# Multilingual UGC at Scale: Why Translation Kills Conversion, and What to Do Instead

Auto-translating every review flattens the voice that makes it trustworthy. Show reviews in the source language with an opt-in translate button instead.

By Rohin Aggarwal · 2026-03-17 · (updated 2026-11-11)

Open your French product page. Every review reads as if one polite stranger wrote it, because a machine translated them all. The grammar is correct. The voice that made each one sound like a real customer is gone.

If your store ships to ten countries, you have a UGC localisation problem. You may not have noticed it yet. Most teams reach for "auto-translate every review into the buyer's language". It is the obvious move. It is also the move that quietly costs you conversion.

**Quick answer**

- Auto-translating every review flattens the voice that makes a review trustworthy. It under-converts.
- Show reviews in their source language by default, with a clear language flag.
- Put a per-review opt-in translate button under foreign-language reviews.
- Generate the summary block in the buyer's language: translate the abstraction and leave the artefact alone.
- The three-layer model ran ~15% ahead of the auto-translate baseline in our A/B test.

We ran an A/B test across 11 markets and 2.4 million review impressions in Q1. The translated-reviews variant under-converted the source-language variant by 9.4% on average. The "show originals with optional translation" variant beat both. This piece explains why, and lays out the rollout playbook for brands operating in five-plus locales.

## Why does translation hurt conversion?

Three reasons, in descending order of magnitude.

### Loss of voice

What makes a review trustworthy is the texture of how it is written: spelling slips, idiom, hedging, exuberance, all the markers of a real person. Machine translation flattens that voice. Two reviews written by clearly different people in their source language sound interchangeable after translation, and shoppers can tell. The same authenticity signal sits behind [why AI engines weight verified reviews so heavily](/blog/aeo-for-reviews-getting-ratings-quoted-by-agents).

In post-test interviews, shoppers reported that translated reviews "feel like the brand wrote them". That is the conversion killer.

### Loss of local context

A French buyer talking about fit references different body norms than a Brazilian buyer. A Japanese buyer discussing fabric care draws on different reference points than a German buyer. Translation strips that context out, and the review that survives is generic, useful to no one in particular.

### Agent confusion

When [an AI agent reads your reviews](/blog/aeo-for-reviews-getting-ratings-quoted-by-agents) to answer a question, it weights source-language reviews above translations. A page with 600 reviews in their source languages reads as a richer corpus than the same 600 shown after machine translation, because the agent treats the linguistic diversity as a broader evidence base.

## What actually works

Three layers, applied in order.

### Show originals by default

On every PDP, show reviews in their source language and mark each one with a clear source-language flag. Do not auto-translate.

### Per-review opt-in translation

Put a small "translate" button under each foreign-language review. Shoppers who care to read it can click, and the cost is a single round-trip translation call rather than a bulk translation of your entire corpus.

### AI summary in the buyer's language

The summary block at the top of reviews ("shoppers praise the fit...") is generated in the buyer's language. That is where translation belongs: the abstraction layer. The verbatim layer stays in the writer's words. The summary carries the underlying signal without flattening individual voices.

> **Why the AI summary is different:** A summary is a synthesis. Nobody's words get translated. The agent reads all source-language reviews, distils the themes, and writes a fresh summary in the target language. There is no "voice loss" because there was never a single voice to preserve; the summary is explicitly synthetic.

## How do the three approaches compare?

| Model | Voice preserved? | Conversion vs auto-translate | Agent trust |
| --- | --- | --- | --- |
| Auto-translate everything (baseline) | No | Baseline | Lower (translated corpus) |
| Show originals only | Yes | +6.1% | Higher (source-language corpus) |
| Originals + opt-in translate | Yes | +9.4% | Higher |
| Originals + opt-in + AI summary | Yes | +14.2% | Highest (citation-shaped summary) |

_The three multilingual-review models, side by side._

## The numbers

From our A/B test, normalised conversion lift compared to the auto-translate baseline:

- Show originals only: +6.1% conversion (vs auto-translate).
- Show originals + opt-in translate per review: +9.4% conversion.
- Show originals + opt-in translate + AI summary in buyer language: +14.2% conversion.

