CRM data governance: 5 roles and the monthly routine

CRM data governance is the written and enforced agreement on who decides what about each piece of sales data, from the meaning of a field to imports, access and deletion. It works when leadership names an owner for each data domain, publishes a few short policies and keeps a committee that decides once a month with the indicators on the table. The risk of getting the design wrong has a number: in a February 2024 press release, Gartner predicted that by 2027 80% of data and analytics governance initiatives will fail due to a lack of a real or manufactured crisis, and analyst Saul Judah summed up the cause by saying that a governance program that does not enable prioritized business outcomes fails (Gartner, 2024). In sales, the crisis already exists, and it shows up in the forecast nobody defends, in the disputed commission and in the report each team rebuilds in a spreadsheet.

Database hygiene, with the quality dimensions, the fields per stage and deduplication, is covered in CRM data quality. The subject here is authority: without someone with the power to decide, the data entry rule becomes a suggestion by the third week.

What is data governance, and what is it not?

The market reference for the subject is the DAMA-DMBOK, the data management guide published by DAMA International, which also has a Brazilian Portuguese edition. DAMA itself states that the guide prescribes no tool and no single model, and that each organization has to adapt the guidance to its own maturity and problems (DAMA International, 2017). I use the DMBOK as a common vocabulary with IT and translate it for the revenue operation into one question: for each piece of data leadership uses to decide, who has the final word and where is it written down. Data management executes, configuring the field and running the merge, and governance decides which field exists, which rule applies and who answers when the rule breaks.

Two misconceptions delay the start. The first treats governance as a synonym for the LGPD, Brazil’s data protection law. The law requires accuracy, purpose limitation and security, and Article 50 of Law 13,709/2018 provides that the controller may implement a privacy governance program adapted to the structure, scale and volume of its operations. That program covers personal data, and sales governance also covers deal value, funnel stage, segment and product, which identify no one and decide the year’s budget. The second misconception hands the subject to IT. IT answers for the platform, security and integrations, and the rule that says whether a deal can enter the proposal stage without a value belongs to whoever answers for the number.

Why does the sales area need data governance?

The lack of governance shows up as an argument about numbers. In a hypothetical example, the sales rep records in the CRM the total value of a 24-month contract signed in January, R$ 240 thousand, and finance reads the same deal as R$ 120 thousand of revenue for the year, because it counts the 12 monthly installments of R$ 10 thousand that fall within the fiscal year. Both readings are correct, and the forecast meeting loses half an hour because nobody wrote down which value the field holds or who decides that.

The same happens with the “Segment” field, created by marketing for campaigns, and the “Vertical” field, created by sales for account portfolios, which live in the same database with different picklists. The signing date one rep records and another corrects turns into a commission dispute at month-end close. The AI agent that prioritizes accounts learns from these contradictions, and leadership finds the problem when the recommendation is already in use.

Who does what in CRM data governance?

Governance for a mid-sized sales operation fits into five roles. In smaller companies one person holds two of them, and what matters is that each role has a written name and the authority the role requires.

Executive sponsor

This is the executive who answers for revenue, usually the sales director, the head of operations or the CEO. The sponsor names the owners, settles conflicts between teams and backs the committee’s decisions when a senior rep complains, and without a sponsor the first portfolio dispute brings the program down.

Data owner by domain

A data domain is a set of information with the same business owner, such as accounts and contacts, leads, deals and pipeline, customers and contracts, products and prices. The owner is the leader of the team that depends most on that data and defines what each field in the domain means, who can change it and what level of quality is acceptable.

Domain	Data examples	Typical owner	Owner’s decisions
Accounts and contacts	Tax ID (CNPJ), segment, size, job title, account owner	Sales leadership	Portfolio rule, company key, who can change the account owner
Leads	Source, campaign, score, consent	Marketing leadership	Qualified lead criteria, list import rule
Deals and pipeline	Stage, value, close date, loss reason	Sales leadership	Entry criteria for each stage, what the value field holds
Customers and contracts	Start date, renewal, account health	CS leadership, with finance	Definition of active customer, sales to CS handoff fields
Products and prices	Catalog, list price, maximum discount	Sales leadership, with finance	Active items, discount approval levels

Data steward

The data steward is the person who keeps the owners’ decisions working day to day. The steward maintains the data dictionary, receives requests for new fields, reviews imports, tracks the indicators and brings to the committee what needs a decision. The role usually comes from revenue operations and, in a company with a single RevOps analyst, also covers CRM administration.

CRM administrator

The administrator executes the configuration: creates the approved field, adjusts validation, builds the access profile, configures the duplicate rule and handles integrations with IT. The content of the rules comes from the domain owners, and this separation protects the administrator: the rep who asks for a new field gets the committee’s answer.

