← Hub
DOC · YOREH-CRS-01 REV 2026.07 · CONFIDENTIAL

Deliverable

Customer
Review Strategy

Reviews as an AI-citation and trust signal.

Prepared by Node AI
Prepared for you, Fin Matson / yoreh.co
Issued July 2026
AT A GLANCE SCOPE READOUT
01
3
Third-Party Plays
02
6
Templates
03
~38
Current Reviews
04
$0–15
Monthly Cost
Key Insight

You already have the engine. Judge.me is installed and sending automated post-delivery requests, and your Klaviyo post-purchase flows are live. This is not a build from zero. The job is to map what runs today, merge it into one sequence so nobody gets asked twice, and add Trustpilot as a light second step on top.


01 /

The Core Thesis: Two Jobs, Three Venues

On-site schema, plus a third-party portfolio
The Point

Two jobs, three venues. Judge.me for on-site schema, and Trustpilot plus Reddit plus editorial listicles for the citations AI actually pulls. Don't bet your whole citation thesis on Trustpilot alone.

Start Here · What You Already Run Corrected per your feedback

An earlier version of this document was written as though Yoreh had no review system. That was wrong, and it matters, because it would have had you build a second one alongside the first.

Live today: Judge.me is installed and running, including automated review requests triggered when orders are delivered. Klaviyo post-purchase flows also exist. Roughly 38 reviews have accumulated. There is no SMS consent collected anywhere.

So the first task is not installation, it is an audit. Map the existing Judge.me automation and the existing Klaviyo flows side by side, then design one merged sequence. The failure mode to design against is a customer receiving two review asks from two systems, which reads as spam and costs you the review you were asking for.

Everything below still applies. Read it as "how the merged sequence should be shaped" rather than "what to install".

Reviews do two different jobs for you, and those jobs need different tools.

Job B is a portfolio, not a single site

The on-site app (Job A) and the third-party portfolio (Job B) do not substitute for each other. Judge.me alone gives you stars in Google but almost nothing AI cites as independent proof. Trustpilot alone gives you one citable venue but no per-product schema and no discovery coverage. No single third-party venue covers every query type, which is why Reddit and editorial are co-primary with Trustpilot, not a footnote to it. The Reddit and editorial tracks are owned by the Content & Editorial Strategy (its off-site engine); this document owns Job A and the Trustpilot venue, plus the collection system that feeds both.


02 /

Why the Third-Party Layer Is High-Leverage

The evidence, and its vendor caveat
The Point

The direction is real and worth acting on. The exact numbers are a vendor's, and the biggest lift belongs to a portfolio, not to Trustpilot alone.

The strongest current evidence is the Seer Interactive "What AI Says About You" study (commissioned by Trustpilot, March 2026): 800,000+ AI responses across 1,926 brands and 15,000+ buying-journey prompts, tested on ChatGPT, Gemini, Perplexity, and Google AI Mode.

~1%
Citation rate with
no third-party profile
~53.5%
With an established
Trustpilot presence
~75.3%
With 80+ reviews
plus regular responses

Review and trust sites are among the largest AI citation sources overall in the study, second only to brand-owned sites. Source: PR Newswire summary, corroborated by independent 2026 AEO research (Alhena on Google store ratings as AI signals).

Read The Numbers Carefully Vendor-commissioned

This is a Trustpilot-commissioned study, so I'd treat the magnitudes as directional, not literal. The 1% to 75% figures describe a mature profile, not your first month (see the cold-start math in section 03).

The study's framing that Trustpilot is the single most-cited review platform is vendor framing. Independent citation research consistently shows Reddit dominating discovery answers, especially on Perplexity, which is exactly why I've built your Job B as a portfolio and not a Trustpilot monoculture.

What survives independent scrutiny is the direction: establishing genuine third-party review presence materially increases AI citability, and Trustpilot is a heavily-cited venue for trust and reputation queries. Enough to build it. Not enough for me to promise you a citation on day 14.


03 /

The Cold-Start Reality

Honest math from a near-zero base
The Point

~38 reviews is a ~1% starting line, not a 53.5% one. Seed the list on day one, plan on sending ~1,000+ invites to net ~80, stage the target over 6 to 9 months, and put the jackets first.

You are not starting from the study's steady state. You are starting near zero, and I want the plan to say so plainly.

A staged target, not a cliff

Do not frame 80 as the launch goal. Stage it.

# Stage Trustpilot reviews Rough window What it buys
01 Seed 25, with responses Month 1 to 2 A profile that reads as real, not empty
02 Build 50, with responses Month 3 to 5 Approaching the study's citation band
03 Threshold 80+, with responses Month 6 to 9 The ~75% citation posture the study describes

04 /

Recommended Stack

~$0 to $15/mo, plus a free Google surface
The Point

Judge.me free for on-site schema, Trustpilot free for the trust venue. The Google Business Profile play is off the table, and the Google surface is reached through Merchant Center seller ratings instead.

