Should lead source live in a custom field, a tag, or the built-in source list?
Use a single-select field with a controlled list. Tags are many-to-many, so a customer can carry three of them and no report can add them up without double counting. Free-text fields drift into dozens of spellings within a quarter. The built-in source list is fine if it is short, enforced, exposed by the API, and set once at first contact rather than edited later.
The structural difference that decides it
A tag is a label with no cardinality rule. A record can have none, one, or six. That is exactly what you want for operational flags like needs-gate-code or dog-on-site. It is exactly what you do not want for source, because source is supposed to sum to one hundred percent of your leads.
A single-select custom field enforces one value from a fixed list. That constraint is the whole point. It makes the field groupable, sortable and summable without a cleanup step in front of every report.
Where free text goes wrong
Within one quarter a free-text source field on a busy account will contain Google, google, Google Ads, GoogleAds, GOOGLE, and Google (Sarah). Every one of those is a separate row in a group-by. The cleanup is a mapping table that someone has to maintain forever, and the mapping table becomes the real source of truth while the field becomes decoration.
If your platform only offers free text, put a normalization step in the integration layer rather than in a spreadsheet. Map on ingest, keep the raw value, and report on the mapped one. That way a new spelling shows up as an unmapped bucket you can see instead of silently splitting a channel in two.
Design the list before you build anything
- Keep it short. Eight to twelve values. If the list scrolls, people pick the first item.
- One dimension only. Do not mix channel with campaign with referrer name. Source is where the demand came from; put the detail somewhere else.
- No value named Other that anyone uses. If Other exceeds a small share, the list is wrong.
- Set at first contact, never edited. Original source is a historical fact. Changing it later rewrites your channel history.
- Confirm the API exposes it. Some platforms surface custom fields in the UI but not in exports or endpoints, which makes the field invisible to reporting.
Why the field is only half the answer
Even a well-designed source field is a human's best guess typed while a customer waits on the phone. It is useful, but it is not measurement. Digital sources should be captured mechanically — tracked numbers, form parameters, click identifiers — and reconciled against what the office staff selected. The gap between the two is the accuracy of your field, and it is worth knowing. See marketing intelligence for how mechanical capture and human-entered source get compared, and integration for where the normalization belongs.
Topics: lead source · custom fields · tags · data hygiene
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.