Before fixing your Salesforce data, assess it
Before your automations and AI can trust your Salesforce data, you need to assess it. Most teams skip that step and go straight to fixing. This is the guide for doing it in the right order, so the fixes hold, and your data is actually ready for what you are asking it to do.
THE SITUATION
Your data problems don't announce themselves
By the time a data issue shows up in a report, a campaign that underperformed, a lead score that made no sense, an AI model that returned the wrong segment, the damage is already done. The record was wrong before it got there. It was wrong when it was created, or when it was updated, or when it was imported six quarters ago. You just didn't have visibility into that.
Marketing Ops and Salesforce Admins are expected to deliver reliable pipeline, accurate reporting, and AI-ready campaigns. But that expectation sits on top of data that most teams have never fully examined. You optimize the subject line. You tune the scoring model. You adjust the routing rules. The underlying data stays unexamined.
That's not a failure of effort. It's a failure of sequence. Most teams only look at data quality after something breaks. They're reactive by default. Not because they don't care, but because no one handed them a clear starting point.
THE SHIFT
You don't have a data quality problem. You have a data confidence problem.
And the reason it persists is almost always the same: teams go straight to fixing, or straight to AI and automation, without ever doing the diagnosis first.
Think about how most teams respond when a campaign underperforms
They tweak the subject line. They adjust the scoring threshold. They revisit the segment definition. What they rarely do is look inside the data itself, at the fields driving that segment, at the fill rates on those fields, at whether the values in those fields actually mean what the field name suggests.That's treating symptoms. It's not a diagnosis.What you actually need is the scan first. An MRI before the surgery. A clear picture of what's healthy, what's damaged, and what needs intervention, before you touch anything. That picture changes everything: it tells you what to fix and in what order, instead of just where to start cleaning.You cannot fix what you haven't assessed. You cannot trust automation or AI running on data you've never checked. That's the whole argument, and it explains why the sequence matters: Assess. Fix. Monitor. In that order.
Think about how most teams respond when a campaign underperforms. They tweak the subject line. They adjust the scoring threshold. They revisit the segment definition. What they rarely do is look inside the data itself, at the fields driving that segment, at the fill rates on those fields, at whether the values in those fields actually mean what the field name suggests.
That's treating symptoms. It's not a diagnosis.
What you actually need is the scan first. An MRI before the surgery. A clear picture of what's healthy, what's damaged, and what needs intervention, before you touch anything. That picture changes everything: it tells you what to fix and in what order, instead of just where to start cleaning.
You cannot fix what you haven't assessed. You cannot trust automation or AI running on data you've never checked. That's the whole argument, and it explains why the sequence matters: Assess. Fix. Monitor. In that order.
THE FRAMEWORK
Three steps. In this order.
This is what gets you to data you can actually trust, for campaigns, for automations, for reporting, for AI. Each step depends on the one before it. That's not bureaucracy. That's just how diagnosis works.
1
Assess: understand what you actually haveBefore you fix anything, you need a diagnosis. This is not optional, and it is not a delay; it's what makes every fix targeted instead of guesswork. Pick the object causing the most pain right now, usually Contacts or Leads, and look at it properly.The key shift: you are not running a cleanup. You are running an analysis. The goal of this step is visibility, not action. You want to see clearly, probably for the first time, what you are actually working with.For each object, you want to understand which fields are actually being used and which are empty or inconsistent. Whether users trust the data in those fields, because low trust is often a bigger problem than low fill rates. Where values contradict what the field is supposed to mean. Which records are complete enough to rely on for campaign, scoring, and AI.
1. Assess: understand what you actually have
Before you fix anything, you need a diagnosis. This is not optional, and it is not a delay; it's what makes every fix targeted instead of guesswork. Pick the object causing the most pain right now, usually Contacts or Leads, and look at it properly.
The key shift: you are not running a cleanup. You are running an analysis. The goal of this step is visibility, not action. You want to see clearly, probably for the first time, what you are actually working with.
