Sample engagement · Salesforce CRM 3-week engagement

A 12-year-old Salesforce org, made honest in three weeks.

Three generations of admins. Forecast accuracy in the low 60s. The same account appearing three times under three different spellings. We rebuilt the object model, deduplicated 4,200 contacts, and shipped a one-page operations playbook the team actually reads.

Executive summary

Audited 144 custom fields, merged 4,200 duplicate contacts across 1,100 accounts, retired 42% of the field bloat, and left behind validation rules so the same drift doesn’t come back. Month-end forecast went from 12 days to 2.

Engagement
3 weeks · fixed scope
Stack
Salesforce · DemandTools · LeanData
Industry
B2B SaaS
At a glance

Four numbers that mattered.

Duplicate contacts
4,200merged
↓ 99%
Across 1,100 accounts
Custom fields
144→ 63
↓ 56%
Required defaults set
Forecast accuracy
63→ 92%
+29 pts
Month-over-month
Time to forecast
12→ 2 days
−83%
Month-end close
01 · The work

Reports nobody trusted, deals that fell through the cracks.

Twelve years of admin turnover left an org that looked busy from the outside — 144 custom fields, 38 picklists, a dozen automation rules — and broken on the inside. Leadership couldn’t answer pipeline questions without three people in a room. We treated the engagement as an audit first, an implementation second.

Problem

An org that nobody owned.

The same account appeared three times under different spellings. 12 of 38 custom fields on Contact had < 5% fill rates. Conversion math depended on which view you opened.

  • 4,200 duplicate contacts across 1,100 accounts
  • 38 custom fields on Contact alone, 12 with < 5% fill
  • Forecast variance ±37% month-over-month
  • No validation rules on lead source, region, or stage
Solution

Audit → merge → guardrail.

Three passes in three weeks. Inventory the drift, fix it once, then prevent it. The engagement was time-boxed with a single handoff document, not an ongoing retainer.

  • Fuzzy-match dedupe in DemandTools, manually reviewed top 200
  • Retired 42% of custom fields, set required defaults on the rest
  • Validation rules + duplicate management policy
  • One-page playbook for the new Salesforce admin
02 · Tools

The stack used on this job.

Boring, proven, and already in the client’s budget. Cleanup engagements aren’t the place to introduce new vendors.

SalesforceCRM DemandToolsDedupe LeanDataRouting Field TripField audit SlackAsync standups NotionHandoff doc LoomAdmin training
03 · The proof

Custom-field inventory, before and after.

Every custom field across the five core objects, with average fill rate, the decision we made, and the resulting state. Retired fields are crossed out — the data was moved or judged irrelevant after review with sales leadership.

Custom field audit · all objects

5 objects· 144 fields audited· 81 retired
Object Custom fields Avg fill rate After cleanup Validation rules Status
Lead 22 31% 9 4 Stable
Contact 38 18% 14 6 Stable
Account 27 52% 19 3 Review Q3
Opportunity 41 44% 21 8 Rebuilt
Custom: Partner Tag 16 3% 0 0 Retired
04 · Outcome

A pipeline leadership can read on a Monday.

The handoff doc is the part clients say they read most. One page, written for a future admin who has never met us.

Before

An org that drifted

  • Forecast accuracy ~ 63% month-over-month
  • 11 lead-source values, 4 of them misspellings
  • Sales ops spent 8 hrs / week on manual cleanup
  • No alert when a deal stalled in stage
  • 4,200 duplicate contacts across the database
After

An org that holds up

  • Forecast accuracy ~ 92% month-over-month
  • 5 lead-source values, picklist-enforced
  • Sales ops cleanup down to 90 min / week
  • Slack alert at 14 days stalled in stage
  • Duplicate prevention rule blocks new dupes at entry

Illustrative outcome for this engagement type.

05 · What’s next

Handed over with a 90-day roadmap.

The cleanup is the foundation. These are the next moves the team can run without us — or that we’d return for on a fixed scope.

Done · weeks 1–3Complete
Deduplicate Contacts & Accounts
SalesforceP0
Retire low-fill custom fields
SalesforceP0
Enforce required lead-source picklist
SalesforceP1
Later · Q32 items
AI-assisted forecast commentary
Claude · SalesforceP2
Quarterly tech-stack review
StrategyP3