# ConvRadar

> ConvRadar is a hosted MCP server that connects Google Analytics 4 to Claude and ChatGPT. It runs conversion audits on GA4 data and publishes conversion benchmarks with their sources.

Benchmark pages state the base of every figure (per session or per product page view), the data period and the source. Numbers on the industry trend pages update monthly from the source file.

## Benchmarks
- [Fashion add-to-cart rate benchmark 2020-2026: 6.3%](https://convradar.com/benchmarks/fashion/add-to-cart-rate): Fashion add-to-cart rate per product page view: 6.3% from Sep 2025 to Aug 2026, down from 6.7% in 2025 and 8.0% in 2023 (Dynamic Yield data).
- [Food and beverage add-to-cart rate benchmark 2020-2026: 9.3%](https://convradar.com/benchmarks/food-beverage/add-to-cart-rate): Food and beverage add-to-cart rate per product page view: 9.3% from Sep 2025 to Aug 2026, down from 9.7% in 2025 and 11.9% in 2023 (Dynamic Yield data).
- [Multi-brand retail add-to-cart rate 2020-2026: 7.1%](https://convradar.com/benchmarks/multi-brand-retail/add-to-cart-rate): Multi-brand retail add-to-cart rate per product page view: 7.1% from Sep 2025 to Aug 2026, down from 8.3% in 2025 and 8.9% in 2023 (Dynamic Yield data).
- [Consumer goods add-to-cart rate benchmark 2020-2026: 4.7%](https://convradar.com/benchmarks/consumer-goods/add-to-cart-rate): Consumer goods add-to-cart rate per product page view: 4.7% from Sep 2025 to Aug 2026, down from 4.9% in 2025 and 6.5% in 2023 (Dynamic Yield data).
- [Mobile vs desktop add-to-cart rate 2020-2026: 6.4% vs 5.1%](https://convradar.com/benchmarks/mobile-vs-desktop/add-to-cart-rate): Mobile add-to-cart rate per product page view: 6.4% from Sep 2025 to Aug 2026, desktop 5.1%; mobile has led every month since Jan 2025 (Dynamic Yield data).
- [Conversion rate benchmarks by industry](https://convradar.com/benchmarks)
- [E-commerce conversion rate benchmarks](https://convradar.com/benchmarks/ecommerce): Ecommerce conversion rate benchmarks: median ~2.2% (typical 1.2-3.4%), add-to-cart ~5%, checkout completion ~30%, AOV ~$68. Compare your store's real numbers with GA4.
- [DTC apparel conversion rate benchmarks](https://convradar.com/benchmarks/dtc-apparel): DTC apparel & fashion conversion rate benchmarks: median ~2.6% (typical 1.4-4.1%), add-to-cart ~5%, checkout ~30%, AOV ~$92. Compare your store's real numbers with GA4.
- [SaaS conversion rate benchmarks](https://convradar.com/benchmarks/saas): SaaS conversion rate benchmarks: registration ~10%, activation ~37%, trial-to-paid ~25%, retention ~90%. See the ranges, then get your real funnel from GA4.

## Articles
- [Singapore Bot Traffic in GA4: What It Is, How to Spot It, and Why It Now Reaches Checkout](https://convradar.com/resources/ga4-singapore-bot-traffic): Thousands of sessions from Singapore in GA4 are automated traffic, not visitors. What bot traffic is, which bots GA4 can and cannot see, the signs that give it away, what it does to your conversion rate, and the newer wave that fires add_to_cart and begin_checkout.
- [Unassigned Traffic in GA4: Two Problems, and Only One of Them Can Be Fixed Backwards](https://convradar.com/resources/ga4-unassigned-traffic): Unassigned is where GA4 files a session when no channel rule matches. A source nobody wrote a rule for can be reclassified for every day already collected. A session with no source at all cannot. How to split the row, the exact rule each value missed, and the primary channel group setting that only fixes the future.
- [GA4's Checkout Journey Report Reads 100% Abandonment on Shopify, and Google Says Why on the Same Page](https://convradar.com/resources/ga4-checkout-journey-report-empty): The Checkout journey report needs begin_checkout, add_shipping_info, add_payment_info and purchase. Google's own help page states that Shopify sends three of those four. The funnel is closed with no toggle, so one absent event zeroes the three steps behind it while Monetization reports the orders normally.
