A Google-style site for a SaaS that turns table QRs into Google reviews

Puanla is a Turkish SaaS with one clear promise for local businesses: measure customer experience, grow your Google reviews. A QR code on the table captures feedback from every customer in seconds, makes it easy for happy customers to leave a Google review, and collects everything in one panel — Google-policy compliant, with no personal data collected. Logic Grid Studio built puanla.app from scratch on Next.js, in the deliberately Google-style interface the product itself promises — white, Roboto, the #1a73e8 blue — with the information architecture, the pricing page and the free-signup flow. On top of the build we delivered the technical SEO and the GEO layer, including a 13-question FAQPage schema shaped for AI answer engines.

The situation

Puanla is a Turkish SaaS for local businesses — cafés and restaurants, barbers and hairdressers, beauty salons, dental clinics, hotels and retail — that puts a QR code on the table and captures customer feedback in seconds. The platform makes it easy for happy customers to leave a Google review, while improvement ideas reach the business through a private feedback channel and everything lands in one panel. Its positioning is deliberate and strict: not a review-gating tool, Google-policy compliant, and KVKK-clean — the rating flow collects no personal data.

Why they came to us

A self-serve SaaS like this is sold by its website: the site has to explain the product in one pass, publish the price without asterisks, and carry a business owner from a search result to a free signup. Puanla needed that site built from scratch — and it needed to be findable both in classic search and in the AI answer engines where owners ask whether requesting reviews is even allowed, a question Puanla's own FAQ answers. Build, SEO and GEO shape each other, so we delivered all three as a single engagement.

Constraints

The launch was single-language Turkish, so every page had to work as the only version — no fallback locale, one market, one reading. The positioning demanded care: Puanla is not a review-filtering tool, and the site had to say so plainly while staying inside Google's policies and the KVKK-clean stance of collecting no personal data. And the pricing model — one annual package at ₺5.000/year + KDV per branch, a 10% discount on every branch for businesses with two or more branches (₺4.500/year + KDV each), no monthly plan — had to be communicated clearly enough that a visitor never has to ask.

What we owned

We owned the whole delivery: a from-scratch build on Next.js, the information architecture — Nasıl çalışır, Özellikler, Fiyatlar, SSS — and the Google-style design system the product itself promises. On top of the build we ran the technical SEO — crafted titles and meta descriptions, sitemap.xml and robots.txt — and the GEO/AEO layer, including a 13-question FAQPage schema in JSON-LD. We also built the conversion paths: the Giriş and Ücretsiz Başlayın CTAs, signup at /kayit, and a WhatsApp bubble for direct questions.

Our approach

We designed the interface the way the product describes itself — Google-style: white, Roboto, the #1a73e8 blue — so the site speaks the same language as the experience the customer's customer will see. The narrative runs in the product's own three steps — scan the QR, share the experience, review & feedback — with trust ticks under the hero (no credit card, Google-policy compliant, KVKK-clean) handling objections up front. Pricing became a single clear package instead of a plan matrix, and the GEO work was FAQ-driven: thirteen questions written the way business owners actually ask them, from how Google reviews affect Maps ranking to whether asking for a review is within Google's rules, all marked up as FAQPage schema.

The outcome

Puanla now runs on a site that carries its feedback-to-review flow end to end: a business owner lands, reads the three steps, sees the single price and signs up free at /kayit without a credit card. Search engines get crafted metadata, a sitemap and robots.txt; AI answer engines get thirteen answer-shaped questions in FAQPage schema, ready to quote. And the positioning boundary — policy-compliant review growth, not review gating — is stated on the page and in the schema, so wherever an answer gets assembled, the correct positioning is there to be quoted.