Custom CRM Fields That Actually Get Used

Custom fields that actually get used: a CRM design guide

Open any mature CRM and you'll find the same archaeological layers: a "Decision Criteria" field three managers added in 2022, half-populated; a "Champion Strength" picklist with options nobody can define; a "Next Step Date" duplicated three times because each RevOps lead wanted their own. Custom fields are the cholesterol of a CRM. They build up quietly, they're hard to remove without a fight, and they slow everything down.

The fix isn't fewer fields. It's stricter rules for what earns the right to exist.

The test a custom field has to pass

Before a field gets added, it should clear four hurdles. Skip any one and you're building decoration, not infrastructure.

It changes a decision. If nobody can name the report, dashboard, routing rule, or coaching conversation that will use the data, the field is a museum exhibit. "Competitor mentioned" only matters if someone is reviewing competitor-tagged losses monthly and adjusting battlecards. Otherwise, delete it.

The answer is knowable at the moment the rep fills it in. Asking an SDR to mark "Budget confirmed" after a 12-minute discovery call is fantasy. Asking them to mark "Budget question asked: yes/no" is a behavior you can actually inspect.

There's one obvious right answer. If two reps would reasonably pick different options for the same deal, the field is noise. "Deal complexity: low/medium/high" fails this. "Number of stakeholders met: integer" passes.

It's cheaper to capture than to reconstruct. A field saves time when it prevents someone (a manager, a marketer, a CS lead) from re-asking the rep later. If the data lives in Gong already, don't duplicate it in a picklist.

A field that fails any of these gets cut. A field that passes all four tends to get used without enforcement, because reps can feel the loop close.

The seven fields that consistently earn their keep

Across well-run B2B orgs, a handful of custom fields show up again and again because they tie directly to a forecast call, a routing decision, or a coaching motion. The list is short on purpose.

1. Source of first conversation (picklist: inbound demo request, outbound cold, event, referral, partner, expansion). Drives attribution, payout splits, and channel ROI. Picklist, not free text.

2. Champion identified (yes/no, plus a lookup to the contact). Forces the rep to name a human, not a vibe. Pairs with a report that flags any stage-3+ deal without a champion record.

3. Economic buyer met (yes/no, with a date stamp). The single best leading indicator most teams have on whether a late-stage deal is real.

4. Compelling event (short text, max 120 characters). If the rep can't articulate why the prospect would buy this quarter in one sentence, the close date is decorative.

5. Next committed customer action (short text plus date). Note the wording: not "next step" (which reps fill in with "follow up next week"), but the action the buyer has agreed to take. "Procurement to send redlines by July 8."

6. Primary use case (picklist tied to your top three or four ICP plays). Lets marketing, product, and CS see which use cases are converting and which are stuck. Keep the list short; if it has 14 options, reps will pick whichever is least embarrassing.

7. Loss reason — primary (picklist, mutually exclusive, paired with a short free-text "what we'd have needed to win"). The free text is where the real intel sits. The picklist is what you slice in dashboards.

Notice what's not on this list: industry vertical (use firmographic enrichment), employee count (same), MEDDIC scores broken into seven fields (a single qualification score works harder than seven half-filled subfields).

Why most custom fields die

Fields decay in predictable ways, and the failure mode usually points to the design flaw.

Free text where a picklist belonged. "Competitor" as free text produces "Salesforce," "SFDC," "salesforce.com," and "SF" in the same column. You can't report on it. Anything you'll want to count belongs in a picklist or lookup.

Picklist where free text was needed. The opposite mistake. Loss reasons especially. A picklist of eight loss reasons gives you neat pie charts and zero insight. Pair the picklist with a one-line free-text prompt and you get both.

Required fields at the wrong stage. Making "Economic buyer name" required to create an opportunity guarantees garbage data, because the rep doesn't know it yet. Make it required to advance to stage 4, when they should know it.

No feedback loop. If the rep fills in "Compelling event" and never hears it referenced in a forecast call, deal review, or pipeline scrub, they learn it's optional. Fields stay alive when managers visibly use them. The moment a manager runs a pipeline review off the field, completion rates climb without a memo.

Duplication via mergers. Every acquisition, every new VP, every consultant rollout adds fields. Nothing removes them. A quarterly field audit, with a default-delete posture for anything under 40% completion that isn't tied to a live report, is the cheapest hygiene a RevOps team can run.

A worked example: pruning a bloated pipeline object

Say your opportunity object has 84 custom fields. Run this exercise across an afternoon.

Pull a usage report: completion rate per field over the last two quarters, plus a list of every dashboard, report, workflow, and routing rule that references each field. Anything under 30% completion and unreferenced goes on the kill list immediately. That alone typically clears 20 to 30 fields.

For the remainder, ask the four-question test above. Fields that fail get archived (hidden from layouts, data preserved) for one quarter. If nobody complains, they're deleted. If a manager surfaces a real use case, the field comes back with an owner and a report attached.

The page layout that survives is usually under 20 custom fields, organized by stage. Reps fill in early-stage fields at creation, mid-stage fields when they advance, late-stage fields before they forecast commit. The form gets shorter as the deal progresses in the rep's mind but longer in the CRM, which is the right direction.

The takeaway

• Run the four-question test (changes a decision, knowable now, one right answer, cheaper than reconstructing) on every existing custom field this quarter. Archive the ones that fail.
• Rewrite "Next step" as "Next committed customer action" with a date. The change in wording alone tightens forecast conversations within a few weeks.
• Pair every loss-reason picklist with a short free-text field asking what would have been needed to win. The picklist feeds dashboards; the text feeds strategy.
• Make field requirements stage-gated, not creation-gated. Required fields at opportunity creation produce fiction; required fields at stage advancement produce data.