains?

Set up cross-domain measurement

Google Analytics Help, undated (accessed 6 November 2026)

Without this setting, Analytics counts one visitor moving between two domains as two users and two sessions. The setting requires the same measurement code from the same data stream on all domains.

Mentioned in: Multilingual website: one or more domains?

Add your business website

Google Business Profile Help, undated (accessed 6 November 2026)

The profile has one website field. Some categories can add extra links for ordering, booking or appointments. There is no mention of language versions.

Mentioned in: Multilingual website: one or more domains?

The EMD Update: Google Issues Weather Report Of Crack Down On Low Quality Exact Match Domains

Search Engine Land, Barry Schwartz, 28 September 2012 (accessed 6 November 2026)

Matt Cutts from Google announced a change that causes low-quality keyword domains to rank lower. This affected 0.6 per cent of English-language searches in the US.

Mentioned in: Multilingual website: one or more domains?

John Mueller On Keyword Domain Names And Branding For SEO

Search Engine Journal, Kristi Hines, 31 December 2023 (accessed 6 November 2026)

Mueller said that a keyword in your domain name provides no recognisable long-term advantage. Report of a post on Reddit.

Mentioned in: Multilingual website: one or more domains?

Google Says Splitting & Merging Sites Takes Longer Than Normal Site Migrations

Search Engine Roundtable, Barry Schwartz, 25 April 2024 (accessed 6 November 2026)

Mueller wrote that splitting and merging sites takes much longer than a standard migration. He does not specify a number of weeks.

Mentioned in: Multilingual website: one or more domains?

Google Says Adding Country Codes To URLs Won’t Help Much

Search Engine Roundtable, Barry Schwartz, 3 January 2025 (accessed 6 November 2026)

Mueller said that a country code in the path provides no noticeable change in regional rankings and that a country extension separates country sites more clearly. This refers to countries, not languages.

Mentioned in: Multilingual website: one or more domains?

Google Questions Need For ccTLDs For Websites To Go International

Search Engine Roundtable, Barry Schwartz, 29 December 2025 (accessed 6 November 2026)

John Mueller from Google wrote on Reddit in December 2025 that he wonders if you really need to distribute your site across country domains. You make it harder for yourself, he wrote, and also for search engines to understand each of those sites.

Mentioned in: Multilingual website: one or more domains?

Own measurement: where do old addresses lead now?

EQwise, measured 7 November 2026

Random sample from archived addresses preserved from the final years of each domain with status 200, excluding images and login pages. Jumper: 40 addresses, 23 to a page, 17 to an error page. Coop: 40 addresses, 8 to a page, 32 to the PLUS homepage. Gogo: 40 addresses, 15 to a page, 2 to the homepage, 23 to an error page. Second sample, 40 old German addresses per brand: Helloprint 32 to a page, 8 to an error page; WE Fashion 24 to a page, 16 to an error page; Batavus all 40 to an error page; Bax Music all 40 to the Dutch homepage. Error page means status code 404. The moments in the article are from web archive recordings and are approximations: batavus.de still had its own pages on 15 September 2025, batavus.be on 13 January 2026, bax-shop.de on 30 May 2024, wefashion.de in December 2022 and helloprint.de on 15 August 2022. hema.be switched between February and April 2023, rituals.be around February 2019. No individual pages were found for action.be and auping.de since 2020. Viewed on the same day: where the old domains of the examples lead, and the folders of Action and Rituals. For Vastgoed Nederland, vastgoedpro.nl was tested; vbo.nl could not be tested.

Mentioned in: Multilingual website: one or more domains?

What migrating from subdomains to subfolders means for SEO results

Marketingfacts, Marc Thierrij, 16 August 2018 (accessed 7 November 2026)

Holland Bikeshop moved its languages from subdomains to folders on the main domain. Year-on-year measurement for June 2018: 27.5 per cent more organic traffic and 5 per cent more orders via organic traffic. The author is an SEO consultant at Maxlead.

Mentioned in: Multilingual website: one or more domains?

Google Says Why Site Migrations Can Exceed 4 Months

Search Engine Journal, Roger Montti, 13 October 2026 (accessed 6 November 2026)

Mueller said that an address change can take longer than four months. In the example, Google chose the old domain again after four months.

Mentioned in: Multilingual website: one or more domains?

