Which Referral Sources Actually Pay: Attribution Through to Fee

Ask a managing partner which referral source is their best and you'll usually get a confident answer.

Ask how they know, and it's normally some combination of who calls most, who they see at events, and a general sense of things. That sense is often wrong, and it's wrong in an expensive direction — the relationship absorbing the most partner time is frequently not the one producing fees.

Volume and value are different rankings

A source can send you thirty enquiries a quarter and produce almost nothing. Another sends four and two of them are significant matters.

The first source feels like the good one. Somebody is calling every week, the relationship is warm, and there's a lunch in the calendar. The second is quiet and easy to under-invest in.

Ranked by fee earned, they trade places. And because nobody's produced that ranking, the firm keeps allocating attention by feel.

Three things make sources look better than they are:

• Volume without fit. They send everything, most of it isn't your practice area.
• Cases that don't convert. Signed, then withdrawn or dismissed. Work done, no fee.
• Referral fees on the way out. A source producing $200k in fees where half goes back out on referral is a different proposition than one producing $150k you keep.

Capture it once, at intake

The whole thing depends on one habit: recording the source as structured data at first contact, every time.

Not a free-text "how did you hear about us" that gets filled in half the time. A required field, from a controlled list, on the intake form.

That list should distinguish things that actually behave differently:

• Named attorney referrals — the individual, not the firm
• Past client referrals — and which past client
• Professional referrals — the specific chiropractor, doctor, agent
• Paid channels — the campaign, not just "Google"
• Organic search, and the page they landed on
• Walk-in and phone, which is usually a proxy for local reputation

Then it has to survive. The source recorded at intake needs to still be on the matter at settlement, which sounds obvious and is exactly where most firms lose it — the enquiry has a source, the matter that gets opened from it doesn't.

Report on fee, not enquiries

Once source lives on the matter through to close, the report you want is one row per source with:

• Enquiries received
• Signed — and therefore the conversion rate
• Matters closed with a fee
• Gross fees earned
• Referral fees paid out
• Net fees
• Average time from intake to fee

That last column matters more than firms expect. A source producing good fees on an eighteen-month cycle is a different cash-flow proposition from one producing less on a six-month cycle, and if you're managing a firm's runway, faster money is worth something.

What to do with the answer

Feed the top of the list deliberately. If two chiropractors produce a third of your net fees, that relationship should get partner time, not whatever's left. It usually gets less than the loud sources.

Have the conversation with the volume sources. A source sending thirty unsuitable enquiries usually isn't doing it deliberately — they just don't know what you want. "We're the wrong firm for X but we're very interested in Y" often reshapes what arrives, and it's a kinder conversation than quietly declining thirty things a quarter.

Track the reciprocity. If you're sending more than you're getting, that's fine as a relationship investment as long as it's a decision. It usually isn't; it's just something nobody counted.

Kill paid channels honestly. Paid legal advertising is expensive and the enquiry-level metrics look fine on channels that produce nothing. Cost per signed matter, and cost per fee earned, are the only two numbers that matter, and they're only computable if source survives to settlement.

Where this breaks

Multiple touches. Someone hears about you from a friend, then searches your name, then clicks an ad. Attributing that to the ad is wrong, and it's the default in most analytics.

For a firm, first-touch is usually the more honest model — the friend caused this — with the search recorded as a secondary. Perfect attribution isn't available and chasing it wastes more time than it's worth. Consistent attribution is enough to rank sources correctly.

Declined matters. Worth tracking too. A source whose cases you decline 80% of the time is costing you screening hours, and that's a real cost even though it never appears in a fee report.

The honest caveat

This tells you where fees came from. It doesn't tell you why, and it's backward-looking. A source that produced nothing for two years and then sent one major case isn't necessarily a bad source — small numbers are noisy, and in a practice area with big variance you can draw a very confident wrong conclusion from eight matters.

Use it to notice the sources you're systematically under-investing in. That's where it's reliable, and that's usually the finding that pays for it.

---

We build intake, matter pipelines and referral attribution for US law firms. See how it fits together, or tell us how you track sources today.