What you run · ~$0/mo, unchanged

Total: $0/mo to start, $15/mo if Judge.me Awesome is added. No annual contract (Judge.me is month-to-month; Trustpilot's 12-month commitment applies only to paid plans).

On switching platforms · don't

The comparison below is kept for reference, but the platform decision is already made and you should not revisit it now. You are on Judge.me, it works, it emits the right schema, and your review history lives in it. Migrating would cost you the history and the working automation in exchange for visual polish.

For completeness: Loox Convert ($49.99/mo) is the one alternative with a real argument, best-in-class visual and video galleries plus social syndication. The schema outcome is essentially the same as Judge.me, so the premium buys presentation, not AI citability. Okendo, Stamped and Yotpo push a brand your size into $119 to $299+/mo bands quickly, and are not worth considering until much larger scale.

The Google surface · seller ratings, not a Business Profile

Google reviews feed Gemini and Google AI Overviews directly, and a Business Profile is a free surface, so an earlier draft of this document recommended claiming one against the Canggu address. That recommendation is withdrawn. The address it rested on is House of Yoreh, which is a former partner's business, not yours. Its two pages were removed from the theme when we applied your feedback.

Why This Changed Withdrawn on purpose

Claiming a Business Profile on that address would hand Google a verified link between Yoreh and a business you no longer have, at the same time as the rest of this engagement is unpicking exactly that association.

It is not a hypothetical cost. The baseline scan found 19 answers describing a coffee shop or cafe attached to the brand, and 18% of answers naming a co-founder who is not part of Yoreh. A Business Profile would be a fresh, authoritative source confirming both, and third-party sources are far harder to correct than your own pages.

What to do instead: reach the Google surface through store and seller ratings via Merchant Center plus an approved review partner (Judge.me syndicates here) once order volume supports it. Same surface, no entity risk. Either way, never incentivize a Google review.

Platform comparison (small Shopify jewelry brand, 2026)
Platform Free tier Entry paid Schema / Google stars Photo / Video AI-visibility role
Judge.me Yes, forever $15/mo flat Yes, on free plan Both, free On-site schema (Job A)
Loox No (trial) $9.99; video $49.99 Yes Photo all tiers; video $49.99 On-site schema, visual leader
Trustpilot Yes, 50 invites/mo $99/mo Off-site trust venue (no per-product schema) Yes Reputation-defense venue (Job B-a)
Google seller ratings Yes, free Free Feeds Gemini / AI Overviews (not product schema) Yes Free Google surface, via Merchant Center
Okendo No $19 (gated) Yes (stars gated to $299) Both Premium, overkill now
Yotpo 50 orders/mo $15+ Yes Both Overkill now
Stamped Removed $199+ Yes Both Overkill now

Pricing sources: Judge.me, Loox, Trustpilot, Yotpo, Okendo, Stamped.


05 /

The Collection System

Deliverability, timing, channel, sequence
The Point

Audit the two systems you already run and merge them into one sequence, fix deliverability, seed the list once, then let the delivery-triggered flow do the rest. Email only. Treat vendor conversion percentages as directional, and point the photo ask at the jackets.

Deliverability foundation (do this before any blast)

The seeding blast and the ongoing flow both die in spam folders if your sending domain is not set up. The unglamorous prerequisite, and it comes first.

Timing: trigger on delivery, never on purchase

A request that fires "X days after purchase" arrives anywhere from 2 to 20 days after the customer actually received the piece. Trigger off Fulfilled Order plus a delay longer than average delivery, or on the actual Delivered event if you add a tracking app (AfterShip, Malomo, Shopify delivered webhook). Jewelry forms an opinion fast but benefits from a few real wears, so 5 to 10 days post-delivery is the sweet spot. Because you have made-to-order pieces and international transit, branch on lead time:

Overshooting slightly is cheaper than asking someone whose piece is still in customs. A/B test 7 / 14 / 21 days per Judge.me's guidance.

Channel: email only

Email carries everything. SMS is out of scope. You told us Yoreh does not collect SMS consent, so there is no compliant list to send to and every SMS element has been removed from this document and from the templates. This is the right answer regardless: a first touch by text on a piece at this price point reads wrong for the brand, and US marketing SMS carries a real compliance load (express written consent, A2P 10DLC carrier registration, quiet hours) that is not worth clearing for the volume it would add.

If you ever do start collecting genuine marketing consent at checkout, revisit it then. Until that exists, treat SMS as unavailable rather than pending.

