CRM Cleanup and Workflow Standardization
Their Zoho CRM had duplicate contacts, deal stages that meant different things to different people, and free-text fields nobody trusted. Every pipeline report was guesswork, so the founder ran the business from memory.
Industry
B2B professional services firm with a small but growing sales team
Role
Systems consultant — CRM audit, data cleanup, workflow automation, documentation
Impact
Pipeline reports went from guesswork to numbers the founder uses in weekly reviews, and the CRM stayed clean months after handover because the hygiene rules and SOPs did the maintaining.
01 - Problem
The problem
The team asked for a sales dashboard. What they had was a data problem. Their Zoho CRM had run for years across several people, each with their own habits. The same company existed three or four times under slightly different names. Contacts came in from spreadsheets, event lists, and email exports with no matching rules, so duplicates were everywhere. Stages had piled up without anyone retiring the old ones — "Proposal Sent" and "Quote Out" both existed and nobody could tell you the difference.
The fields were the deeper problem. Industry, lead source, and deal type were all free-text, so "Referral", "referral", "ref from client", and "word of mouth" were the same thing four ways. Any report grouped by those fields was meaningless, so the team quietly stopped trusting the CRM — reps kept real notes in personal spreadsheets and only updated Zoho when chased. The founder was forecasting and hiring off a pipeline number everyone privately knew was wrong. A dashboard on dirty data just makes guesswork look official, so the engagement became a CRM cleanup first, with automation layered on only once the records could be trusted.
02 - System Flow
How the system moves
Full data audit
Zoho export into Google Sheets
Deduplicate records
Matching rules plus review queue
Standardize fields
Controlled picklists replace free text
Re-map pipeline
Stages match real sales process
Automate hygiene
Make scenarios validate new records
Weekly exception report
Flagged records to Google Sheets
SOPs and handover
Field guide and team training
03 - Build
What I built
Audit and deduplication pass
I exported every module — contacts, accounts, deals — from Zoho into Google Sheets and ran a structured audit: duplicate detection on email, phone, and normalized company name, plus a column-by-column check of which fields were actually usable. Exact duplicates merged with rules that kept the most complete value and preserved deal history. Anything ambiguous went into a review queue we worked through together, because merging the wrong two contacts is worse than leaving a duplicate.
Field standardization with controlled picklists
I mapped every free-text value in lead source, industry, and deal type to a short agreed list, then converted those fields to picklists in Zoho. The mapping lived in Google Sheets — old value, new value — so the team could challenge every decision before I wrote anything back. Values that genuinely didn't fit got an "Other" option with a required note, so the picklists stayed honest instead of becoming a dumping ground.
Pipeline re-mapped to the real sales process
I sat with the founder and reps, walked through how a deal actually moves from first conversation to signed agreement, and rebuilt the stages to match — each with a written definition and exit criteria. Overlapping legacy stages were retired and every open deal migrated through a mapping table, with the ambiguous ones reviewed one at a time. Most cleanups skip this, and it's why they decay: if the stages don't describe reality, people stop updating them.
Data hygiene automation in Make
Once the structure was solid, Make scenarios kept it that way. New and updated records run through validation: incoming contacts are checked against existing records before creation, required fields are enforced, and anything that fails lands on a weekly exception report in Google Sheets instead of rotting in the CRM. This is workflow automation doing the boring job it's best at — catching drift while it's still one record, not three hundred.
SOPs and handover
I wrote short, practical SOPs: what each field means, when a deal moves stages, how to handle a suspected duplicate, and who owns the weekly exception report. I trained the team and documented every Make scenario — what it watches, what it does, what to check if it errors — so they don't depend on me to keep their own internal tools running. A system the team can't maintain isn't finished.
04 - Impact
Business impact
Pipeline reporting stopped being guesswork — the founder runs weekly reviews off live Zoho reports instead of asking each rep what's really going on.
Reps abandoned their shadow spreadsheets because the CRM finally matched reality, leaving one source of truth instead of five.
What used to be a quarterly afternoon of manual list-scrubbing before any campaign or report is now a few minutes reviewing the weekly exception sheet.
Stage definitions with exit criteria ended the recurring "is this deal actually close?" debates — a deal in a stage now means one specific thing.
New records enter clean by default, so the cleanup didn't decay the way the team's previous tidy-ups had.
The standardized data became a foundation to build on — segmented outreach and an operations dashboard turned into realistic next steps.
05 - Edge Cases
What had to be handled
Near-duplicates that weren't exact matches — the same person with a work and personal email — were caught by normalized matching on name plus company and routed to human review rather than auto-merged.
Merging risked losing data, so merge rules kept the most complete value per field, concatenated notes, and re-pointed all deals and activities to the surviving record before anything was deleted.
Free-text values that didn't map cleanly went to "Other" with a required note, and the exception report surfaces "Other" usage so genuinely new categories get added deliberately.
Open deals in retired stages couldn't be bulk-moved, so each legacy stage had an explicit mapping and ambiguous deals were reviewed one at a time with the rep who owned them.
Future bulk imports were the likeliest way to re-pollute the CRM, so the SOP routes all imports through a Make scenario that dedupe-checks and validates rows before creation.
If a Make scenario fails or Zoho's API is unavailable, records queue rather than disappear — the scenario retries and unresolved failures hit the weekly exception sheet so nothing is silently dropped.
06 - Improvements
What I would improve next
Build the reporting dashboard the team originally asked for — now that the data is trustworthy, an operations dashboard on Zoho is a build, not a fantasy.
Push validation upstream with API integration from the website forms and inbox, so leads arrive structured instead of being cleaned after entry.
Add a light lead-scoring layer using the now-reliable source and industry fields to help reps prioritize.
Schedule a quarterly automated audit — a Make scenario that re-runs the duplicate and field-quality checks and emails a short health report, so drift gets caught even if habits slip.
Related capabilities