- [GA4 Behavioral Modeling: Sessions Get Estimated, session_start Doesn't, and Most Funnels Divide One by the Other](https://convradar.com/resources/ga4-behavioral-modeling): GA4's consent-mode modeling estimates users, sessions and new users, and leaves event counts untouched. So session_start in a report is unmodeled and inflated, the same event is modeled down inside a funnel exploration, BigQuery gets neither, and the conversion rate you quote depends on which of the two numbers you grabbed.
- [AI Bot Traffic in GA4: The Crawlers Aren't There, and the Agent Says It's Chrome](https://convradar.com/resources/ai-bot-traffic-ga4): GPTBot and ClaudeBot never run your GA4 tag, so filtering AI bots out of GA4 removes nothing that was in it. What does land in GA4 is an agentic browser reporting itself as Chrome 141, and it moved one store's conversion rate from 1.300% to 1.112% with nothing on the site changed.
- [Is GA4 Data Accurate? Four Edits It Makes After Collection, Two of Them Silent](https://convradar.com/resources/is-ga4-data-accurate): GA4 estimates sessions to ±3.25% and users to ±1.63%, folds high-cardinality rows into (other), hides rows under Google signals, and samples past 10 million events. Two of the four get flagged in the interface. The Data API flags sampling, thresholding and the roll-up on every response.
- [Checkout Conversion Rate Benchmarks (2026): Mobile Doesn't Lose It Inside the Checkout](https://convradar.com/resources/checkout-conversion-rate-benchmark): Checkout conversion rate benchmarks for 2026: 45% average, 59% top quintile, 66% top decile. Mobile completes checkout at 44% against desktop's 49%, so three quarters of the mobile conversion gap happens before the checkout page ever loads.
- [GA4 Engagement Rate Benchmarks: Everyone's 56% Was Measured on a Different Stopwatch](https://convradar.com/resources/ga4-engagement-rate-benchmark): GA4 engagement rate benchmarks by channel and industry, and the reason they don't compare: the engaged-session timer is a per-property setting between 10 and 60 seconds, the channel spread runs 42.62% to 72.50%, and most published bounce-rate figures are Universal Analytics numbers.
- [Add-to-Cart Rate Benchmarks (2026): Your Number Is 4.6% and 9.8%, and Both Are Right](https://convradar.com/resources/add-to-cart-rate-benchmark): Add-to-cart rate benchmarks for 2026 by sub-vertical, in both denominators: per session and per product view. Why the published figures disagree by 2x, and the reach rate that converts one into the other.
- [Supermetrics Alternatives in 2026: An Honest Side-by-Side](https://convradar.com/resources/supermetrics-alternative): A useful comparison of Supermetrics alternatives in 2026, Coupler.io, Windsor.ai, Porter Metrics, Improvado, ConvRadar. The real invoice arithmetic, and no fake winners.
- [Why Your ChatGPT MCP Connector Goes Quiet (and When It's Actually Broken)](https://convradar.com/resources/chatgpt-mcp-slow): ChatGPT MCP feels slow because a single tool call dies at 60 seconds. Ten timed runs against a live analytics connector, what the pauses mean, and the three ways it really fails.
- [Conversion Rate by Traffic Source (2026): The Channel Bands, and Why 'Paid Beats Organic' Is Usually Branded Traffic](https://convradar.com/resources/conversion-rate-by-traffic-source): 2026 conversion rates by traffic channel split into two bands, search, email, direct and referral near 5%, paid and organic social near 2%. Plus the two ways the numbers lie: mixing sources, and GA4 hiding branded paid search.
- [New vs Returning Visitor Conversion Rate (2026): It's 1.7×, Not the 3-5× Everyone Repeats](https://convradar.com/resources/new-vs-returning-conversion-rate): New visitors convert about 1.7× lower than returning ones, not the 3-5× the internet keeps repeating. The 2026 benchmark by cohort, why a low blended rate is usually a traffic-mix problem, and how to split it in GA4.
- [Review Mining: Turn Customer Reviews into Landing Page Copy](https://convradar.com/resources/review-mining): Review mining, done right: mine category discussion instead of your own reviews, rank pains by frequency, and gap-check your landing page against them.
- [Lead-Gen Conversion Rate Benchmarks by Industry (2026): Session-to-Lead, Not the Landing-Page Number](https://convradar.com/resources/lead-gen-conversion-rate-benchmarks): Whole-site session-to-lead conversion benchmarks by industry for 2026, the real 1-4% bands, why the famous 6.6% landing-page figure isn't your number, and the phone-call half everyone forgets to count.