Users

Sales reps, SDRs, marketing analysts and account managers record the data at the source, follow the standard for their stage and tell the steward when a rule gets in the way of the operation. The committee reads these requests every month, because a rule designed without the user ends up bypassed.

The RACI matrix for sales governance

RACI is the matrix that defines, for each decision, who executes (R, responsible), who approves and answers for the result (A, accountable, with only one person per row), who is consulted beforehand (C) and who is informed afterwards (I). The table covers the decisions that cause the most conflict in the CRM.

Decision	Sponsor	Domain owner	Data steward	CRM administrator	Users
Create or change a field	I	A	R	R	C
Change funnel stages or criteria	A	R	C	R	C
Add a new picklist value	I	A	R	R	C
Approve a data import	I	A	R	R	I
Merge duplicates with different owners	I	A	R	C	C
Grant or change an access profile	I	A	C	R	I
Set retention and deletion periods	A	R	R	C	I
Connect a new integration or AI agent	A	C	R	R	I
Change the definition of an indicator	A	R	C	I	I

In the retention row, legal and the data protection officer are mandatory consultees, and in the integration row so is IT. Keep a single A per row, because two people approving means nobody approves.

What minimum data policies does the CRM need?

Five policies cover most of the conflicts in a sales operation, and each fits on one page, because reps follow a policy they can read in three minutes.

Field creation and changes

Every new field goes through a request with name, definition, type, picklist values when there are any, who fills it in, at which stage and which report or automation will use it. The steward checks whether a field with the same function already exists, the domain owner approves and the administrator creates it, along with the matching entry in the dictionary. A field with no report or automation using it does not get created, and a field unused for two quarters goes to review.

Data entry standard

The policy says who approves the required fields per stage, the picklist values and the validated formats. The technical content of these rules is in the data quality article; governance adds the domain owner’s signature on each rule and a channel to request changes, which keeps a manager from creating a parallel list.

Duplicates

The policy records the key for each object, which record survives the merge and who decides when the two records have different owners. That last point affects portfolios and commissions, so the decision rests with the owner of the accounts domain, with a deadline to respond.

Retention and deletion

The policy defines, for each type of record, the retention period, the event that triggers a review and the record’s destination: delete, anonymize or keep with a recorded legal basis. The sponsor sets the periods with legal and the data protection officer, because they depend on contracts, legal obligations and the purpose of each piece of data. I write as an operations consultant, with no claim to legal advice; the use of this data in AI is covered in AI and the LGPD.

Access

The policy starts from least privilege: each profile sees and changes what its function requires. It defines who can export the database, who can delete records, who sees contract values and how access is removed when someone leaves the company. The design of profiles by role is in CRM permissions.

How do the data committee and the monthly routine work?

The data committee is the meeting where the domain owners decide what was left pending during the month. It brings together the sponsor, the domain owners from sales, marketing and CS, the steward, the CRM administrator and, when the agenda calls for it, someone from IT, finance or legal. One hour a month is enough when the steward sends the agenda two days ahead and each request arrives with the steward’s recommendation.

Block	Time	Led by	What comes out of the meeting
Database health indicators	15 min	Data steward	Domains below target and one person responsible for each gap
Field, picklist and stage requests	15 min	Steward, with the domain owners	Each request approved, rejected or returned with a deadline
Imports, integrations and agents	10 min	CRM administrator	What was loaded, the errors and what needs approval
Ownership and access conflicts	10 min	Sponsor	Recorded decision on portfolios, duplicates and profiles
Retention, deletion and data subject requests	5 min	Steward, with the data protection officer	The semester’s list and status of requests
Decision log and next steps	5 min	Steward	Decisions in the log, with owner and date

Which indicators show database health?

The committee needs a few indicators, measured the same way every month. The ones in the table measure whether governance works, and the quality dimensions, detailed in the data quality article, come in only through the critical fields: those that feed the sales forecast, commissions, campaign segmentation or handoffs, which at the start number ten to fifteen. The targets are a reference, and the committee adjusts them after the first three measurements.

Indicator	How to calculate	Starting reference	Owner
Critical fields with owner and definition	Critical fields with a complete dictionary entry, over total critical fields	100%	Data steward
Unused fields	Fields with no data and no use in reports in the last 180 days	Falling every quarter	CRM administrator
Completeness of critical fields	Records with the stage’s critical fields filled in	Above 90%	Each domain owner
New duplicates	Duplicate companies and contacts created in the month	Falling month over month	Data steward
Imports within the rule	Imports approved before loading, over total	100%	Leads domain owner
Reviewed profiles	Users whose access profile was reviewed in the quarter	100%	CRM administrator, with IT
Requests answered on time	Field and data subject requests answered within the agreed time	Above 90%	Data steward
Decisions executed	Logged decisions completed by the planned date	Above 80%	Sponsor

How do you start data governance small?