Sequence: 3 to 4 touches, plus a separate photo reminder
# Touch Timing (from post-delivery trigger) Job
01 Initial request Day 0 Ask for the review, single CTA, in-email star widget, product image
02 Reminder +3 to 5 days Gentle nudge to non-responders, new subject line
03 Last-chance (optional) +7 to 10 days Final soft ask, optional compliant incentive
04 Photo/video reminder Separate flow, +1 to 3 days after a text review One-tap photo upload, no rewriting required

Only continue for people who have not yet reviewed; separate the photo ask from the initial ask (asking for a photo up front lowers initial response); cap at 3 to 4 touches. The separate media-reminder is the single biggest structural lever for photo volume (Judge.me media reminder).

The Merged Sequence · Build Spec Email only · one ask per customer
STEP 0 (do this first): AUDIT
  |- List every Judge.me automated request currently firing (trigger + delay)
  |- List every Klaviyo post-purchase flow currently firing
  |- Decide which system OWNS the review ask. Turn the other one's ask OFF.
     (Recommended: Judge.me owns the ask, Klaviyo owns everything around it.)

MERGED FLOW (whichever system owns it)
TRIGGER: Delivered   (or Fulfilled Order + buffer if no tracking app)
  |
  |- Flow filter: has NOT already submitted a review (re-evaluate at each step)
  |
  |- TIME DELAY, then CONDITIONAL SPLIT: made-to-order tag OR international?
  |     |- YES -> longer buffer (~18 to 24 days total)
  |     |- NO  -> shorter buffer (~10 to 14 days total)
  |        (delay BEFORE the split; no delay right after a split)
  |
  |- EMAIL: Touch 1 (initial request)
  |- Time Delay +3 to 5 days -> split on "reviewed?" -> EMAIL Touch 2 (reminder)
  |- Time Delay +7 days      -> split on "reviewed?" -> EMAIL Touch 3 (last-chance)

SEPARATE FLOW A: photo prompt
  TRIGGER: Review submitted (text-only)
  |- Time Delay 1 to 3 days -> EMAIL: "add a photo" (one-tap, higher incentive for media)

SEPARATE FLOW B: Trustpilot second step
  TRIGGER: Review submitted, rating 4-5
  |- Time Delay 2 to 3 days -> EMAIL: Trustpilot invite (Template 06, no incentive)
     NOTE: this is a second-step ask to an already-happy reviewer, not a parallel
     campaign. It is NOT review gating: every customer still gets the Touch 1 ask
     above regardless of how they feel, and nobody is diverted from reviewing.

The rule that governs all of this: one review ask per customer per order. Two systems both asking is the single most likely way this goes wrong, and it is invisible until a customer complains. The Step 0 audit exists to prevent exactly that.

On engines: you already run both, so this is not a purchase decision. Judge.me's built-in request and media-reminder emails are the simplest owner for the ask itself. Klaviyo is better for the branching (delivery-aware timing, the made-to-order branch, the Trustpilot second step). Pick one owner for the ask, and do not let both fire.

Photo and video: the highest-leverage goal for jewelry

Visual UGC de-risks an expensive, alternative-aesthetic buy, and jacket-on-real-Oura photos double as citable community content for the off-site engine. Vendor-reported conversion lift (treat as vendor claims, not independent fact): PowerReviews reports video reviews lifting conversion 40 to 60%, photo reviews 30 to 45%, and products with customer photos converting up to ~91% higher. The direction is well-supported; the exact percentages are the vendor's. Get them via the separate one-tap media-reminder, a larger incentive for photo/video than text (Loox claims a fixed-amount credit drives ~40% more visual reviews than a percentage discount, again a vendor figure), and by prompting the exact shot. Prioritize prompting jacket buyers, where your product pages are barest.


06 /

Compliance

Non-negotiable
The Point

Genuine reviews only, incentives never tied to sentiment, no gating, and disclose. Add the EU Omnibus verification note, and keep AggregateRating at the product level so nobody chases Organization stars.

FTC Consumer Reviews and Testimonials Rule (effective Oct 21, 2024; up to $51,744 per violation)

Google goes further: incentives for Google reviews are banned outright. Incentivize on-site (Judge.me/Loox) reviews only. Never attach an incentive to a Google or Trustpilot review request.

EU Omnibus Disclosure Ships worldwide

You ship worldwide, so this applies. Under the EU Omnibus Directive, if you display consumer reviews you must state whether and how you verify that reviews come from real purchasers.

Add a short, plain review-policy statement near your on-site widget (for example: "Reviews are from verified customers who purchased this piece; incentivized reviews are labelled"). A cheap labelling obligation, easy to trip if reviews are shown with no verification statement at all.

Dev Note · Do Not Chase A Non-Result Schema placement

Put AggregateRating at the product level, not the Organization level. Google will not show star rich results for an Organization-level AggregateRating (self-serving "rate our company" markup is ineligible).

Product-level review markup on your jacket pages is fine and is the whole point of Job A. Flag this to your dev up front so they don't spend a day trying to make Organization stars appear.