- [Ecommerce Conversion Rate Benchmarks (2026): The Bands by Sub-Vertical, Not the 2% Average](https://convradar.com/resources/ecommerce-conversion-rate-benchmarks): US ecommerce conversion rate benchmarks for 2026 as p25/median/p75 bands by sub-vertical, electronics, apparel, beauty, marketplace, subscription, plus how to find where your own number actually lands.
- [GA4 Dashboards vs Asking AI: When You Actually Need Looker Studio](https://convradar.com/resources/ga4-dashboard-vs-ai): GA4 dashboard vs AI: a dashboard earns its keep for two jobs, monitoring a few numbers and reporting to people who won't log in. For everything else, ask AI.
- [Which AI Traffic Converts Best? ChatGPT vs Perplexity vs Gemini vs Claude](https://convradar.com/resources/which-ai-traffic-converts-best): Which AI traffic converts best, ChatGPT, Perplexity, Gemini, or Claude? The public benchmarks disagree by 10x for the same platform. Here's how to read your own GA4 by source instead.
- [Freemium Conversion Rate Benchmarks: the 8% Median Nobody Hits](https://convradar.com/resources/freemium-conversion-rate-benchmarks): Freemium conversion rate benchmarks for 2026, free-to-paid by motion (freemium, free trial, reverse trial, demo), the two-crowd distribution hiding behind the 8% median, payers-per-1,000-visitors arithmetic, and which numbers GA4 can even see.
- [The CRO Audit Prompt That Isn't Allowed to Guess](https://convradar.com/resources/cro-audit-prompt): Viral CRO prompts produce confident audits with invented numbers. This is the six-phase conversion audit prompt that reads live GA4 data, verifies causes on the actual page, and shows its arithmetic.
- [GA4 Ask Advisor Not Showing? Why It's Missing and What to Do About It](https://convradar.com/resources/ga4-ask-advisor-not-showing): Ask Advisor missing from your GA4? It's an English-only beta on a staggered rollout. How to check each gate, the one thing you can actually fix, and how to get GA4 answers today.
- [How to Use Conversion Benchmarks (Without Fooling Yourself)](https://convradar.com/resources/how-to-use-conversion-benchmarks): A conversion benchmark can't grade you, it can only show where you leak. How to pick the right peer set, benchmark each funnel step instead of the finish line, and spot the ways teams misuse the numbers.
- [Why Your GA4 Attribution Is Wrong (and How to Find the Channel Getting Credit It Didn't Earn)](https://convradar.com/resources/ga4-attribution-wrong): GA4 attribution is wrong more often than you'd think: inflated Direct, self-referrals stealing credit, and the 2026 model changes. How to find which channel got robbed, and fix it.
- [Cart Abandonment in GA4: Find Where Carts Die (and Why)](https://convradar.com/resources/cart-abandonment-ga4): GA4 has no cart abandonment report. What it has: two built-in funnels, an honest rate formula, five ways the number lies, and a path from 'which step' to 'why'.
- [Coupler.io Alternatives in 2026: An Honest Side-by-Side](https://convradar.com/resources/coupler-io-alternative): A useful comparison of Coupler.io alternatives in 2026, Windsor.ai, Supermetrics, Improvado, Two Minute Reports, ConvRadar. Which fits which job, with no fake winners.
- [GA4 for SaaS: Track Signups, Trials, and Activation (and Where GA4 Stops)](https://convradar.com/resources/ga4-for-saas): GA4 for SaaS: which signup, trial, and activation metrics GA4 can actually measure, the cross-domain trap that quietly breaks your funnel, and how to pull it all through AI.
- [Free Conversion Audit: What a Real One Finds (and How to Get Yours)](https://convradar.com/resources/free-conversion-audit): Most 'free conversion audits' are a sales call in disguise. Here's what a real conversion audit of your own GA4 finds: the funnel drop-off, the lost revenue, the fix. Plus how to get the quick audit free in 60 seconds, and what the paid written report adds.
- [GA4 vs Shopify: Why the Numbers Don't Match (and How to Reconcile Them)](https://convradar.com/resources/ga4-vs-shopify-numbers-dont-match): GA4 says one revenue number, Shopify says another, and they're both right. Which direction the gap runs, what size is normal, and a like-for-like reconciliation that lands on the real tracking loss instead of chasing zero.
- [Is It Safe to Connect Google Analytics to an AI? What the Connector Sees](https://convradar.com/resources/connect-ga4-to-ai-safe): Is it safe to connect Google Analytics to ChatGPT or Claude? What a read-only MCP connector can and can't touch, where your GA4 data actually goes, and the risks worth checking.
