Lead source attribution

Lead source attribution is the practice of assigning each new lead to the marketing channel, campaign, or touchpoint that produced it. If 1,000 leads arrive in a month and the team can trace 400 to paid search, 350 to organic search, and 250 to referrals, then attribution has turned an undifferentiated count into a breakdown the business can act on. Each lead carries a source field, and the sum of those fields tells you where demand actually originates.

The metric matters because a lead total on its own hides every important decision. Two months can both produce 1,000 leads while one is healthy and one is fragile. If this month leans on a single paid channel that is getting more expensive, the headline looks fine while the foundation weakens. Attribution exposes that concentration. It connects spend to output and lets you compare channels on equal terms.

Attribution is the input layer beneath lead conversion rate and cost per lead. Without a reliable source on every lead, downstream channel economics are guesswork. With it, you can trace a closed deal back through the pipeline to the first campaign that surfaced the buyer.

At its simplest, attribution share for a channel is the leads credited to that channel divided by all attributed leads, expressed as a percentage. The harder part is deciding how credit is assigned when a lead touches several sources before converting. The inputs below define a clean, comparable attribution layer.

1. 1

   Source capture

   Every lead must carry a source from the moment it enters the system. UTM parameters, referrer data, and form fields feed this. A lead created without a source should fail validation, not default silently to direct.

2. 2

   Attribution model

   Decide whether credit goes to the first touch, the last touch, or is shared across touches. First touch rewards demand creation, last touch rewards demand capture. The model you pick changes which channels look strong.

3. 3

   Attribution window

   Set the period during which a touch can claim a lead. A 30 day window credits a campaign for a lead that converts within 30 days of the click. Too short and you miss slow buyers, too long and you over-credit old activity.

4. 4

   Deduplication

   Match repeat visitors and multiple form fills to a single person so one buyer is not counted as three leads across three channels. Without this, attribution shares add up to more than the real lead count.

Worked example. In a month you record 1,000 attributed leads after deduplication. Paid search is credited with 400, organic with 350, referral with 250. Paid search attribution share is 400 divided by 1,000, or 40 percent. If paid search spend was 20,000 pounds, your cost per attributed lead from that channel is 50 pounds. The same arithmetic across every channel gives you a like-for-like comparison no single lead total could provide.

A metric tree decomposes total attributed leads into the channels that produce them, then decomposes each channel into the levers that drive its output. This moves attribution from a pie chart you look at into a structure you can diagnose.

The first level splits leads by source family: paid, organic, referral, and outbound. Each family then breaks down further. Paid leads are a function of spend, click volume, and landing page conversion. Organic leads depend on ranking pages, traffic, and form conversion. Referral leads come from partner introductions and existing customer advocacy. When a channel moves, the tree tells you whether the cause was reach, conversion, or spend.

This is where lead source attribution becomes a decision tool rather than a report. KPI Tree connects each branch to the team that owns it, with RACI roles on every node so the accountable owner is named. When paid lead volume drops, the person responsible for that branch is pushed the change, not left to discover it in a monthly review.

There is no single correct channel mix, because the right blend depends on the market and the motion. What benchmarks well is coverage and concentration. A healthy attribution layer credits a source to almost every lead and does not depend entirely on one channel that could be cut off overnight.

Read these together rather than in isolation. A channel can show a high attribution share simply because tracking on other channels is weak, which inflates its slice of a shrinking attributed total. Coverage first, then mix, then the economics of each source.

Improving attribution is partly a data exercise and partly a modelling one. The data work raises coverage so more leads carry a true source. The modelling work makes the credit you assign reflect how buyers actually move through your channels.

The highest leverage improvement is usually coverage, not model sophistication. A perfect attribution model applied to leads where a third have no source still produces a misleading picture. Fix capture first, then refine how credit is shared. KPI Tree supports this by tracking the unattributed share as its own node with an accountable owner, then checking whether the fix actually moved coverage in the following period.

1. 1

   Treating direct as a real channel

   Direct and unknown are usually tracking failures, not a marketing source. Counting them as a healthy channel hides the work needed to recover the true origin of those leads.

2. 2

   Defaulting to last touch without thinking

   Last touch is the easiest model and it systematically under-credits the channels that create demand early. If your sales cycle spans weeks, last touch alone will mislead budget decisions.

3. 3

   Letting attribution shares exceed 100 percent

   When the same buyer is counted across several channels without deduplication, shares add up to more than the real lead count. Resolve identity before assigning credit.

4. 4

   Stopping attribution at the lead stage

   A source field that never reaches the closed deal cannot tell you which channels produce revenue. Carry the source through the full funnel so lead quality, not just lead volume, drives spend.

5. 5

   Comparing channels on volume alone

   The biggest source of leads is not always the best. Without joining attribution to conversion and revenue, you reward the channel that fills the top of the funnel cheaply while starving the one that closes business.