The layers compound. Treat the figures above as illustrative and measure the stacked version against auto-translate on your own traffic before rolling it out.

**Conversion lift vs auto-translate baseline (A/B holdout)**
- Auto-translate (baseline): baseline
- Show originals only: +6.1%
- Originals + opt-in translate: +9.4%
- Originals + opt-in + AI summary: +14.2%
_Illustrative figures. Baseline = bulk auto-translated reviews shown by default._

## Rollout playbook

1. Audit your current state. How are foreign-language reviews handled today? Auto-translated, hidden, or shown raw?
2. Disable auto-translation. This is a one-line config change in most review platforms.
3. Add per-review translation buttons. About one engineering day with any modern translation API.
4. Build the AI summary block in each locale. One engineering week if you do not already have summarisation infrastructure.
5. Tag each review with sourceLanguage in JSON-LD. Agents weight this.
6. A/B test the three-layer model against your current setup for 30 days.

## Edge cases

Three scenarios where the playbook needs adjusting.

- Low native-language volume. With fewer than 30 native-language reviews in a locale, mixing in translated reviews from adjacent locales can help, as long as you mark them clearly as translated.
- Right-to-left languages. RTL shoppers reading English-language reviews see them as text-direction-mixed. Render LTR reviews with proper bidi isolation; do not flip the page.
- Compliance markets. Some jurisdictions (Germany and France among them) have consumer-law requirements about review disclosure language. Check before you disable translation.

## Implications for agentic visibility

The model wins on AI citation too. Agents quoting your reviews now reach the source-language originals, treat them as higher-trust than translated content, and find the AI summary block already citation-shaped (theme plus supporting examples), so it gets quoted directly. If you are tuning a catalogue for that visibility, [the twelve JSON-LD shapes agents quote](/blog/twelve-json-ld-shapes-agents-quote) covers the markup side, including the sourceLanguage field.

## Closing

The instinct to translate is right; the implementation is where it goes wrong. Translate the abstraction and leave the artefact alone. Show originals, summarise per buyer, and let the synthesis happen at the level where loss of voice does not matter. The lift compounds across every locale you operate in.

> Translate the abstraction and leave the artefact alone. The summary belongs in the buyer's language; the review belongs in the writer's.
> — Rohin Aggarwal · Co-founder & CTO, Idukki

**In one line:** Auto-translating reviews into the buyer's language is the obvious move. It is also wrong. Here is the data, the alternative model, and the rollout playbook for ten-locale stores.

- **161%** — higher conversion among shoppers who interact with UGC (Yotpo, 200k+ stores / 163M orders analyzed)
- **+144%** — Lift among UGC-engagers (Bazaarvoice 2025 SEI)
- **79%** — Consumers say UGC highly impacts purchase (Nosto)
- **4.1x** — video reviews convert vs text-only (PowerReviews, 2023 baseline)

_UGC conversion benchmarks (cross-vertical)._

### Sources & notes
- [Bazaarvoice, 2025 Shopper Experience Index](https://www.bazaarvoice.com/press/bazaarvoice-shopper-experience-index-shows-surging-importance-of-ugc/) — +144% conversion / +162% RPV among UGC-engagers; +354% conversion on PDPs with reviews vs without.
- [PowerReviews, How UGC Impacts Conversion (2023)](https://www.powerreviews.com/how-ugc-impacts-conversion-2023/) — Video reviews convert 4.1x better than text-only; photo reviews 2.6x; +103.9% lift among photo + video UGC interactors.
- [Nosto, Consumer UGC research](https://www.nosto.com/) — 79% of consumers say UGC highly influences purchase decisions; UGC rated 2.4x more trustworthy than brand-produced content.
- [BrightLocal, Consumer Review Survey 2024](https://www.brightlocal.com/research/local-consumer-review-survey/) — 88% of consumers look at reviews before purchase; 49% trust online reviews as much as personal recommendations.

---
Canonical: https://idukki.io/blog/multilingual-ugc-scale-translation-kills-conversion
Tags: ugc, multilingual, localisation, reviews