How Long Should An SEO Migration Take? [Study Updated]

Search Engine Journal, Dan Taylor, 9 January 2025, updated 13 January 2025 (accessed 11 November 2026)

Study of 892 domain migrations: an average of 523 days until the new domain matched the estimated organic traffic of the old one. Traffic was estimated using third-party tools, not measured in the analytics of the sites themselves. The data was collected on 22 October 2024.

Mentioned in: Multilingual website: one or more domains?

Here’s an update on our use of country code top-level domains

Google, The Keyword, 15 April 2025 (accessed 9 November 2026)

Google has been redirecting its own country domains to google.com since April 2025. The company wrote that country domains are no longer necessary.

Mentioned in: Multilingual website: one or more domains?

From ccTLDs to one gTLD: case study and guide

searchVIU, Eoghan Henn, 7 May 2018, updated 2 November 2026 (accessed 9 November 2026)

An agency describing such migrations writes that they occur rarely and that companies keep their figures to themselves.

Mentioned in: Multilingual website: one or more domains?

Search Quality Evaluator Guidelines

Google, version 11 September 2025 (accessed 9 November 2026)

The guidelines for Google’s evaluators state that reputation is based on the experience of real users and the opinion of experts, and that what a site claims about itself does not count. Google evaluators research the company itself for a business that publishes on many websites. Those guidelines are instructions for humans, not a description of the algorithm.

Mentioned in: Multilingual website: one or more domains?

Categories – FAQ

Trustpilot Help Center, undated (accessed 9 November 2026)

Review platforms like Trustpilot give each domain its own ranking and price each domain separately.

Mentioned in: Multilingual website: one or more domains?

Trustpilot Pricing & Plans

Trustpilot Business, undated (accessed 9 November 2026)

A review subscription at Trustpilot on 9 November 2026 for the Plus package starts from $319 per month per domain, billed annually. That is from $3,828 a year per domain. Whether VAT is added is not stated on the page.

Mentioned in: Multilingual website: one or more domains?

Merge request form

Trustpilot Help Center, undated (accessed 9 November 2026)

Merging profiles is permitted under certain conditions, including for the same content in other languages, but extra costs may apply.

Mentioned in: Multilingual website: one or more domains?

Google store ratings: An overview for Merchants

Google Merchant Center Help, undated (accessed 9 November 2026)

Google seller ratings (the stars under your adverts) belong to a domain and a country. The domain in the advert must match the ratings, you generally need at least 100 unique reviews, and with one domain per country, you repeat the enquiry per domain.

Mentioned in: Multilingual website: one or more domains?

International study, May 2025, 75,000 brands

Commercial search data provider, blog post, 26 May 2025 (accessed 9 November 2026)

It looked at Google’s AI Overviews. Mentions of a brand elsewhere on the web were most strongly correlated with visibility (0.66 on a scale to 1: a clear correlation). Referring domains (sites linking to you) and backlinks (the links themselves) were more weakly correlated (0.30 and 0.22).

Mentioned in: Multilingual website: one or more domains?

International study, November 2025, 129,000 domains

Commercial search data provider, blog post, 24 November 2025 (accessed 9 November 2026)

It looked at ChatGPT. Sites with more referring domains received more citations (times ChatGPT mentions a site as a source). Sites on multiple review platforms received 4.6 to 6.3 citations, compared to 1.8 without.

Mentioned in: Multilingual website: one or more domains?

Pricing & Plans

CookieYes, undated (accessed 9 November 2026)

On 9 November 2026, the CookieYes basic package cost €9 per month or €90 a year per domain, excluding local tax. You get one free plan per account. Each additional domain has its own plan: an extra €90 a year for the basic package.

Mentioned in: Multilingual website: one or more domains?

International study, April 2024, 20,000 keywords in 15 markets

Data from a commercial search data provider, published 29 April 2024 (accessed 9 November 2026)

In an international study from April 2024, covering 20,000 keywords from webshops in 15 markets, country domains held 56 per cent of the top three positions.

Mentioned in: Multilingual website: one or more domains?

To the topics

Ads and Clicks

2024 Zero-Click Search Study: For every 1,000 EU Google Searches, only 374 clicks go to the Open Web. In the US, it’s 360.

SparkToro (data from Datos), 2 July 2024

Approximately 1% of all clicks on Google searches go to paid ads. Per 1,000 searches, 360 clicks (United States) and 374 clicks (EU) go to the open web, and 58.5% (United States) and 59.7% (EU) end without a click. This involves a click panel (September 2022 to May 2024) and measures the share of all clicks, not the share of users. Ad blockers can depress the ad figure, the author states.

