How to run a 60-minute RevOps diagnostic on your own HubSpot

Most teams don't have a marketing problem. They have a definitions problem. Before you spend another quarter on campaigns, spend an hour finding out whether the system underneath can tell you what already works.

1. Why an hour is enough

You are not auditing everything. You are looking for the two or three places where a record changes meaning as it moves between teams. Those are the points where pipeline quietly leaks, and they show up fast once you know where to look.

2. Start with the data model

Open your contact and company properties and count how many are actually used in a workflow, a report or a view. In most portals a third of them are dead weight from an old process, and at least one duplicate field is being written to by two different sources.

• Which properties drive routing or scoring?
• Which are set by a form, an integration and a human, all writing to the same place?
• What happens when two of them disagree?

Three sources, one field, no agreed order. This is what a disagreement looks like before anyone calls it a reporting problem.

Figure 1 One property, three writers

Three systems write the same three fields and none of them is the owner. The geometry is the whole finding: everything converges on one place, and nothing decides what happens when they disagree.

• FormWrites on submit
• Enrichment integrationWrites on sync
• Rep, by handOverwrites later

One contact property Lifecycle stage, lead source, owner. Last write wins: no rule, no owner.

3. Follow one lead end to end

Pick a real inbound lead from last week and trace it: the form it filled, the properties it set, the workflow that fired, the owner it landed with, and how long it sat before someone touched it. Do this once and you will usually find the gap without needing a report.

4. Check what a stage really means

Ask marketing and sales, separately, to define an MQL and an SQL. If the two answers differ, every conversion metric between them is fiction. This is the single most common finding, and the cheapest to fix.

Nothing here needs new tooling. It needs both teams to say the sentence out loud and agree it is the same sentence.

Figure 2 Why the conversion rate between the two is fiction

On the left, marketing and sales each hold their own definition, so the arrow between them runs from one population to a different one. On the right there is one sentence, one property and one writer, so the arrow runs inside a single set.

Two definitions

Two populations, and a rate computed across both

Marketing counts an MQL when the form is filled and the score clears a threshold.

Sales counts one when a rep has actually spoken to somebody worth speaking to.

The conversion rate between the two is arithmetic on two different populations.

One definition

One set, with an entry rule and an exit rule

One written sentence per stage, agreed in the room, with an owner against it.

One property in the CRM that carries it, with one system allowed to write to it.

An entry and an exit rule, so a record cannot sit in a stage it no longer qualifies for.

5. Rebuild one number by hand

Take the pipeline figure your dashboard reports for last month and rebuild it from the deal records. If you cannot land on the same number, you have found why nobody trusts the reporting.

6. Decide what to fix first

Rank what you found by how much revenue it touches, not by how easy it is. Clean definitions before automation, routing before scoring, and reporting last, because the report is only as honest as the model beneath it.

Figure 3 The order to fix in

Each one depends on the one above it. Reporting sits last not because it matters least but because it is reading everything above it.

1. First Definitions MQL, SQL, stages
2. Then Routing Owner, territory, fit
3. Next Scoring & automation Only once the base holds
4. Last Reporting Honest by then

The order matters more than the effort. Fix the base and the campaigns you already run start reporting the truth.