Mida

What it costs, on both billing paths

Two clauses from the terms belong next to that table. No refunds are provided for partial billing periods, and cancellation takes effect at the end of the current billing cycle — so on the annual path, the moment to be sure is before the payment, and the 15-day trial is the instrument for that. The terms also state the tested-user allowance resets on the first of each calendar month, with mid-cycle upgrades prorated.

What is actually included

• A no-code visual editor plus an HTML/JS/CSS code editor, for standard A/B and split-URL tests
• MidaGX, the AI feature that generates test variations from a description — 30 credits a month free, unlimited on Growth
• Server-side testing via SDKs for iOS, Android and Flutter, and multi-armed-bandit optimisation, both Growth-only
• Targeting rules by segment, event, country, device, OS, browser, traffic source, returning visitor, language, time of day and cookies
• Integrations the vendor counts at 74+, including GA4, GTM, Shopify, WordPress, Webflow, HubSpot, Segment and Slack
• Campaign scheduling, one-click winning-variation deployment and content personalization, included in Growth rather than sold separately

The pricing detail the page does not publish

Mida markets its pricing as usage-based — 'pay for tested traffic' is the headline on the pricing page, above a selector running from 100K to 2M monthly tested users. What the page does not publish is a price for any tier above the first. When we selected higher tiers in a browser, the displayed Growth price and its stated 100,000-user allowance did not change. Mida's own listings on G2 and Capterra, both marked as provided by the vendor, describe the plan as 'starting at $399.00, usage based, per month'.

None of that is a contradiction, exactly — a starting price is consistent with everything we read. But it means a buyer with 500,000 tested users a month cannot price this product from public information, and any third-party page quoting a single figure for that buyer is guessing. If you are above the entry tier, the number that matters exists only in a sales conversation. Get it in writing, and get the overage behaviour — what happens the month you exceed your tier — in writing at the same time, because the public page does not state that either.

What customers say — three platforms, three very different bases

This is a case where quoting one rating would mislead in either direction, so here are all three as read on 13 August 2026. Capterra: 4.6 from 51 reviews — 37 five-star, 10 four-star, 3 three-star, 1 one-star. G2: 4.8 from 2 reviews, both posted in May 2026. Trustpilot: 3.2 from a single one-star review posted in August 2024. Mida's homepage shows the first two as badges and does not mention the third.

Context for weighing them. The Capterra base is the only one large enough to mean much, and many of its recent reviews carry Capterra's label for vendor-invited reviews with a nominal incentive offered — disclosed on each review, and worth knowing when reading a run of five-star entries. The recurring praise across platforms is consistent: responsive human support, easy setup, and the lightweight script. The recurring complaints are consistent too: a visual editor that reviewers describe as occasionally buggy or wonky, an interface a few call spartan, and reporting that pushes advanced analysis out to external BI tools. The lone Trustpilot reviewer, who arrived via an AppSumo lifetime deal for Howuku — the product Mida evolved from — reported A/B test data failing to collect and slow support responses in 2024.

Who this actually suits

If you run a small site and want to learn A/B testing properly, the Sandbox tier is a genuinely good place to do it — the traffic allowance is generous and the limits that bind are on test count, not visitors. If you are a marketing team or small agency whose traffic fits inside 100,000 tested users, Growth on annual billing is a straightforward, fully-unlocked tool at $3,588 a year, bought with a 15-day trial and the knowledge that refunds are not part of the deal. Who should look elsewhere: anyone whose traffic is well above the entry tier and who needs public, predictable pricing to budget against — that number is not published — and anyone who needs deep native reporting, which reviewers consistently say is the thin part of the product.