For each object, you want to understand which fields are actually being used and which are empty or inconsistent. Whether users trust the data in those fields, because low trust is often a bigger problem than low fill rates. Where values contradict what the field is supposed to mean. Which records are complete enough to rely on for campaign, scoring, and AI.
Where to start
- Pick one object: Contacts or Leads, whichever is causing the most visible problems right now
- Identify the 10 fields that drive your most important campaigns, automations, or scoring models
- Check fill rates, look for impossible values (future birthdates, invalid emails), review picklist distributions for values that shouldn't exist
- Document what each field is supposed to mean versus how the data is actually being used; the gap between those two things is where most problems live
2
Fix: act on what the assessment surfacedNow you have the diagnosis. The fix is targeted because you know exactly what is wrong and where. This is not a bulk cleanup; it's a prioritized response to what the assessment found. That's the reason the order matters: fixing without assessing means cleaning the wrong things first, in the wrong sequence, for the wrong reasons.One rule that is easy to miss: deduplicate before you validate. Merging records first makes everything else cleaner and faster. Validating contact data on records that will later be merged wastes the work.
2. Fix: act on what the assessment surfaced
Now you have the diagnosis. The fix is targeted because you know exactly what is wrong and where. This is not a bulk cleanup; it's a prioritized response to what the assessment found. That's the reason the order matters: fixing without assessing means cleaning the wrong things first, in the wrong sequence, for the wrong reasons.
One rule that is easy to miss: deduplicate before you validate. Merging records first makes everything else cleaner and faster. Validating contact data on records that will later be merged wastes the work.
Where to start
- Start with the highest-impact fields: the ones feeding your next campaign, automation, or AI model, not the longest list of problems
- Deduplicate first: merge records before you validate anything
- Verify contact data (emails, phone numbers) on the records that matter most before they reach Marketing Cloud or an AI agent
- Fix the documentation in parallel: assign field owners, write field descriptions, and record what "good" looks like for each field. This pays back fast when the next project starts
3
Monitor: keep it clean so the fix holdsA one-time cleanup reverts. New records come in. Fields get repurposed. Teams stop following the rules. Without monitoring, you are back to guessing within three months, often without noticing until something breaks.The concept worth building around: sentinel fields. These are the five to ten fields that, if they degrade, signal broader problems in your data. You do not need to monitor everything. You need to monitor the right things and respond when the numbers move.Data quality only improves if someone is watching it. Monitoring turns a one-time fix into a program.
3. Monitor: keep it clean so the fix holds
A one-time cleanup reverts. New records come in. Fields get repurposed. Teams stop following the rules. Without monitoring, you are back to guessing within three months, often without noticing until something breaks.
The concept worth building around: sentinel fields. These are the five to ten fields that, if they degrade, signal broader problems in your data. You do not need to monitor everything. You need to monitor the right things and respond when the numbers move.
Data quality only improves if someone is watching it. Monitoring turns a one-time fix into a program.
Where to start
- Identify your sentinel fields, the ones whose degradation signals broader problems
- Schedule recurring checks on fill rates and key quality metrics for those fields
- Set thresholds: what does "good" look like for each sentinel field, and what triggers action when it drops?
- Make the results visible to the team: quality improves when it's shared, not when it lives in one admin's reports
Recommended videos from our Demo HUB
YOUR ROADMAP
What to do and when
The framework above gives you the what. This gives you the when. Not everything needs to happen this week, and trying to do it all at once is one of the main reasons these programs stall. The destination is a Salesforce org you can confidently hand to an automation, a reporting system, or an AI.
THIS WEEK: Get your first diagnosis
Pick one object, Contacts or Leads, whichever is causing the most visible pain right nowIdentify the five to ten fields that drive your most important campaigns, automations, or scoring models. Run a basic health check on those fields: fill rates, impossible values, picklist distributions. Document what each field is supposed to mean versus what the data actually shows.The goal is not to fix anything yet. The goal is to see clearly, probably for the first time, what you are actually working with.Expected outcome: A short list of the highest-risk fields in your most important object. You now know where the problems are.
