Executive Summary
I created an account, walked Nivoda's registration flow the way a new retailer would, browsed the live marketplace, and read what retailers say on Trustpilot and in Professional Jeweller. This is what I found, organized as four problems I would start solving in my first weeks, each with the evidence, the fix, and the metric it should move.
The four findings:
- The front door asks for trust before it shows value: the signup form front-loads ID warnings and a pre-checked three-channel marketing consent before a single diamond is shown.
- The step from signup to first purchase could use a clearer guided path: buying runs through verification first, and a new retailer has little in-product sense of where they are in that journey or what unlocks the first order.
- There's room to define what an "activated" Nivoda retailer looks like: today every new retailer gets the same default experience, sorted price low to high, without a clear activation target to design toward.
- The repeat-order engine leans on the retailer's memory more than the product: saved searches, holds, and enquiries are unconsumed demand the product doesn't loop back on.
Three things already work well and are worth protecting: the honest country-first registration step, the discount-to-list price anchor on every stone card, and the retail tools that make Nivoda's feeds a switching-cost moat.
One caveat up front: I don't have the funnel data, and my account never transacted. Some of these findings may already be on the roadmap, or look different once real numbers are attached. Treat this as a demonstration of how I work, not a verdict on the product.
Why this exists
Nivoda's numbers are public and impressive: past $1 billion in cumulative diamond sales, 100+ countries, a $51M round led by Northzone, and a 2026 expansion into finished jewelry. The growth role owns the funnel that has to keep that compounding: acquisition through conversion and early retention.
I created an account, walked the registration flow the way a new retailer would, browsed the live marketplace, and read what retailers say about Nivoda on Trustpilot and in Professional Jeweller. This document walks each of the four problems in turn, with the evidence, the fix, and the metric it should move.
1. The front door asks for trust before it shows value
What I saw. The homepage says "Join for Free" and "Join 10,000+ jewelers for free." Clicking it leads to a form that, before showing a single diamond, asks for first name, last name, business email, phone number, and a password with five rules, and warns me to have my passport, driver's license, or residence permit ready. A helper text under the email field says "Use your business email to unlock 2M+ products faster." It hints that a personal email (a Gmail address, say) leads to a different experience, without spelling out what actually changes.
Why it matters. Verification is non-negotiable in this industry, and the country-first step ("Select your country so we can apply the right verification process") is genuinely well done. There's an opportunity, though, in the difference between requiring verification to transact and introducing it on the very first screen, before any value has been shown. Sean Ellis's formula1 is a useful lens: conversion tracks desire minus friction. Right now the form asks for a lot of friction at the moment desire is least established, before the prospect has seen the inventory depth, the discount-to-list pricing, or a 360° video.
What I'd do. Instrument every step of registration and measure drop-off per field. Then test flipping the funnel: let a prospect run a search and see real stones with prices before the full form, and move the ID warning from the account-creation screen to the verification step itself, where it's relevant. The "business email" nudge should either state its benefit concretely or be tested against removal. And the marketing consent checkbox for email, SMS, and WhatsApp comes pre-checked, which tends to grow the list faster than it grows list quality or protects sender reputation, the kind of number that can look strong on a dashboard while working against the funnel behind it.
The metric. Landing-page visitor to completed registration, segmented by traffic source, plus registration to verified account within 7 days.
2. The step from signup to first purchase could use a clearer guided path
What I saw. Nivoda's promise is that buying is "as easy as online shopping," and the homepage even previews the retailer view with per-stone prices and an Add to cart button on every card. Getting there, though, runs through business verification first, something the signup flow surfaces early with its friendly "Have your ID ready" notice. What I couldn't get a feel for from the outside was the journey in between: where a new retailer stands in verification, what's left, and what buying looks like once it's done. (My own account was on the supply side, so I'm describing the pre-purchase path a new retailer walks toward that first order, not the verified buyer screen itself.)
Why it matters. For a marketplace that positions buying as effortless, the stretch between signing up and placing a first order is the conversion moment, and it leans on the new retailer staying motivated through verification. A little visible progress and reassurance there tends to go a long way, since interest can cool while someone waits without knowing how long the wait is or what comes next. It's the same pattern I've seen elsewhere: when a necessary step isn't explained, people experience it as a stopping point rather than a next step.
What I'd do. A gentle way in would be to make the verification journey visible and encouraging for new accounts: a simple progress view of what's done, what's left, and how long it usually takes, plus a friendly reminder of what it unlocks. Pairing that with a short, well-timed follow-up while interest is still fresh, and giving people a preview of the buying experience during the wait, could help more new retailers carry their momentum through to a first order. Much of this is likely wiring rather than new capability, since the verification status already exists in the system.
