Lead source performance

Lead source performance measures how well each channel produces leads that convert and generate revenue, not just how many leads it produces. A channel that delivers 500 cheap leads which never close performs worse than one that delivers 100 leads where a quarter become customers. Performance judges a source on the full journey from lead to revenue, then sets that against what the channel costs to run.

The metric matters because lead volume is a vanity number on its own. Marketing teams that optimise for raw lead count reward whichever channel fills the top of the funnel most cheaply, which is rarely the channel that produces the best customers. Performance corrects this by joining three things every volume report leaves out: the conversion rate of each source, the value of the deals it closes, and the cost of running it. Only with all three can two channels be compared fairly.

Lead source performance sits one level above lead source attribution. Attribution tells you where leads came from. Performance tells you which of those origins deserve more budget. It draws directly on lead conversion rate and average deal size to turn a list of sources into a ranking the business can fund against.

Performance combines volume, quality, and cost into a single comparable figure per channel. The expression above multiplies a channel lead count by its conversion rate and average deal value to get revenue produced, then divides by source cost to get value returned per pound. The inputs below define a fair calculation.

1. 1

   Leads per source

   Take the deduplicated, attributed lead count for each channel. This is the starting volume, and it must come from a reliable source field so channels are compared on the same basis.

2. 2

   Conversion rate per source

   Measure the share of each channel leads that become customers, not the blended rate across all channels. Conversion is where quality differences between sources show up most sharply.

3. 3

   Average deal value per source

   Use the revenue from deals that closed from each channel. Some sources produce small deals at high volume while others produce a few large ones, and a blended figure hides that.

4. 4

   Source cost

   Include the full cost of running the channel: ad spend, content production, tooling, and the people time it consumes. Leaving out cost makes free-looking channels appear better than they are.

Worked example. Paid search produces 400 leads at a 5 percent conversion rate and a 4,000 pounds average deal, for 80,000 pounds of revenue, against 20,000 pounds of spend. Its performance is 80,000 divided by 20,000, or 4. Referral produces 100 leads at a 20 percent conversion rate and a 6,000 pounds average deal, for 120,000 pounds, against 5,000 pounds of cost. Its performance is 24. Referral produces a quarter of the leads and six times the value per pound. Volume alone would have ranked them the wrong way round.

A metric tree decomposes lead source performance into the factors that make a channel valuable, then connects each factor to the team that influences it. The root is blended performance across all channels. The first level splits into the four drivers: volume, conversion, deal value, and cost. Each then breaks into the levers beneath it.

This structure shows why a channel performs the way it does, not just that it ranks high or low. If a source underperforms, the tree tells you whether the cause is thin volume, weak conversion, small deals, or high cost. Each of those points to a different fix owned by a different team. Marketing can lift volume, sales can lift conversion, pricing affects deal value, and operations affect cost.

KPI Tree assigns RACI ownership to each node so the accountable owner is named, not implied. When a channel conversion rate drops, the tree pushes the change to the person responsible for that branch. The verified impact loop then confirms whether the intervention they made actually improved the channel performance figure, rather than assuming the effort worked.

Performance benchmarks vary by motion, so the useful comparison is relative rather than absolute. Rank your own channels against each other and watch the spread. A healthy portfolio has a few high-performing sources you can scale and a clear view of which low performers to fix or cut.

Read conversion alongside cost rather than on its own. A 2 percent conversion channel with very low cost per lead can outperform an 8 percent channel that is expensive to run. The performance figure is what reconciles the two, which is why ranking on a single sub-metric is misleading.

Improving performance means lifting one of the four drivers for a channel, or reallocating budget away from sources where no driver can be moved. The most efficient gains usually come from conversion and fit rather than from chasing more volume.

The discipline that makes these gains stick is checking that a change actually moved the number. KPI Tree models each channel as a branch with its four drivers and an accountable owner, then runs a verified impact loop after every intervention, so a follow-up improvement or a budget shift is measured against the performance figure it was meant to lift rather than assumed to have worked.

1. 1

   Ranking channels on lead volume

   The largest source of leads is rarely the best source of revenue. Judging performance by volume rewards channels that fill the funnel cheaply while ignoring whether those leads ever close.

2. 2

   Using a blended conversion rate

   Applying one conversion rate across every channel hides the quality difference between sources. Conversion must be measured per channel or the performance figure is meaningless.

3. 3

   Leaving cost out of the comparison

   A channel that looks free because it has no media spend still consumes content, tooling, and sales time. Excluding those costs flatters the wrong sources and distorts the ranking.

4. 4

   Stopping at the lead, not the deal

   Measuring performance at lead creation misses what happens after. A source that produces many leads that stall in the pipeline performs worse than the lead count suggests.

5. 5

   Judging a slow-building channel too early

   Organic and referral compound over time. Cutting them on an early performance read, before content or advocacy has built, removes the cheapest long-term sources from the portfolio.