WEEKS 2-4: Fix what matters most, document in parallel
Use what the assessment surfaced. Start with the fields that feed your next campaign or automation, not the biggest cleanup, the most urgent one. Deduplicate first, then validate. Fix the documentation in parallel: assign field owners, write field descriptions, record what "good" looks like. This step pays back fast when the next project starts.Expected outcome: Your highest-impact fields are clean, validated, and documented. Your next campaign runs on data you have actually checked. You know which fields to trust and which to treat with caution.MONTH 2: Extend the assessment across the org
Repeat Phase 1 on the next most important object. Broaden the field list. Start building a picture of data health across the org, not just the object you fixed first. This is where you move from "we fixed a problem" to "we understand our data."Share what you found. A one-page summary of field health, risks identified, and fixes made builds internal trust in Salesforce data — and starts making data quality a team concern instead of an admin burden.Expected outcome: A documented baseline across more than one object. Leadership and stakeholders can see the evidence.ONGOING: Make it a program, not a project
Set up recurring checks on the fields that matter most. Define thresholds: what does "good" look like for each sentinel field, and what triggers action when it drops? Set alerts. Review trends. Catch problems before they reach campaigns, automations, or AI, not after.This is what makes trusted data real. A cleanup without monitoring reverts. A monitored org improves over time and becomes the kind of foundation that automations and AI can actually rely on.Expected outcome: Data quality is no longer reactive. You are not the person who finds out something was wrong after a campaign went out. You are the person who knew before it did.Assess. Fix. Monitor. That is the sequence. That should be your mantra. By the time your next major campaign or AI initiative lands, you could already know which data to trust, and have the evidence to prove it.THIS WEEK: Get your first diagnosis
Pick one object, Contacts or Leads, whichever is causing the most visible pain right now. Identify the five to ten fields that drive your most important campaigns, automations, or scoring models. Run a basic health check on those fields: fill rates, impossible values, picklist distributions. Document what each field is supposed to mean versus what the data actually shows.
The goal is not to fix anything yet. The goal is to see clearly, probably for the first time, what you are actually working with.
Expected outcome: A short list of the highest-risk fields in your most important object. You now know where the problems are.
WEEKS 2-4: Fix what matters most, document in parallel
Use what the assessment surfaced. Start with the fields that feed your next campaign or automation, not the biggest cleanup, the most urgent one. Deduplicate first, then validate. Fix the documentation in parallel: assign field owners, write field descriptions, record what "good" looks like. This step pays back fast when the next project starts.
Expected outcome: Your highest-impact fields are clean, validated, and documented. Your next campaign runs on data you have actually checked. You know which fields to trust and which to treat with caution.
MONTH 2: Extend the assessment across the org
Repeat Phase 1 on the next most important object. Broaden the field list. Start building a picture of data health across the org, not just the object you fixed first. This is where you move from "we fixed a problem" to "we understand our data."
Share what you found. A one-page summary of field health, risks identified, and fixes made builds internal trust in Salesforce data — and starts making data quality a team concern instead of an admin burden.
Expected outcome: A documented baseline across more than one object. Leadership and stakeholders can see the evidence.
ONGOING: Make it a program, not a project
Set up recurring checks on the fields that matter most. Define thresholds: what does "good" look like for each sentinel field, and what triggers action when it drops? Set alerts. Review trends. Catch problems before they reach campaigns, automations, or AI, not after.
This is what makes trusted data real. A cleanup without monitoring reverts. A monitored org improves over time and becomes the kind of foundation that automations and AI can actually rely on.
Expected outcome: Data quality is no longer reactive. You are not the person who finds out something was wrong after a campaign went out. You are the person who knew before it did.
Assess. Fix. Monitor. That is the sequence. That should be your mantra. By the time your next major campaign or AI initiative lands, you could already know which data to trust, and have the evidence to prove it.
Assess
Fix. Monitor.That is the sequence. The rest follows from it.
Assess. Fix. Monitor.
That is the sequence. The rest follows from it.
