Partnerships and referral fees that hold up

Two companies agree to send each other business. There is a lunch, sometimes two. Somebody says the word "synergy" without noticing they said it. Logos go on each other's websites under a heading like "Partners," a Slack channel gets created, and for about six weeks people check it. Then a referral shows up that nobody remembers agreeing to pay for, or doesn't show up at all despite eleven meetings that all seemed promising, and the partnership becomes the kind of thing that quietly stops being mentioned in the pipeline review.

That failure pattern is common enough to take seriously as evidence rather than bad luck. Most partnership advice treats the problem as matchmaking — find the right partner, get the intro, build the relationship — and that part genuinely is the easy half. Two companies with overlapping customers and non-competing products find each other constantly, at conferences, through mutual investors, in the comments of a mutual customer's post. What almost never gets discussed, and what actually kills the arrangement, is mechanism: the specific moment a referral is supposed to happen, and the specific rule for deciding whose customer it was when both sides plausibly touched the deal. A fee schedule without an attribution rule does not prevent a dispute. It guarantees one, usually by month four, over a deal large enough that neither side wants to be generous about it.

The trigger has to be a sentence, not a feeling

Ask most people running a partnership what causes a referral and you get something like "when it's a good fit" or "when they need what we do." That is not a trigger. A trigger is the specific situation, describable in one sentence, in which your partner's contact says your name.

A working version sounds like this: when a client of the accounting firm asks about payroll software during the year-end close conversation, the accountant names the payroll vendor. That is testable, teachable, and — this is the part that matters — repeatable by someone who was not in the room when the partnership was agreed. A vague trigger does not survive the account manager who set it up moving to a different role; it lived entirely in one person's judgment, and when that person leaves, the referrals stop even though nobody decided to.

Most partnerships never get this specific because writing the sentence forces an uncomfortable admission: for a lot of intended partnerships, the trigger moment does not actually occur very often. The accountant is asked about payroll during onboarding, not during year-end close, and onboarding happens once. If you cannot write the sentence, you have not found a partnership. You have found two companies that like each other, which is a different and less useful thing.

Being referred takes less generosity than people assume, and more specificity

Assuming the trigger is real, the second failure is that most companies make themselves hard to refer without noticing they've done it. A partner willing to say your name still needs to say something specific, and most partners won't study your product deeply enough to explain it well from memory. They will refer you in exactly the words you gave them, or badly, or — most often — they will hesitate because they aren't sure what to say and let it pass.

The fix is unglamorous and almost nobody does it properly: one sentence describing what you do in language the referring company's client would use, one link that goes straight to the thing relevant to that referral rather than a homepage, and one named contact who will personally answer within a day. Not a form, not a general inbox — a person, because a partner vouching for you is putting their own credibility on the line and wants to know it lands somewhere responsive. Companies that hand partners a full deck and three case studies instead have made the referral harder — nobody reads a deck in the thirty seconds where the trigger moment occurs.

The fee schedule decides what behavior you get, whether you meant it to or not

Once the mechanism is real, the money has to match it, and the shape of the fee changes what the partner actually does — not just how much they're paid.

A flat fee per qualified referral rewards volume. It is simple to administer and works well when deal sizes within the partnership are roughly similar, because a flat number that's fair for a typical deal becomes either an insult or a windfall once sizes vary widely. It also does nothing to reward a partner for sending a better lead over a worse one, which is fine if you qualify leads yourself and mildly dangerous if you can't.

A percentage of revenue, usually of first-year contract value, rewards bigger deals and aligns incentives for as long as it applies. The problem is where it stops. A percentage that runs forever means paying a partner in year five for an introduction made in year one — which sounds generous until the partner is earning more from that account, annually, than the sales rep who has to renew it.

A first-year-only percentage, ending after twelve months, is the compromise most software companies land on: it rewards the introduction without becoming a permanent royalty. The trade-off is that it gives partners a reason to push hard in year one and lose interest after — fine for a one-time referral, a problem if you wanted help with renewals too.

Structure	Rewards	Where it breaks down
Flat fee per referral	Volume, simplicity	Unfair when deal sizes vary widely
Percentage, uncapped	Bigger deals, ongoing alignment	Becomes a permanent royalty on old work
Percentage, first year only	Bigger deals, clean end point	Partner interest can drop after month twelve

None of these is correct in the abstract. The question is not "what's standard" but "what behavior do I want from this partner, and does the fee reward that or a different one." A partner you want introducing you once wants a flat fee or a capped percentage. A partner you're hoping will sell alongside you for years wants something closer to uncapped, or they will correctly notice they're carrying more of the relationship than they're paid for.

Write the attribution rule before there is anything to attribute

Here is the sentence that should exist in every partnership agreement and almost never does: what happens when the customer was already talking to us.

This is the single most common referral dispute, and it is predictable enough that it should be solved in the contract rather than in a tense call four months from now. A partner introduces a company you'd already emailed eight weeks earlier and never heard back from. Did they refer that customer, or did you already have them? Both sides have a genuinely defensible position, which is exactly the condition under which disputes get ugly.

The rule needs three parts, written down before the first referral, not negotiated after the first disputed one. First, a definition of what counts as a referral at all — an introduction, a named contact passed along, not a vague mention that a company exists. Second, a timing rule: if you already had an open conversation with that contact within some agreed window, commonly ninety days, the referral doesn't count — and the window needs to be a number, not "recently." Third, a source of truth: whichever CRM record has the earliest timestamped contact wins, and both sides agree in advance to trust that record over memory. None of this is exciting to write. All of it is cheaper to write once than to litigate once.

Reciprocity has to be structural, not sentimental

The last failure mode is quieter than the other two. A partnership where referrals flow mostly one direction survives on goodwill for a while and then stops, without anyone announcing it — the referring side gradually finds other things to spend attention on, because a partner who sends five leads a quarter and receives none eventually notices, even if they never say so out loud.

The instinct is to fix this with a conversation reminding the underperforming side that the partnership is supposed to be mutual. That rarely works, because the imbalance usually isn't a motivation problem, it's a mechanism problem: one side's customers naturally need the other company more often than the reverse does. A payroll company's clients need an accountant constantly; an accountant's clients need a new payroll vendor rarely, only at a switch. No amount of goodwill evens that out. The honest fix is to pay for the imbalance directly — a higher fee flowing one way — or to accept the partnership is an acquisition channel for one side and a service relationship for the other, and stop pretending otherwise.

This is the same discipline that applies to any channel you're relying on for growth: know what it actually produces before you build a program around what you hope it produces, a test we've applied more broadly in acquisition channels that survive scrutiny. Referral specifically has its own blind spot worth reading before you formalize anything — most companies already have a referral rate happening informally that nobody has measured, in the referral channel nobody measures. Formalizing a partnership before you know your baseline is how you end up paying for referrals that would have happened anyway.

What this comes down to

Structure the partnership around the dispute you know is coming, not the launch lunch. Write the trigger sentence first, and if you cannot write it, you do not have a partnership yet. Make the referral easy with one sentence, one link and one named person, because the moment a partner hesitates over what to say is the moment you lose the referral. Pick a fee structure for the behavior you actually want, and put the attribution rule in writing before the first referral, because it will otherwise get written during the first dispute, under worse terms, by whoever is angrier. Compare that against SEO and paid search on a small budget: those channels fail loudly, in a dashboard, within a quarter. Partnerships fail quietly, over a slow erosion of attention, and by the time it's obvious the relationship is gone, so is the goodwill needed to fix it.