Lead Source
A lead source is the specific channel, campaign, or referral through which a prospect first discovered your business and entered your funnel.
Key takeaways
- Keep a fixed channel vocabulary; free-text source fields degrade within a year.
- Write the source once at first identified visit and never overwrite it.
- Source is only actionable once joined to channel cost and downstream outcome.
- Store channel, medium and campaign separately so reports can roll up cleanly.
- Links shared privately arrive as direct and understate word-of-mouth channels.
In depth
A lead source is a stored field, and its usefulness depends entirely on the vocabulary behind it. Most setups keep three levels: a broad channel such as paid search or referral, a medium describing the format, and a campaign identifier naming the specific effort. The values are written once, at the first identified visit, from URL parameters and the referring domain. Because it is a single field on a single record, it can answer where this contact came from but never how much each touch contributed.
Granularity is the main design decision. Too coarse and every paid campaign collapses into one bucket you cannot optimise; too fine and the report has four hundred rows, each with three leads and no statistical weight. The workable compromise is a fixed list of channels for reporting, with campaign detail stored underneath for drilling. A free-text source field always degrades: within a year the same channel exists under five spellings and every report needs manual cleaning.
Source earns its keep when it is joined to two other numbers: what the channel cost and what its leads eventually became. That join turns a list of origins into a cost per qualified lead by channel, which is the figure budget decisions actually need. A scorecard funnel helps here because the source recorded at quiz entry travels with the scored result, so cost per lead and cost per high-scoring lead can both be computed from the same records.
A single source field compresses a journey into one word, and for anything with a long consideration period that word is usually wrong in an interesting way. Referral traffic often carries no identifying header at all. Dark social, meaning a link pasted into a private chat, arrives as direct. And the field records where someone entered, not what convinced them, so a channel that reliably produces first visits may be doing very little of the persuading.
Example in practice
How to measure it
Build one table with a row per source and four columns: leads created, qualified leads, spend, and qualified conversion rate. Cost per qualified lead is spend divided by qualified leads, and it is the column that should drive the budget conversation. Sort by it rather than by volume, and hide rows below a minimum lead count so tiny sources do not top the ranking by accident.
Separately, keep a hygiene metric: the percentage of new leads with a recognised source value. Watch it weekly, because it degrades silently when a landing page is republished without tags or a new campaign launches untagged. A sudden rise in unknown sources is nearly always a broken link somewhere, and finding it the same week is far cheaper than reconstructing a quarter later.
Common mistakes
The recurring problem is letting each team invent its own values. Sales types the source by hand, the ads team uses lowercase, and an integration writes a different label for the same channel. Reports then need cleaning before every review. Publish a short controlled list, validate the field on write so unknown values are rejected, and give one person the authority to approve new entries.
The second is judging channels on lead count alone. A source producing many cheap leads outranks one producing few expensive ones, right up until someone looks at what closed. Rank by cost per qualified lead, and where the sales cycle allows, by cost per won deal. Until the outcome column exists, a source report tells you where volume came from and nothing about whether it was worth buying.
Frequently asked questions
What is the difference between lead source and lead attribution?
Lead source is the raw origin of a single lead, such as an organic search or a referral. Lead attribution is the model that decides how credit is shared when a lead touched several sources.
How is lead source captured automatically?
It is typically read from UTM parameters and the referrer header when the visitor first lands. A quiz funnel can persist these values with each result so the source is stored on the lead without manual entry.
Why track lead source if I already track total leads?
Total leads tell you how many you got, but lead source tells you where your best leads come from. That breakdown is what lets you shift budget toward the channels with the lowest cost per qualified lead.
What is the difference between lead source and channel?
Channel is the broad category, such as paid search, organic search, email or referral. Source is usually the more specific origin within it, such as the particular site or campaign. Many tools use the words loosely, so define both in your own documentation and stick to the definitions, otherwise two reports will disagree while both being correct.
Should lead source ever be overwritten?
The original source field, no. Keep it as a write-once record of how the contact first arrived. If you also want the most recent origin, add a separate latest-source field that updates freely. Overwriting the first value destroys the only record of what opened the relationship, and that loss cannot be reconstructed afterwards.
How do I track lead source for offline channels?
Give each offline effort its own doorway. A dedicated short URL, a QR code pointing at a unique landing page, a distinct phone number or a code the visitor enters all produce a trackable entry. Where none of that is possible, a short self-reported question on the form is imperfect but better than filing every event lead as direct.
How many lead sources should my taxonomy have?
Few enough that a person can hold the list in their head, which in practice means somewhere around ten to fifteen top-level values. Everything more specific lives in the campaign field beneath. If a value receives fewer leads than you would act on, merge it upward; a category you never make a decision about does not need its own row.
Why do so many of my leads show organic search as the source?
Often because people discover you elsewhere and then search for your name to return, so branded search absorbs credit that another channel earned. Split branded from non-branded search queries in your reporting. The non-branded portion is genuine discovery; the branded portion is mostly a return path for demand created somewhere your source field cannot see.
Can two leads from the same company have different sources?
Yes, and that is normal in committee purchases. Keep the source on the contact record where it belongs and add a separate account-level field for the source of the first contact from that company. Reporting on the account field avoids counting one deal several times, while the contact field still shows which channels reach which roles.