You can spot it in about four minutes of a monthly review. Marketing says the quarter produced 312 leads. Sales says they saw about 40 worth calling. Nobody in the room can reconcile the two, so everybody agrees to look into it, and then nobody does.
That gap looks like a reporting problem. Underneath it is a definitions problem, and it quietly costs more than most of the things on your budget line.
Rarely through one bad decision. It is usually four years of reasonable ones.
Somebody set up lifecycle stages during onboarding and picked the defaults. A workflow was built to move contacts to MQL when they downloaded anything. Six months later a campaign needed a different rule, so a second workflow was added rather than the first one changed. Then an integration started creating contacts at Lead. Then somebody imported a list from an event and mapped Lifecycle Stage to a column that was not lifecycle stage.
Now the funnel chart is a work of fiction and everyone has privately stopped believing it. The tell is when people start keeping their own spreadsheet. Once that happens you are paying for HubSpot and running the business on Excel.
Budget decisions get made on vibes. If nobody trusts sourced pipeline, the marketing budget is defended with anecdotes and last year's number. That is how good programmes get cut in a bad quarter.
Sales follow-up gets slower. When routing rules have drifted, enquiries sit. We audited a portal last year where 11% of form submissions had no owner for more than three days. Nobody had done anything wrong. A rep had left and the rule still pointed at them.
Your best channel is invisible. If first touch is being overwritten by a later interaction, the thing that actually started the deal disappears from the report. We have seen webinars and research reports carry a year of pipeline and show as nothing.
Onboarding new people takes twice as long. A portal that only makes sense if somebody explains it is a portal with a person-shaped dependency in it.
Write the definitions down first, with sales in the room. In a document, before anyone touches HubSpot. What is an MQL here. What has to be true for a contact to be an SQL. Who decides. What happens to a contact that goes cold. Half an hour of argument now saves a year of it later, and the argument is the point. If marketing and sales cannot agree on the definition in a meeting, no workflow is going to reconcile them.
Audit the workflows before you build anything. Export the list. For each one, name the owner and the reason it exists. In most portals we look at, between a quarter and a third can be turned off with nobody noticing. Turn them off before you add more.
Enforce lifecycle in one place. One workflow sets lifecycle stage. Everything else reads it. If an integration or an import needs to set it, that is a conversation to have rather than a default to accept.
Fix imports at source. Most data drift arrives through imports. Agree a mapping template, use it every time, and make one person responsible for approving imports over a certain size.
Then build one dashboard both teams use. Not a marketing dashboard and a sales dashboard. One. If a number appears on it, both teams have agreed what it means.
The audit is a week. The definitions workshop is half a day and a fortnight of follow-up because somebody always needs to check with somebody. The rebuild is three to five weeks in a mid-sized portal, longer if you have a Salesforce sync in play, because you have to agree the system of record for each object before you touch anything.
Budget for the reporting looking worse before it looks better. When you restate lifecycle stages honestly, the funnel usually gets narrower. That is the number being right, and it is worth saying out loud before the board sees the chart.
This is most of what our RevOps work consists of. It is unglamorous and it is the thing that decides whether anyone believes the rest of the marketing.