The Conversion Math Behind Lead Attribution

Signed cases are a multiplication problem, not an addition problem

Here is the equation that governs every dollar your firm spends on marketing:

meaningful opportunities × completion rate × qualified rate × close rate

Four terms. Multiplied, not added. That distinction is the whole tip.

When you add, a weak input costs you a little. When you multiply, a weak input costs you proportionally across everything upstream of it. A firm generating plenty of meaningful opportunities with a broken completion rate — calls that ring out, forms that never get a reply, after-hours inquiries that die in a voicemail box — is not "doing okay on marketing and needs to tighten up intake." It is destroying a fixed percentage of every ad dollar, every month, quietly.

Most firm owners look at the first term and the last term. They know roughly how many inquiries came in, and they know how many matters they signed. The two middle terms — completion and qualified rate — are where the leakage lives, and they are the two terms almost nobody measures directly.

Why most firms can't see the middle of the equation

This isn't a discipline problem. It's an instrumentation problem, and the data backs that up.

In CallRail's 2026 survey, 52% of firms used a CRM, but only 22% used call tracking or recording and 11% used virtual receptionists.

Read those three numbers as a system and the picture is obvious. Just over half of firms have somewhere to store a lead. Barely a fifth have the ability to observe what actually happened on the phone. And roughly one in nine has staffed coverage for the calls that arrive when nobody's at a desk.

So the CRM fills up with records that say a lead exists. It does not say whether the call connected on the first ring or the fourth, whether the caller had a matter you actually take, whether the person who answered asked the qualifying questions, or which campaign produced that call in the first place. You end up with a database of outcomes and no record of process. That is enough to compute the first and last terms of the equation and nothing in between.

A CRM tells you a lead existed. Call tracking tells you what happened to it. Most firms have the first and not the second, then wonder why their numbers don't explain anything.

Improve one stage at a time

The multiplication structure has a useful consequence: you don't need to fix everything. A modest lift in a single term moves the whole product. And because the terms multiply, the cheapest available improvement is usually not "more leads."

Work the equation in order and ask a specific diagnostic question at each stage:

• Meaningful opportunities. Not raw traffic, not form fills. How many inquiries came from people with a matter in a practice area you serve, in a geography you serve? Volume that fails this test inflates your top-line number and depresses every rate below it.
• Completion rate. Of those opportunities, how many actually reached a human — or got a response fast enough that the person was still paying attention? This is where the 22% call tracking number bites. If you don't record calls, you are guessing.
• Qualified rate. Of the ones you reached, how many met your actual criteria? Venue, statute of limitations, damages, conflict, fee structure. If this rate is low, the problem is usually upstream in targeting, not downstream in sales skill.
• Close rate. Of the qualified ones, how many signed? If this is your weak term, it's a follow-up cadence and consultation problem — and it's the only term where "try harder" is a legitimate answer.

Pick the weakest term. Change one thing. Measure. Then move to the next. Changing three things at once across a multiplicative system guarantees you'll never know which one worked.

The first move: store first-touch source separately from most recent touch

Here's the concrete action, and it's smaller than you think.

In your CRM, create two distinct fields: first-touch source and most-recent-touch source. Never let one overwrite the other.

Most intake systems keep a single "source" field, and whatever arrives last wins. A prospect finds you through an organic search for a practice-area-plus-city page, reads for a week, then clicks a branded paid ad and calls. Single-field attribution credits paid search and erases the page that actually created the opportunity. Do that across a few hundred matters and you will systematically defund the channel doing the work while scaling the one taking credit.

Two fields fix this immediately, and they're what make real multi-touch and data-driven analysis possible. First-touch tells you what generates demand. Most-recent-touch tells you what closes it. Those are different jobs, and they deserve different budgets.

Then track cost per signed matter before and after

Once the fields exist, establish a baseline. Compute cost per signed matter — total marketing spend divided by signed matters — for a clean period before you make any change. Then make one change, let it run, and compute it again.

This is the step that separates operators from people with opinions. Without a before-and-after number tied to a specific change, every improvement is anecdotal. You'll have a story about how things feel better since you switched vendors. Stories don't survive a bad quarter. Attributable numbers do.

Cost per signed matter is the right metric because it absorbs all four terms at once. Improve any of them and the number falls. It also prevents the classic trap of celebrating a cost-per-lead reduction that came from buying cheaper, worse leads — those show up as a higher qualified-rate failure and a flat or worse cost per signed matter.

What this unlocks for coverage strategy

There's a second-order benefit worth naming. Once first-touch source is stored separately and you can compute cost per signed matter by source, you can finally answer the geographic question: which cities and practice areas actually produce signed matters, and which ones produce noise?

Firms building out serious page inventory — service-and-city coverage across every practice area in every market they serve — generate a lot of entry points. Without split attribution, all of that collapses into "organic" and you learn nothing about which markets deserve more investment. With it, you get a ranked list of cities by cost per signed matter, which is the single most useful document in a growing firm's marketing folder.

That's the sequence: instrument the middle of the equation, split the source fields, baseline cost per signed matter, improve one stage, re-measure. Not glamorous. Extremely hard to argue with.

See how BOSSEO approaches Lead Attribution: bosseo.com/attribution.