- [SaaS Conversion Rate Benchmarks (and Which Ones You Can Actually Measure)](https://convradar.com/resources/saas-conversion-rate-benchmarks): SaaS conversion rate benchmarks by funnel stage and monetization model, signup, trial-to-paid, freemium, activation, NRR, with the line marking what GA4 can measure and what lives in Stripe.
- [Why Did My GA4 Traffic Drop Suddenly? Real Drop vs Tracking Glitch](https://convradar.com/resources/ga4-traffic-drop): GA4 traffic dropped overnight? Half the time GA4 just stopped counting. The visitors never left. How to tell a real drop from a tracking artifact: by shape, by segment, and by one Search Console cross-check.
- [Google Analytics MCP Without a Service Account: Three Paths That Work](https://convradar.com/resources/google-analytics-mcp-without-service-account): Stuck at gcloud auth on the official Google Analytics MCP? What the service account is actually for, the two ways through Google Cloud, and the one path around it.
- [How to Run a GA4 Audit with AI (and What It Actually Finds)](https://convradar.com/resources/ga4-audit-with-ai): Run a GA4 audit with AI in about four minutes: the exact prompt, a real audit output with findings, where AI beats a $3,000 consultant audit, and where it doesn't.
- [GA4 AI Assistant (Ask Advisor): What It Can and Can't Do](https://convradar.com/resources/ga4-ai-assistant): What GA4's AI assistant (Ask Advisor, Insights, predictive metrics) actually does in 2026, how to use it, and where it stops short of a real decision.
- [AI Traffic Not Converting? Check the Landing Page](https://convradar.com/resources/ai-traffic-not-converting-landing-pages): AI traffic not converting? It usually lands on deep pages built for SEO, not orientation. Find which AI landing pages leak in GA4, then fix it.
- [How to Connect GA4 to ChatGPT Without Code (2026 Setup)](https://convradar.com/resources/connect-ga4-to-chatgpt): Connect Google Analytics (GA4) to ChatGPT with no code. The live MCP setup, enable Developer Mode, add the connector, authorize read-only GA4, plus what to ask and how it differs from Claude.
- [How to Track AI Traffic in GA4 (and Whether It Converts)](https://convradar.com/resources/track-ai-traffic-ga4): Track AI traffic in GA4: find the AI Assistant channel, build a custom channel group for AI referral traffic, and see whether ChatGPT and Perplexity convert.
- [GA4 Ecommerce Reports with AI: Product Performance, Revenue, and Category Analysis](https://convradar.com/resources/ga4-ecommerce-reports-ai): How to pull GA4 ecommerce reports through AI, product performance, category analysis, AOV shifts, revenue attribution. Real prompts, real workflow, no GA4 UI required.
- [Google Analytics MCP vs ConvRadar: Honest Picks by Audience](https://convradar.com/resources/google-analytics-mcp-vs-convradar): Google's official Google Analytics MCP server is solid but requires Python, gcloud, and service accounts. Here's who it's actually for, and what the no-code path looks like.
- [Windsor.ai Alternatives in 2026: An Honest Side-by-Side](https://convradar.com/resources/windsor-ai-alternative): A useful comparison of Windsor.ai alternatives in 2026, Coupler.io, Supermetrics, Improvado, ConvRadar. Which fits which job, with no aggression and no fake winners.
- [Claude vs ChatGPT for Analytics: An Honest 2026 Comparison](https://convradar.com/resources/claude-vs-chatgpt-for-analytics): A practical Claude vs ChatGPT comparison for analytics work, where each one is stronger, what stays the same regardless of which you pick, and how to choose for GA4.
- [GA4 Mobile Conversion Rate: Why It's Half of Desktop and How to Fix It](https://convradar.com/resources/ga4-mobile-conversion-rate-fix): Mobile conversion rate typically runs at half of desktop. Diagnose the real cause in GA4 with AI, then apply the specific fixes that actually move the metric.
- [CRO with AI: A Practical Loop from Diagnose to Verify](https://convradar.com/resources/cro-with-ai): A practical loop for conversion rate optimization with AI, diagnose, hypothesize, test, verify. Where AI actually helps, where it doesn't, and the prompts that move work along.
- [Why Did My GA4 Conversion Rate Drop? A 10-Minute AI Diagnostic](https://convradar.com/resources/ga4-conversion-rate-drop): A step-by-step diagnostic for a GA4 conversion rate drop using Claude or ChatGPT. Define the drop, find where it concentrates, correlate with what changed, in ten minutes.
