A referral lifecycle that keeps trust intact

The project started inside a payment flow that already worked. Our import and export customers were paying their suppliers on the platform, and from the supplier's side the whole experience was three steps: a payment email arrived, they clicked a link, they chose how to receive funds. For most suppliers, that email was their first contact with our product. No campaign, no form, no demo, just money showing up correctly.

The question I was handed: could that moment become adoption? These suppliers were already using the product in the most convincing way possible. Getting them to open their own account looked like the shortest funnel I'd ever seen.

My first pass treated them like leads, because that's what our funnel knew how to do. New contact, Salesforce record, nurture emails, sales follow-up on a timer. The complaints arrived before the results did, from our own customers, the importers who were paying these suppliers. They did not enjoy watching people they do business with get worked by our sales team. One message, lightly paraphrased, ended version one:

We're glad to introduce you to our suppliers. Please don't have anyone call them.
an existing customer, politely, and they were right

That's the trap in a referral flow: two relationships are in play at once. Push a supplier too hard and you damage the new relationship and the existing customer's. Their name is effectively on the email. So I tore up the standard funnel and rebuilt the lifecycle around a different question: what would make them want an account?

Two relationships are in play: the supplier's trust is borrowed from the importer whose name is on the email, so a cold call damages both at once.

The success signal is a completed payment

My first rebuild decision was defining success honestly. Registration and KYC completion were necessary steps on the way to the outcome the business cared about. The real signal was a supplier completing a real payment on the platform after clearing KYC. A behavior, in other words. I stopped measuring the lifecycle by registration speed or KYC time at all.

That definition forced me to respect the mindset these users arrived in. They came through a payment link to receive money. Trust didn't start at zero. One side of their business relationship was already our customer. But it was borrowed trust, and borrowed trust does not survive a cold call.

Learning before commitment

The core design was a light version of the product for newly referred suppliers, mine to spec and product's to build. The limits were on usage. Volumes were capped, but everything that made the product worth having was visible from day one: fast transfers, flexible FX rates, a clean operational dashboard. I'd like to print the exact caps here, but it's payments data and compliance reads this blog. The principle survives without them: cap quantity while keeping every capability. A limited account that still demonstrates value sells the upgrade for you.

Around that light version I ran email journeys keyed to user state. State lived in Salesforce, sends went out through HubSpot, and the Zapier bridges I built moved status changes between the two: KYC initiated, KYC cleared, and first payment completed each flipped a supplier from one journey to the next automatically. Early-journey messaging stayed educational, and KYC was always positioned as the step that opens full value. It was never framed as a finish line or as success by itself.

Delaying sales on purpose

The most consequential call I made was keeping sales out deliberately, because our customers had told us to, in writing. Privacy, relationship boundaries, control over their own supplier network: the concerns were concrete, and respecting them mattered more than accelerating outreach.

So I ran the early stage on three things: product-led exposure inside the payment flow itself, retargeting audiences I built from payment-link visitors, and email to opted-in suppliers only. Messaging stayed on practical value, instant payments, better FX flexibility, less operational drag, and never pushed conversion. The goal was for suppliers to reach their own conclusion.

⛨ Compliance note: no supplier received marketing email without an explicit opt-in, enforced as a hard consent gate in the CRM that every journey checked before sending. In a referral flow the sender's reputation is borrowed from the existing customer. Consent is a legal requirement, and here it is also how you avoid spending someone else's trust.

Sales entered on exactly one trigger: the supplier making an active move toward a full account. When someone initiated registration and started preparing for KYC, Zapier flipped their Salesforce stage and a rep picked up the account with full context. At that point the conversation felt like someone helping you through KYC rather than someone selling at you. Same rep, same product, completely different reception. The difference was that the user opened the door first.

Every journey change hangs off a state rather than a send date, and the rep arrives to help with KYC instead of selling into it. Borrowed trust does not survive a cold call.

Lifecycle design as sales-time protection

Delaying sales was a trust decision, and the internal economics pointed the same way. Sales time is expensive, and a supplier who received one payment last Tuesday is mostly not ready for a conversation. Letting automation handle the early education meant reps only touched accounts that had already raised a hand, instead of working a list of people who never asked to be on one.

The discipline underneath all of it was treating referred suppliers as their own segment, with their own journeys, their own triggers, and their own definition of done. For scale context: the cross-border payments line this work fed grew from $2M to $76M in its first year and past $250M in its second. I won't claim the lifecycle did that, product and pricing did the heavy lifting, but that line grew through referrals the entire way, and you don't get two years of referral growth if a supplier's first week with you feels like a timeshare pitch.

The line this lifecycle fed: $2M to $76M in year one, past $250M in year two. That kind of growth only compounds when a supplier's first week doesn't feel like a pitch.

Questions I keep getting

What improves lifecycle performance fastest?

In this project, one change beat everything else: moving triggers off the calendar and onto user state. A “day 3” email is a guess; a “KYC cleared yesterday” email is an answer. Once sends keyed off Salesforce status instead of send dates, the same copy worked noticeably harder.

Why do onboarding and email systems break at scale?

For us the failure mode was sync. KYC status lived in Salesforce, email ran in HubSpot, and any lag between them meant congratulating a supplier on a step they hadn't taken, or nudging one who'd already finished. The unglamorous Zapier work I did keeping the two systems honest with each other mattered more than any subject line I wrote.

What should teams automate first?

The state changes. Ours were KYC initiated, KYC cleared, and first payment completed, each moved the supplier between journeys, and exactly one of them handed the account to sales. Automate those transitions before anything creative. They're where users predictably stall, and where a mistimed manual step costs trust you can't buy back.

What I'd tell a peer building this

Four calls carried this system, and I'd make all four again. Define success as a behavior (for us, a real payment after KYC) and delete every metric that rewards rushing it. Cap usage while keeping every capability. Make consent a hard gate in the CRM. And give sales exactly one entry trigger, owned by the user. The journeys, the Zapier plumbing, the retargeting: everything else was in service of those four decisions.