A program that starts by mapping all of the company’s data spends two quarters before the first decision and loses its sponsor along the way. Start with the domain that hurts most, almost always deals and pipeline, because it supports the forecast and commissions, and run the first committee with that scope only.

In the projects I follow, governance earns leadership’s trust when its first decision settles an old dispute, such as what the value field holds or who keeps the duplicated account. Once the deals domain is stable, the committee adds accounts and contacts the next quarter and leads after that, each with an owner, critical fields, definitions and indicators.

Common mistakes in sales data governance

Starting with the document

The company approves a long policy and nothing changes in the CRM, and the team keeps recording as before. The policy has to come from decisions about real fields and become configuration the same week, through validation, closed picklists or access profiles.

Handing governance to IT

IT takes on the subject because nobody else would and starts deciding the meaning of business fields without knowing the operation, and the sales team works around the rules. IT takes part as the owner of the platform and security, with the domain owners deciding the content.

Treating governance as an LGPD project

The compliance project deals with personal data, ends with a report and leaves out deal value, stage and product. Use the same committee for both fronts, with the data protection officer consulted on decisions about personal data, and keep the routine after the compliance project closes.

Naming an owner without authority

The operations analyst gets the title of data owner and cannot refuse an import requested by a director. A domain owner has to be a leader with decision power over the area, and the analyst stays as steward, with the sponsor telling the team that the steward’s decisions stand.

A committee with no record

The meeting happens, people agree and nobody writes anything down. The next month the same subject comes back, and the administrator configures what they remember from the conversation. The decision log costs five minutes per meeting and ends that repetition.

Training: who needs to learn what

A rule becomes routine when each role knows what to do with it. Domain owners learn to read the indicators, to decide field requests with criteria and to hold their own team to the data entry standard in the pipeline meeting. The steward and the administrator learn to maintain the dictionary, to assess imports and to translate decisions into configuration. Sales reps, SDRs and CS learn the policy for their stage in the company’s own CRM, with the report each field feeds. In the teams I follow, adherence grows when the rep sees during training that the field they fill in decides their commission.

30-day roadmap for leadership

1. Week 1: name the sponsor and the steward. Write down the names, what the steward approves alone and what goes to the committee, and tell the team.
2. Week 1: choose the first domain. Start with deals and pipeline and list the ten to fifteen critical fields.
3. Week 2: write the definitions. Each critical field gets a definition, an owner, who fills it in, at which stage and which report uses it, directly in the data dictionary.
4. Week 2: build the RACI. Adapt the matrix in this article, with a single approver per decision, and get the sponsor’s signature.
5. Week 3: publish the five policies, one page each. Fields, data entry, duplicates, retention and access, with legal reviewing the retention policy.
6. Week 3: take the baseline. Measure the health indicators for the chosen domain before any correction.
7. Week 4: hold the first committee. Use the agenda in the table, decide the accumulated requests and open the decision log.
8. Week 4: prepare each role. Show the team, in the CRM, what changed at their stage and why, and schedule the next committee.

The same structure of owner, committee and log serves AI governance, and many companies use the data committee to approve the agents that read the CRM.

If your board argues about which number is right before discussing the result, that is the work of the in-company CRM adoption training. I run the program with the sponsor, the domain owners, the steward and each role on the team, in the CRM you already use. The group leaves with the CRM best-practices guide, including the definitions of the critical fields and the data entry policies, and with the portfolio and database management manual, including the roles matrix, the committee routine and the database health indicators.

Frequently asked questions

What is CRM data governance?

It is the set of roles, policies and routines that defines who decides about each piece of sales data, from fields and picklists to imports, access and deletion. It gives leadership an owner for each domain and a committee that decides once a month, with a record.

What is the difference between data governance and data quality?

Quality measures the state of the database: completeness, accuracy, duplication, freshness and standardization. Governance defines who decides the rules that produce that quality and who answers when they break.

Who should own the CRM data?

Each domain has its own owner, the leader of the team that depends most on that data, such as sales leadership for deals and marketing leadership for leads. The data steward, usually from revenue operations, keeps the decisions working, and the CRM administrator configures what was approved.

Is data governance required by the LGPD?

The LGPD requires principles such as accuracy, purpose limitation and security, and Article 50 provides that the controller may implement a privacy governance program. Sales governance covers that program for personal data and goes further, because it includes value, stage, segment and product, which weigh on business decisions.

Does a small company need a data committee?

It needs the routine, even if the committee is the sales director, the operations person and the marketing lead in a half-hour monthly meeting. What cannot be missing is a written owner for each domain and the decision log.

How do you prepare the team to follow data governance?

Train by role, in the company’s own CRM: domain owners learn to decide and to read indicators, the steward and the administrator learn to maintain the dictionary and the configuration, and each user learns the policy for their stage. Start with the leaders, because they are the ones who hold the team to the standard, and measure through the change in the database health indicators.