RevOps/No. 08/6 min read

You don’t have a reporting problem. You have a data problem.

Three dashboards, three different pipeline numbers, and the instinct is to buy a fourth. The fix isn’t a new reporting layer. It’s scoring the data underneath.

SM

Sage M.

LSR Marketing · July 2, 2026

If your pipeline dashboards disagree with each other, the fix is almost never another reporting tool. It’s a data quality score: a recurring measurement of how complete, consistent, and current the CRM properties your reports depend on actually are. Score the data first. Then decide if you still have a reporting problem. You probably won’t.

We see the same sequence on repeat. Marketing builds a funnel report. Sales builds theirs. The numbers don’t match, so leadership buys an attribution tool or another BI layer to play referee. Six months later there are three versions of pipeline and none of them survive a board question. Every new layer inherited the same broken inputs.

Why every new dashboard fails the same way

A report is just a query over a handful of properties. Lifecycle stage. Deal stage. Amount. Close date. Original source. Owner. If those fields are half-empty, or mean different things to different teams, no visualization layer can fix it. The new tool renders the disagreement in nicer colors.

This is expensive in ways nobody itemizes. Gartner pegs the average cost of poor-quality data at $12.9 million a year, and the sharper finding in the same research is that most organizations don’t measure data quality at all. They feel it in missed forecasts and automation that misfires, then they blame the tools.

The definitional rot is the sneakiest part. When “Lead” and “MQL” catch the same contacts, every conversion rate downstream is fiction. We wrote a whole field note on the lifecycle stages most HubSpot accounts get wrong, and the reporting damage from that one mistake is exactly the kind of thing a new dashboard can’t repair.

Score your data before you trust a single chart

The fix starts with measurement, and it’s less work than you think. One afternoon. One spreadsheet. No new tooling.

Pick the properties your reports actually run on

Not all 400 properties in your portal. The 12 to 20 your revenue reports query. For most B2B teams the list looks like: Lifecycle stage, Lead status, Original source, Contact owner, Deal stage, Amount, Close date, Create date, plus whatever your routing and segmentation depend on. Usually Industry, Country, and company size. If a property doesn’t feed a report, a route, or an automation, it’s out of scope.

Run three checks: complete, consistent, current

Complete is a populated percentage. What share of open deals have an Amount? What share of contacts that hit MQL have an Original source? Pull the numbers per property. Most teams guess 90% and find 60%.

Consistent is one meaning per field. Picklist values that don’t overlap. No free-text Industry field with fourteen spellings of “financial services.” No two teams using Lead status for different jobs. If you have to ask someone what a value means, the field fails.

Current is decay. Owners who left the company six months ago. Open deals with close dates in the past. Contacts whose last activity predates your current ICP. Data doesn’t stay fixed; it rots quietly until a report steps on it.

Average the checks across your property list and you have a data quality score per object. It will be lower than anyone expected. That’s the point. Now it’s a number you can move instead of a feeling everyone argues about.

Fix the point of entry, not the quarterly cleanup

A one-time cleanup feels productive and starts decaying the day it ships. The durable fixes are structural, and they all live at the point of capture.

Validation at entry. HubSpot supports validation rules on properties: min and max values on numbers, regex patterns on text, formatting rules on phone fields. Turn them on for the properties in your scope. Every bad value you block at entry is a row you never clean.

Required fields at stage gates. A deal doesn’t move to Proposal without an Amount. A contact doesn’t hit MQL without an Original source. The rep fills it in when the context is fresh, not during a quarter-end archaeology dig.

Automation stamps, humans decide. Dates and stage timestamps get set by workflows, not memory. People are good at judgment and terrible at data entry. Structure the system so they only do the first one.

One owner per property. Every field in scope has a named owner and a written definition. This is the data model work underneath everything else, and it’s where most of our Marketing Operations & CRM engagements start. Most CRM problems are data problems. You can’t automate your way out of a broken object model.

Report the score next to the pipeline number

Here’s what changes when the score becomes a standing metric instead of a one-off audit: it gets defended. Put CRM data quality in the same monthly review as pipeline and win rate. When completeness on Amount drops eight points, someone notices in a week instead of a fiscal year.

And the downstream conversations get shorter. Forecast reviews stop being arguments about whose number is right, because there’s one set of inputs and a visible measure of how much to trust it. That’s the real deliverable: not cleaner data for its own sake, but reports that survive the follow-up question.

If you want the starting structure, our CRM Data Model Worksheet is the same one we use to map properties, owners, and definitions in client engagements.

One more reason to do this now. Every automation you run, every lead score you compute, every AI tool you’re about to buy reads these same properties. Broken in, broken out. The teams getting real output from AI this year aren’t the ones with the best prompts. They’re the ones whose data was worth reading.