Trustpilot / migration note: you cannot port historical reviews into Trustpilot; its reviews must be genuine and invitation-driven on its own domain. Your ~38 existing reviews seed the on-site app (Judge.me) only; Trustpilot is built fresh via the day-one seeding blast plus ongoing invitations.


07 /

Six Paste-Ready Templates

Your voice: spare, confident, "art not content"
The Point

Four on-site touches and one Trustpilot invite, all email. Merge tags are Klaviyo-style; subject lines under 40 characters; the Trustpilotinvite is sentiment-neutral by design.

Template 01 · Initial Request Email · Touch 1
SubjectHow's the {{ product_title }}?
PreviewTwo minutes, if you have them.
{{ first_name }}, You didn't pick something everyone else is wearing. That's the point. Now that you've lived with your {{ product_title }} for a bit, we'd like to know how it's holding up, on the hand, in daily wear, all of it. [ ★ ★ ★ ★ ★ ] ← tap to leave your review [ Leave a review ] It takes about two minutes, and it helps the next person who doesn't blend in find the piece. Yoreh
Template 02 · Reminder Email · Touch 2 · +3 to 5d
SubjectStill thinking it over?
PreviewNo pressure. Two minutes when you're ready.
{{ first_name }}, Quick one. If you've had a moment with your {{ product_title }}, we'd still love to hear what you think, any rating, honest as you like. [ Leave a review ] If it's not landing the way you hoped, tell us that too. We'd rather know. Yoreh
Template 03 · Photo Reminder Email · Separate flow
Subject{{ first_name }}, add a photo
PreviewShow us how you wear it.
{{ first_name }}, Thanks for the review of your {{ product_title }}. It means a lot. One more thing, if you're up for it: add a photo. Show us how you style it, on the hand, on the neck, in your world. [ Add a photo ] (one tap, no rewriting) Real pieces on real people is the only marketing we actually believe in. Art, not content. Yoreh
Template 04 · Last-Chance + Incentive Email · Touch 3 · Optional
SubjectOne last thing
PreviewAn honest review, any rating.
{{ first_name }}, Last time we'll ask. If you leave an honest review of your {{ product_title }}, any rating, we'll send you {{ credit_amount }} off your next piece as a thank-you. [ Leave a review ] To be clear: the reward is for the review, not for a good one. Tell us the truth. Yoreh Reviewers receive a store credit for leaving an honest review, regardless of rating.

The italic disclosure line is required by the FTC rule. Keep it on any incentivized touch. Incentives are for on-site reviews only, never for Trustpilot or Google.

Template 05 · Removed Was the SMS variant

This slot held an SMS variant of the review request. It is removed, because Yoreh does not collect SMS marketing consent, so there is no list it could legally be sent to. Nothing replaces it: the email sequence carries the whole job.

Template 06 · Trustpilot Invitation Email · No incentive
Subject60 seconds on Trustpilot?
PreviewYour honest take, whatever it was.
{{ first_name }}, We're a small independent studio, and independent reviews are how people decide whether we're the real thing. Whatever your experience with your {{ product_title }} was, good, mixed, or not for you, a quick word on Trustpilot goes a long way. [ Review us on Trustpilot ] Share your honest experience, whatever it was. That's the whole point. Yoreh

Sentiment-neutral by design (no "vouch for us," no "if it earned it"), and no incentive on the Trustpilot or any Google invite, per platform rules and the compliance section.


08 /

Setup Checklist

Deliverability first, jackets weighted throughout
The Point

Audit what already runs, merge it into one sequence, then add Trustpilot as a second step. Email only. No Google Business Profile. Product-level schema, no Organization stars.

Audit first (before changing anything)
Deliverability
On-site (Judge.me, already installed)
Third-party (Trustpilot)
Google (seller ratings only)
Automation (the merged sequence)
Ongoing

09 /

Success Metrics

Staged, jacket-weighted, honest about timing
The Point

Judge.me jacket stars are the earliest visible win. Trustpilot citation is a later-scan outcome, not a launch metric. Reddit and editorial are measured in the content document.


§ /

Sources

Evidence, pricing, compliance

Seer/Trustpilot "What AI Says About You" (March 2026, Trustpilot-commissioned, magnitudes directional): PR Newswire; independent corroboration Alhena. Platform pricing: Judge.me, Loox, Trustpilot. Collection practice: Judge.me review requests, Judge.me media reminder, Omnisend product-review email. Conversion figures (vendor claims): Loox academy, PowerReviews UGC conversion. Compliance: FTC Reviews Rule Q&A, Google UGC policy, EU Omnibus Directive (review-verification disclosure), Gmail/Yahoo bulk-sender requirements (SPF/DKIM/DMARC). Cross-reference: Yoreh Content & Editorial Strategy (Reddit + editorial off-site engine).