What's new at Mida

• Faster reports and test delivery — reports with many secondary metrics now load up to 180× faster on complex queries, and we rebuilt visitor lookups, conversion queries, and the experiment-serving path, so your dashboard and your site's test delivery both respond quicker.

• Confidence could occasionally display inverted between control and variant — fixed, with a permanent safeguard added.
• Comparisons sometimes used the wrong baseline.
• Changing a test's primary goal now correctly recomputes past conversions.
• Metrics no longer show 0 when filtered by number of orders.
• Custom Script goals were missing from revenue reporting.
• Multi-page editing no longer overwrites your test's URL settings.
• Shopify purchases are now deduplicated by order ID for cleaner revenue attribution.

• A more intuitive segmentation analysis report — adjust the variables in your post-test segmentation and see how each one affects your results, with Mida suggesting your next step automatically.
• Simpler page targeting with URL patterns — wildcards (*) are handled automatically, no more separate advanced URL rules. An Advanced/Classic mode remains for complex setups.
• Switch between projects and organisations directly from the sidebar.

• Brand-new mida.so homepage and global navigation, rebuilt around MidaGX with a dedicated download entry.
• Refreshed pricing page — clearer Sandbox, Growth, and Enterprise tiers with the full feature comparison restored.

• New AOV confidence algorithm — a single unusually large order is far less likely to crown a false winner.
• Goals & metrics: one unified table, bounce rate promoted to primary metrics, placeholder dashes instead of misleading zeros, faster loading.
• Smarter test screenshots — captures the changed elements, side by side with control.
• Per-experiment revenue trend graphs.
• Faster geolocation at ~10× lower cost.
• A dedicated Metrics section with a Last-24h estimated event count next to each conversion goal — a pulse check on your tracking.
• Add team members and tags directly inside the experiment setup flow, before you publish.
• Integrations reframed around goals — every connection now speaks in outcomes, not plumbing.
• Redesigned checkout and billing pages.

• Conversions showing 0 in goal detail — the critical one, now resolved.
• Date-matching bug that could misalign report data.
• Active reports showed the wrong end date.
• Offline conversions failed to match their visitor.
• EU accounts could sync to the wrong plan.
• Whitelabel plans briefly showed the Mida watermark.

Plus 12 more fixes across charts, variant naming, page verification, and report scrolling.

• Reports load noticeably faster — smarter caching across all result views.
• Auto-redirect after the welcome screen.

• Completed tests kept collecting conversions after they ended.
• Metrics could show 0 while data was still loading, or when a report snapshot failed to parse.
• Bounce rate: accuracy and speed of calculation.
• Test duration calculation; secondary-goal live data not showing the full table.
• Revenue figures now blur while loading instead of flashing wrong numbers.
• Sign-in security: sessions fully invalidated on logout; cross-region Google sign-in errors now surface properly; tighter cross-site request protection.
• New tests correctly default to Bayesian analysis.

• Security hardening sweep: injection vulnerabilities patched, honeypot bot detection added, and email verification infrastructure upgraded.
• Platform stability: database connection fixes and higher capacity under load.
• Agencies: client projects are created automatically, with pooled usage billing.
• Data retention extended from 12 to 18 months.

• Website verification failed on sites using Cookiebot or server-side Google Tag Manager.
• Event-based goals weren't filtering correctly in results.
• Secondary goals can now match your Purchase goal.
• Edge cases in AI-generated experiments and Sunny AI's custom event creation.

Plus 8 more fixes across event validation, EU access, and report internals.

• MidaGX: viewport support, image generation, and better sharing.
• Password changes now require re-authentication; stricter email validation.

• First-touch attribution recorded incorrectly.
• Redeploying a test no longer overwrites its creation date.
• Renaming a goal now refreshes everywhere immediately.
• Plan downgrades reset usage counters correctly.

Plus 7 more fixes across integrations, error handling, and account security.

• Experiment table: custom tags, member assignment, and hover-to-see variant conversions.
• Completed tests now snapshot their final results — historic reports open instantly and never drift.
• Traffic estimation shows how many visitors your targeting rules will actually reach.

• Purchase events with different capitalization were counted as different events.
• Post-test segmentation errors and a wrong breakdown field.
• Hypothesis ICE score calculation.
• Anti-flicker script raised false alarms during verification.

Plus 6 more fixes across AI summaries, caching, and traffic estimation.

• Refreshed dashboard — clearer key metrics, faster decisions.
• Smoother test-creation flow.
• Export events with full detail; add anticipated events as goals directly.
• Multi-Armed Bandit allocation got smarter.

• Conversion-rate calculation bug — re-derived and corrected.
• Purchase events failing to export.
• Variant JS/CSS occasionally failed to load.
• Invited users were sometimes placed into a brand-new empty workspace.

Plus 8 more fixes across the Shopify launch, sharing links, and statistics.

• Sharper geo-targeting: state-level selection, case-insensitive matching, and a preview mode to test location rules.
• Bounce rate methodology upgraded; faster report queries.
• AI models upgraded across the platform.
• Smarter spam detection on signups.

• GIF uploads to variants failed.
• Event filters erroring in reports; empty conversion data.
• Geolocation missed some countries.
• Access revocation and invites to unregistered users.

Plus 7 more fixes across reports, trials, and AI generation.

• Visitor processing roughly twice as fast under load.
• Screenshot capture for EU-hosted accounts.

• Segment rules weren't checked against new visitors.
• Invalid dates in reports; revenue goals missing from goal checks.
• Goal status check bug; order-count filtering.

Plus 4 more fixes across surveys and result exports.

• More accurate visitor geolocation with better fallbacks.
• More reliable pageview-goal matching.

• WordPress export failed when a revenue goal was selected.

• Clearer page-targeting validation with helpful error messages.
• Per-visitor data export; simplified report labels and colored variant outlines.

• Bayesian is now the default statistical engine; Frequentist reporting shows a clear Yes/No with p-value.

• Default confidence level moved from 95% to 80% for faster conclusive results.

• Experiment list refresh: cleaner layout, campaign-type filter, sort by base conversions.

• Confidence intervals show lower and upper bounds.

• Revamped experiment reports with a Confidence Interval bar.

• Line-based conversion charts, a Visits comparison chart, and a daily activity heatmap.