Featured in: SEO and GEO, Internet marketing FAQs

In 2026, Less than One Third of Google Searches Still Send a Click

SparkToro (data from Similarweb), 9 June 2026

In the first four months of 2026, 68.01% of Google searches in the United States ended without a click, compared to 60.45% in 2024. This is a different panel than in the 2024 study (which showed 58.5% without a click for the United States) and assumes two-thirds mobile and one-third desktop, so the two studies are not directly comparable.

Featured in: SEO and GEO, Internet marketing FAQs

2026 Search Advertising Benchmarks

WordStream by LocaliQ, 1 June 2026 (updated)

The average CTR for search ads is 6.64%, across Google AdsGoogle Ads Google Ads is Google's advertising system, formerly known as AdWords. This allows you to place advertisements in search results, on YouTube, and on other websites. and Microsoft Ads combined and over 20 industries. The source does not mention countries or a figure for display ads, and it is a vendor measuring its own client campaigns.

Featured in: SEO and GEO, Internet marketing FAQs, Internet marketing plan

We Analyzed 4 Million Google Search Results. Here’s What We Learned About Organic CTR

Backlinko, 16 April 2025 (originally 27 August 2019)

Analysis of approximately 4 million Google search results using Search Console data. The first organic result received an average of 27.6% of clicks, and moving up one spot increased CTR by an average of 2.8%. Other sources measure lower; the figure depends on the research methodology.

Featured in: SEO and GEO, Internet marketing FAQs

Alphabet Inc., Form 10-K for fiscal year 2025

Alphabet Inc. (SEC), 2026 (annual report for 2025)

Google’s advertising revenue (Google advertising) was $294,691 million in 2025, compared to $264,590 million in 2024 and $237,855 million in 2023. This includes all of Google’s advertising revenue, so YouTube and the Google Network as well, not just Google Ads.

Featured in: Internet marketing plan, Internet marketing FAQs

Conversions versus key events in Google Analytics

Google Analytics Help, continuously updated (accessed 4 November 2026)

In Google Analytics 4, the actions that matter for your business are called key events; in Google Ads, they are called conversions.

Mentioned on: More visitors to your website, Increase conversion

To the topics

Usability

Hamburger Menus and Hidden Navigation Hurt UX Metrics

Nielsen Norman Group, 26 June 2016

In a test with 179 participants on 6 websites (desktop and smartphone), discoverability decreased by more than 20% with hidden navigation, tasks took at least 39% more time on desktop and 15% more time on mobile, and perceived difficulty was 21% higher. The study is ten years old and no more recent NNG research on website navigation has been found.

Featured in: WordPress website development, Internet marketing FAQs

“Get Started” Stops Users

Nielsen Norman Group, 20 August 2017

User research (usability studies) shows that a generic button like “Get Started” makes visitors stop, and it’s better to state exactly what follows. This is qualitative research without revenue figures.

Featured in: WordPress website development, Internet marketing FAQs

Web Content Accessibility Guidelines (WCAG) 2.2

W3C, 12 December 2024

WCAGWCAG WCAG stands for Web Content Accessibility Guidelines, which are the international standards that define when a website is fully usable for everyone, including people with disabilities. has three levels: A (the lowest), AA, and AAA (the highest).

Featured in: WCAG

Decree of 9 November 2024 … entry into force of the Implementation Act on Accessibility Requirements for Products and Services

Staatsblad 2024, 341, 13 November 2024

The Implementation Act on Accessibility Requirements for Products and Services entered into force on 28 June 2025.

Featured in: WCAG

To the topics

Design and Trust

Attention web designers: You have 50 milliseconds to make a good first impression!

Lindgaard, Fernandes, Dudek and Brown, Behaviour & Information Technology 25(2), 115-126, 2006

Three experiments show that people form a judgement about the visual appeal of a webpage within approximately 50 milliseconds, and that this judgement aligns with what they find after longer viewing. This is about first impressions, not purchases.

Featured in: WordPress website development, Internet marketing FAQs

How do users evaluate the credibility of Web sites? A study with over 2,500 participants

Fogg et al., Stanford Persuasive Technology Lab (ACM DUX 2003), June 2003

Over 2,500 participants evaluated the credibility of websites. The site’s appearance (“design look”) was mentioned most often, in 46.1% of comments. The study is from 2003.