- [GA4 Prompts for Claude and ChatGPT: The Library I Actually Use](https://convradar.com/resources/how-to-write-ga4-prompts-for-claude): A library of GA4 prompts for Claude and ChatGPT, weekly audits, funnel diagnosis, anomalies, mobile, landing pages, channels, with the exact wording and why each one works.
- [How to Connect GA4 to Claude Without Code (5-Minute Setup)](https://convradar.com/resources/how-to-connect-ga4-to-claude-without-code): Connect GA4 to Claude with no code in five minutes. Step-by-step setup, what you can ask once it's live, and how to use the same connector with ChatGPT and other AI assistants.
- [GA4 MCP Server: A Practical Guide to Every Option in 2026](https://convradar.com/resources/ga4-mcp-server): Compare every GA4 MCP server option, Google's official, hosted SaaS like ConvRadar and Coupler.io, plus the BigQuery + LLM path. Pick the one that fits.

## Documentation
- [Ask your data](https://convradar.com/docs/ask-your-data): No fixed report. The connector describes what data exists, then aggregates whatever metrics and dimensions your question needs.
- [Funnel leak diagnosis](https://convradar.com/docs/funnel-leak-diagnosis): Walks the funnel step by step, finds the biggest drop, then splits that step by segment to find who actually dropped.
- [Anomaly detection](https://convradar.com/docs/anomaly-detection): Scans daily metrics for statistically significant spikes and drops, then checks whether the traffic mix behind them changed.
- [Segment comparison](https://convradar.com/docs/segment-comparison): Puts two segments side by side on the same metrics with computed deltas, plus the ready-made device and country splits.
- [Traffic mix](https://convradar.com/docs/traffic-mix): Source and medium breakdown with sessions, users and conversion rate per channel.
- [Product analysis](https://convradar.com/docs/product-analysis): Every SKU classified as underperforming, watchlist, healthy or no-purchase, with per-item view, cart and purchase rates.
- [Paid geographies](https://convradar.com/docs/paid-geographies): Region by source combinations scored for pause or scale, with confidence bands and automatic window expansion when the data is thin. Web report, no MCP tool.
- [Visual proof](https://convradar.com/docs/visual-proof): Captures real desktop and mobile screenshots through an anti-bot browser so the model looks at the page instead of guessing from HTML, then records what it observed. Screenshot capture counts as a page check: 30 a month on Pro, 1 to try on Basic.
- [Hypothesis library](https://convradar.com/docs/hypothesis-library): A curated catalogue of CRO hypotheses built from published research. Every finding is matched to one, and you can mark what you are testing so the next session resumes.
- [Change diary](https://convradar.com/docs/change-diary): Log what you shipped and when. Later, the same connector reads the pre and post windows and runs a z-test on the difference instead of eyeballing a chart.

## Tools
- [Conversion Audit](https://convradar.com/free-audit): A quick audit of your GA4 traffic in chat on the Free plan. The written audit with ranked fixes is part of Pro.
- [Page Scan](https://convradar.com/page-scan): Paste any URL and see where the page loses conversions. No GA4, no login.
- [Page speed check](https://convradar.com/page-speed): Measured speed and a read of your key pages for what costs you conversions. About 60 seconds, no GA4.
- [Voice of Customer](https://convradar.com/voice-of-customer): Part of Pro: what customers in your category complain about and whether your page answers it.
- [Benchmarks](https://convradar.com/benchmarks): Conversion benchmarks by industry, with the source and base of every figure.

## Optional
- [Fashion and apparel add-to-cart rate, monthly data (CSV)](https://convradar.com/benchmarks/fashion/add-to-cart-rate/data.csv)
- [Food and beverage add-to-cart rate, monthly data (CSV)](https://convradar.com/benchmarks/food-beverage/add-to-cart-rate/data.csv)
- [Multi-brand retail add-to-cart rate, monthly data (CSV)](https://convradar.com/benchmarks/multi-brand-retail/add-to-cart-rate/data.csv)
- [Consumer goods add-to-cart rate, monthly data (CSV)](https://convradar.com/benchmarks/consumer-goods/add-to-cart-rate/data.csv)
- [Mobile, desktop and tablet add-to-cart rate, monthly data (CSV)](https://convradar.com/benchmarks/mobile-vs-desktop/add-to-cart-rate/data.csv)
- [Sitemap](https://convradar.com/sitemap.xml)