Reporting and Attribution for Contractors

Why "Number of Leads" Is a Misleading Number

A channel that produces 50 cheap, poorly-qualified leads that convert at 2% can lose money next to a channel that produces 10 expensive, well-qualified leads that convert at 40%. Lead count alone hides which channel is actually worth the spend; only tracking from lead source through to a closed job answers that.

The Minimum Fields That Matter

• Original lead source -- the very first channel that brought the contact in, captured once and never overwritten.
• Latest lead source -- the most recent touch, useful for understanding what re-engaged a returning contact.
• Original campaign -- which specific ad, page, or referral, not just the broad channel.
• Won/lost outcome -- tying the source all the way through to whether the job actually closed, not just whether a lead arrived.

Without the first field specifically -- captured at first touch and never overwritten by a later interaction -- attribution collapses to whatever channel touched the contact last, which usually over-credits retargeting and under-credits the channel that actually created the interest.

A Worked Example

Say a contractor runs both a Google Ads campaign and a referral program. If attribution is only captured on the most recent touch, a customer who first heard about the business from a referral but later clicked a retargeting ad gets recorded as a "Google Ads" lead -- silently making the ad spend look more effective than it is, and the referral program look worthless, when the opposite may be true. This is illustrative, not a measured case study.

What to Check First

• Do you know, for your last 20 closed jobs, which channel actually brought each one in originally?
• Is that source captured at first contact, or reconstructed from memory afterward?
• Can you see win rate broken out by source, not just lead count by source?

Limitations

Attribution is inherently imperfect for offline businesses -- word-of-mouth referrals in particular are hard to capture with certainty. The goal is directionally honest data, not perfect certainty.