Featured in: WordPress website development, Internet marketing FAQs

The effects of website design on purchase intention in online shopping: the mediating role of trust and the moderating role of culture

Ganguly, Dash, Cyr and Head, International Journal of Electronic Business 8(4/5), 302, 2010

Website design increases purchase intention via trust in the webshop and reduces perceived risk. Purchase intention was measured in a questionnaire, not actual conversion.

Featured in: WordPress website development, Internet marketing FAQs

McDonald’s Unveils ‘i’m lovin’ it’ Worldwide Brand Campaign

McDonald’s Corporation, press release via PR Newswire, 2 September 2003 (accessed 21 October 2026)

McDonald’s launches the global ‘i’m lovin’ it’ campaign from Munich, ‘ich liebe es’ in German, for customers in more than a hundred countries.

Mentioned on: Lockup

McDonald’s lovin’ hip new slogan

Associated Press in Deseret News, 3 September 2003 (accessed 21 October 2026)

The new strapline ‘I’m lovin’ it’ replaces ‘We love to see you smile’ and is set to run in more than a hundred countries by the end of that month.

Mentioned on: Lockup

Lockups

The University of Iowa, Brand Manual (accessed 21 October 2026)

The brand guidelines dictate that lockups are not used on university websites: the logo is already in the header bar and the name of a department is displayed as plain text.

Mentioned on: Lockup

To the topics

Marketing Budgets

Gartner CMO Spend Survey 2025, as reported by Marketing Brew

Gartner (via Marketing Brew), May 2025

The marketing budget in 2025 was an average of 7.7% of revenue, and 31% of the budget went to paid media. The survey included approximately 400 CMOs in North America, the UK, and Europe, almost all from companies with over $1 billion in revenue (fieldwork February and March 2025). So these are large companies and the figure is not a standard for a small business. This page follows the Marketing Brew article because Gartner’s own press release was inaccessible.

Featured in: Internet marketing plan, Internet marketing FAQs

How to Get the Most From Your Marketing Budget

U.S. Small Business Administration (blog by guest author Rieva Lesonsky), 9 July 2019

The blog states that companies in 2018 spent an average of 7.9% of their revenue on marketing: 6.3% for B2B products, 6.9% for B2B services, 9.6% for B2C products, and 11.8% for B2C services. The figure is secondhand, as the blog cites research from Web Strategies. It refers to all companies, not small businesses, and says nothing about the share of digital.

Featured in: Internet marketing plan, Internet marketing FAQs

The CMO Survey, March 2023 edition

Duke University Fuqua School of Business, Deloitte and American Marketing Association, March 2023

According to 314 US marketing leaders, an average of 53.8% of the marketing budget went to digital channels and 17% to social media. These are self-reported figures from US companies at a single point in time.

Featured in: Internet marketing plan, Internet marketing FAQs

To the topics

AI agents

How we built our multi-agent research system

Anthropic (engineering blog), 13 June 2025

A team consisting of a lead agent (Claude Opus 4) with helper agents (Claude Sonnet 4) performed 90.2% better than a single agent alone on Anthropic’s internal research test. By having the helper agents search in parallel, the turnaround time for complex questions decreased by up to 90%. Conversely, such a system uses approximately 15 times as many tokens as a standard chat conversation. This refers to the developer’s own test, using models from 2025.

Mentioned on: AI swarm for discoverability

Towards a science of scaling agent systems: When and why agent systems work

Google Research (with MIT), 28 January 2026

A controlled comparison of 180 AI agent setups. On tasks that can be broken down, such as financial analysis, a team with a central director improved performance by 80.9% compared to a single agent. On tasks that build on each other step-by-step, performance was 39 to 70% worse. Agents without direction increased errors by up to 17.2 times, while with a director, this remained at 4.4 times. If a single agent already solves a task correctly more than about 45% of the time, a team rarely yielded any benefit. These are test environments, not measurements on websites.

Mentioned on: AI swarm for discoverability

Improving Factuality and Reasoning in Language Models through Multiagent Debate

Du et al., MIT and Google Brain (arXiv), May 2023

Agents discussing and checking each other’s answers made fewer factual errors than a single model. In writing biographies, accuracy rose from 66% to 74%, and on the MMLU knowledge test from 64% to 71%. The research used language models from 2023; later studies show that the gains do not apply to every task.

Mentioned on: AI swarm for discoverability

To the topics