The metric. Registration to verified, verified to first order, and time-to-first-order. That second conversion, turning a verified account into a first purchase, is likely one of the highest-leverage numbers in the funnel.
3. Defining what an "activated" Nivoda retailer looks like
What I saw. From the outside, the first-run experience treats every new retailer identically: same search page, same defaults, sorted by price low to high, which surfaces the lowest-priced, least representative stones first rather than what a given retailer actually stocks. Retailers come in different shapes and forms. An engagement-ring independent, a 10-store chain feeding its website by API, and a jewelry manufacturer buying melee have different first orders, different cadences, and different aha moments.
Why it matters. Facebook found "7 friends in 10 days."2 Slack found 2,000 team messages.3 These thresholds matter because they convert retention from a hope into a design target. Reading Nivoda's retailers' own words suggests where its version lives: what they rave about is the first time a stone arrives quality-checked, on time, without a $150 FedEx bill and a week of supplier back-and-forth. The behavioral predictor is likely upstream of that: something like completing a first search with media filters on, placing a first hold, or requesting a first "eye clean?" check within the first week.
What I'd do. Run cohort analysis on existing buyer data to find the early behaviors that best predict a retailer still ordering in month 3. Then rebuild the first week around driving that specific behavior: segment-aware defaults at signup (natural vs. lab-grown vs. melee focus), a first-search experience that guides while the retailer does the real task rather than a tour that narrates, and a default sort that reflects buyer intent instead of ascending price.
The metric. Percentage of new verified accounts hitting the activation threshold in week 1, validated against month-3 retention.
4. The repeat-order engine leans on the retailer's memory more than the product
What I saw. Nivoda's model means a retailer's need is spiky and demand-driven: a customer walks into their store, and now they need a 1.5ct oval in a specific budget. The product has saved searches, holds, and enquiries, all signals of unconsumed demand, but from what I can observe, the loop back to the retailer relies mostly on the retailer remembering Nivoda exists at the right moment. Meanwhile the trade-credit offering (30/60/90-day terms through the i80 facility) is a structural retention weapon that lives mostly on the marketing site.
Why it matters. Early retention is in this role's mandate for a reason: in a marketplace, month-2 buyer retention compounds into GMV faster than any acquisition win. The intent signals already exist, which makes the trigger design fairly natural to build. A saved search matching a newly listed stone, a held stone about to expire, a price drop on an enquired stone. These are Fogg-model4 signal triggers aimed at users who already demonstrated motivation, which is the only kind of notification that builds a habit instead of an unsubscribe.
What I'd do. Build and test the top three demand-recapture loops: new-match alerts on saved searches, hold-expiry nudges with one-tap extension, and price-movement alerts on enquired stones. Position credit terms inside the product at the moment of a large first order, not just on the website. And for accounts gone quiet after a first order, a win-back built around what actually changed in inventory for their segment, not a generic "we miss you."
The metric. Month-2 active buyer retention and repeat-order share of GMV, cohorted by first-order category.
What's already working
Three things I'd protect rather than change. The country-first registration step sets verification expectations honestly and early. The stone cards themselves are excellent merchandising: discount-to-list on every card is a permanent price anchor that does silent selling on every search. And the retail tools (virtual showroom, website feeds via API) are a textbook "come for the tool" wedge: a retailer whose website inventory runs on Nivoda's feed has switching costs no discount can beat, which makes the feeds a retention asset worth treating as part of the growth funnel, not a side product.
Where I'd start
Item 2 first. It's relatively low-effort, it sits right at the moment a new retailer is deciding whether to push through to that first order, and it's the most likely of the four to be shaping conversion today. Item 1's instrumentation runs in parallel because everything else depends on knowing the real drop-off numbers. Item 3 is the biggest prize and needs data access before the design work is credible. Item 4 builds the machine that keeps the wins compounding after the first quarter.
A note on scope
This was built from one registration walk-through, a browsing session on a non-transacting, supplier-registered account, and public sources (Trustpilot, Professional Jeweller, funding coverage, Nivoda's own site). I never saw the verified buyer experience, the funnel data, or session recordings, which is exactly the data I'd want in week 1 before committing to this priority order. Treat the sequence as a hypothesis. The method is the point.
Notes
- Sean Ellis, the growth marketer who coined the term "growth hacking" and popularized the product-market-fit survey; "conversion tracks desire minus friction" is a common distillation of his approach to onboarding and activation. ↩
- Facebook's "7 friends in 10 days," the early activation threshold its growth team found best predicted long-term retention. ↩
- Slack's widely cited benchmark that teams exchanging around 2,000 messages were far more likely to stick. ↩
- The Fogg Behavior Model (BJ Fogg, Stanford): a behavior occurs when motivation, ability, and a prompt converge at the same moment. Triggers work best when aimed at users who already have the motivation. ↩