Pipeline Leak Audit · System 01 · what your sales records show

What Your Sales Records Can Tell You

A deal moves to the next step in your sales software for one of two reasons: the customer really moved forward, or someone updated the record. A salesperson pushes a deal forward to keep it off the “stuck deals” report. A week of updates gets typed in on a Friday afternoon, dated whenever felt right. A deal jumps straight from first meeting to proposal because filling in the steps in between felt like paperwork nobody would read. In a step-by-step report, every one of those looks exactly like real progress.

The idea in one line: build a deal-by-deal list from what the system recorded, mark the parts that could reflect a salesperson’s habits rather than the customer’s decisions, and treat anything you can’t confirm as a question to answer.

What your sales software already records

All three of the sales systems small businesses use most can give you a usable history of how each deal moved. None of them hands it over ready to use.

Salesforce

Records every change of step automatically, and keeps the date of the most recent one. To make it usable you still need to set it up: make close and loss reasons required, keep one consistent list of steps, and build the right report.

HubSpot

The most complete out of the box. It records the date each deal entered and left every step, how long it’s been at its current step, and flags stuck deals automatically, with no setup needed.

Pipedrive

Good at showing the current step, activities, and how long things take. A full line-by-line history of every past move is harder to get, and usually means its reporting add-on rather than a plain download.

Whichever you use, one thing stays true. The date on a step tells you when the record changed. On its own, it doesn’t tell you when the customer moved, and the gap between those two is where most sales reporting goes wrong before anyone starts analyzing it.

Six ways the record misleads you

None of these is deliberate. Each is a salesperson doing their job under time pressure, in a system that was never going to notice.

What to check before you trust a number

Skipped steps: a deal jumps from first meeting straight to proposal. A summary report assumes everything in between happened. Count how many deals ever reached each step, and how many got there directly from the step before; the difference is how often steps get skipped.

Moving backwards: deals get moved back, reopened, or revived after going cold. “Time at this step” then understates how long the deal has really been open. Count how often each deal went back, and track its total age separately.

Late updates: salespeople update the system at the end of the week, the end of the month, or when a manager asks. The date shows when the system was told, not when the customer decided. Don’t treat it as the real decision date unless an email, call, or meeting backs it up.

Blank or vague loss reasons: a salesperson closes a lost deal quickly and picks whatever’s fastest, or leaves it blank. “Lost to a competitor,” “timing,” and “no decision” become unreliable exactly where you need them most.

Work that never got logged: the call was on a mobile, the email is in a personal inbox, the follow-up was a text nobody recorded. “No activity” can mean nothing happened, or that nobody wrote it down. Those are different findings.

A deal that looks busy but has nothing booked: a note was added, an email was sent, but nothing with a date is on the calendar. This is the most useful kind of “stuck,” and any report that only checks for recent activity misses it.

None of these six means you should distrust your sales records altogether. They’re why a sales audit starts by sorting out what the record can support, before anything gets added up.

Build the deal list

One row per deal, showing every step it moved through. The rest of this audit runs on this list. Download or rebuild these details.

What the list needs

The deal: a reference, the customer, who owns it, where it came from, and its expected value

Every move: from which step, to which step, the date, and how many times it went back to an earlier step

Where it stands: open, won, or lost, the close date if closed, and the loss reason plus any note

Activity: the last time anything was logged, the last time you spoke with the customer if you can tell, the next thing booked, and how many calls, emails, and meetings there were

People: how many people at the customer you’ve spoken with, whether you know who signs off on the money, and when your main contact was last in touch

This is a spreadsheet. Download it where your sales software offers a history report; where it doesn’t, rebuild the last three months by hand from each deal’s activity history.

Rebuilding this deal by deal is slow by hand, and it’s exactly the kind of sorting an AI assistant does quickly, as long as you keep it collecting facts and holding back on conclusions.

Do this step with an AI assistant

Download whatever deal-history or activity report your sales software offers, and send it in one chat with these two messages, in order. Most of the value is in the second: it asks the assistant to point out what the data can’t support, rather than filling the gap with a guess.

“Here’s a deal-history report from my sales software. Turn it into one row per deal: current step, every move with its date, how many times it went backwards, the last and next activity dates, and the loss reason if it was lost. Don’t interpret it yet.”
“For each deal, point out anything the data can’t confirm: a step skipped on the way to where it is now, a move that happened right after a long gap with nothing logged, or a loss reason that’s blank or vague. List those separately from the deals where the record looks clean.”

Decide what “stuck” means for each kind of deal

“Stuck” only means something compared with how that kind of deal usually moves.

Two weeks of silence is normal for a big contract that takes a quarter to approve, and a real warning sign on a small deal that usually closes in ten days. So set the rule separately for each kind of deal you sell. A deal is stuck when it’s open, has nothing booked, has had no contact with the customer for longer than usual for that kind of deal, and has been at its current step longer than that kind of deal usually takes. HubSpot’s built-in “stuck” flag works on a similar idea (a deal sitting at a step much longer than that salesperson’s won deals usually did), which is a sensible starting point if you don’t want to set your own rules yet.

Write your rules down before you flag anything

Kind of deal: ______ (by size, product, or anything else that changes how long a sale takes)

Usual time at this step: ______ (the middle value from deals of this kind you won)

Usual gap between contacts with the customer: ______

Stuck means: open, nothing booked, no contact for longer than the gap above, and at this step longer than usual for this kind of deal

A deal that meets this rule isn’t necessarily dead. It needs someone to make a decision about it, which is what System 03 is for.

A worked example

A 14-person firm selling services to other businesses runs its sales in HubSpot and believes it’s been winning fewer deals. The history report covers the last six months and 61 deals.

Laid out as a deal-by-deal list, three things show up that the summary report never did:

What the list showed

9 of the 61 deals skip “Needs Analysis” entirely, jumping from first meeting straight to proposal, and all 9 belong to the same salesperson.
14 deals sit at “Proposal Sent” with nothing booked and no contact for more than three weeks, well past the usual 9 days for this kind of deal.
22 of the 31 deals marked lost this period have “Other” as the reason, with no note.

What that changes

The skipped step isn’t a broken sales process. It’s one person’s habit, worth a five-minute conversation.
The 14 stuck proposals are the finding worth sizing: real money sitting with nobody assigned to move it.
With 70% of losses marked “Other,” the question this audit was meant to answer (why are we losing?) can’t be answered yet, so fixing that field goes to the top of the list.

None of that showed up in the falling win-rate the firm started with. The trend said “something is worse.” The deal list said where, and separated one person’s habit, a real backlog, and a record-keeping problem: three different things a single percentage had mixed into one.

Once you have a deal list you trust and a rule for “stuck” that fits each kind of deal, you have a list of deals worth a second look, and a question mark next to the ones you can’t trust yet. What you don’t have yet is a reason for each one. System 02 covers the common reasons a deal gets stuck, and how to tell one person’s habits from a problem with the whole process.

Next in the Pipeline Leak Audit · System 02Why Deals Actually Stall 8 min read