Before you blame the software, check the usual suspects. Most CRM problems trace back to the same handful of issues, and you can find nearly all of them in ten minutes using screens you already have open. This is the audit we run on every account before we touch a setting.
Before You Start: Three Rules
A ten-minute audit only works if you stay disciplined. Find the first broken thing, spend an hour on it, and you never see the other seven.
- Find, don't fix. Write every problem on one list as you go. Fixing comes after the ten minutes, in priority order.
- Use the last 30 days. Set every filter and report to the same window, or you will draw conclusions from numbers that do not line up.
- Audit as a rep, not an admin. Admins see everything. Where you can, look through a normal user's account. That is the system your team works in.
Minute 1: Lead Response Time
This is the most expensive number in your CRM, and almost nobody measures it honestly. Filter to leads created in the last 30 days and pick ten at random. For each, compare the timestamp on the record with the first outbound message or call. Not the automated confirmation. The first real attempt to reach that person.
What good looks like
Every lead gets an automated acknowledgment within seconds and a human attempt inside five minutes during business hours. Under 60 seconds is where top performers operate. If your ten samples average hours, you have found your biggest leak.
Fastest fix: build one workflow triggered on form submission that sends an instant SMS to the lead and a notification to whoever owns the queue. That one automation usually changes more than the rest of this audit combined. Our guide to choosing the right trigger covers which event to fire on.
Minute 2: Unassigned Leads
Filter your contacts and open opportunities by owner and look for the unassigned or blank bucket. Then do the half most people skip: filter by owner for anyone who has left or changed roles. Those records are just as orphaned, and harder to spot because they look assigned.
What good looks like
Zero unassigned records, and every owner on the list still works there and still works those leads. A lead should get an owner the moment it is created, not when somebody notices it.
Fastest fix: bulk-reassign the orphans today, then add an assignment step to the workflow that creates the lead. Round-robin distribution is the simplest version, and one of the first things we set up in an automation build. If you route by territory, product, or deal size, write that rule down where your team can read it.
Minute 3: Duplicate Records
Duplicates are quiet. They throw no errors. They just split one person's history across two records, so your rep sees half the conversation and your reports count the same lead twice. Sort contacts by last name and scan for repeats, then search two or three of your most common email domains for the same person appearing twice.
What good looks like
One record per person, with email and phone treated as the unique identifier. A few duplicates in a large database is normal housekeeping. If the same lead appears three or four times, something upstream is creating records instead of updating them.
Fastest fix: deal with the source before you clean the data, or you will be doing this again next month. The usual culprits: a form that creates a new contact on every submission, an import that ran without matching on email, two integrations writing the same lead in from different directions. Consolidate what you found, keeping the record with the real history, then fix the intake path. Duplicate cleanup is standard in our migration process for this reason.
Minute 4: Dead Automations
Open your workflow list and sort by last activity or enrollment count. Two failures to look for: workflows published but not enrolling anyone in weeks, and workflows built, tested, and never turned on.
For each one that looks dead, check three things in order. Is it published and active? Is the trigger pointed at a form, tag, or stage still in use? Are the filters excluding everyone who would otherwise qualify? Renamed forms and retired tags break more automations than anything else.
What good looks like
Every published workflow has enrolled someone recently, or has an obvious seasonal reason why it has not. Nothing sits in draft that your team believes is running.
Fastest fix: publish the ones that should be live, repoint the broken triggers, and archive the rest. Archiving matters more than it sounds: a list full of dead entries is how conflicting automations get built, because nobody can tell what is already running.
Minutes 5 and 6: Stage Definitions
Count your pipeline stages, then ask a harder question about each one: what has to be true for a deal to move into it, and what has to be true for it to move out? If two people on your team would answer differently, the stage is not a stage, it is a feeling. "Proposal Sent" is verifiable. "Warm" is not.
Then look at where the deals actually sit. If most of your open pipeline is bunched into one or two stages, the rest are decoration. If a stage has not held a deal in a month, it does not exist in practice.
What good looks like
Five to seven stages, each named after a completed action, each with a written entry criterion, and deals spread across them rather than piled at one end. More than seven creates confusion. Fewer than five hides what is happening.
Fastest fix: rename vague stages to action-based ones and merge the ones nobody uses. Give each an expected duration while you are in there, because that is what makes stale deal alerts possible. Our pipeline guide walks through stage design in more depth.
Minute 7: Required Fields
Required fields are the most common reason a deal will not move and nobody can explain why. Open a real opportunity and drag it forward a stage. If it refuses, or demands three fields your rep has no way of knowing yet, you have found a bottleneck your team has been working around for months by leaving deals in the wrong stage.
While you are there, open your main intake form and count the fields. Every extra one is a reason not to finish it. Ask of each: will somebody use this in the next 24 hours?
What good looks like
Intake forms ask for the minimum needed to make contact and route correctly, usually name, phone, email, and lead source. Anything more detailed is required later in the pipeline, at the stage where the rep would naturally have learned it.
Fastest fix: strip required fields off early stages and forms, and re-apply the important ones deeper in the pipeline. Keep lead source required everywhere. It is the field that tells you which marketing to keep paying for, and almost always the first one people delete.
Minute 8: Notification Noise
Count how many events send you an alert, then ask a rep how many CRM notifications they read versus dismiss. The honest answer is usually that most get dismissed, which means the one that mattered got dismissed too.
Sort your alerts into two piles. Pile one needs a human within the hour: a new inbound lead, a reply from a prospect, a missed call. Pile two is useful eventually: stage changes, completed tasks, form submissions from existing contacts.
What good looks like
Pile one goes to the phone as SMS or push and stays rare enough that people still react to it. Pile two goes to a daily digest, or nowhere at all. If a notification does not name a person to act and a thing to do, it should not interrupt anyone.
Fastest fix: turn off everything in pile two today. Then route pile one to the specific owner rather than the whole team, because an alert that goes to everyone belongs to no one.
Minute 9: Reporting Basics
Open your dashboard and answer four questions without exporting anything: how many leads came in last month, where they came from, what share booked an appointment, and how many deals closed. If you cannot answer all four in under a minute, your reporting is decorative.
The usual failure is not a missing dashboard, it is a missing lead source. If many of your leads show a blank source, no report will tell you which channel to spend more on.
What good looks like
One dashboard, a handful of widgets, and every lead carrying a source. Leads in, source, appointments booked, deals closed, and speed-to-lead is enough to run a business on. Fifty widgets is not more insight, it is somewhere to hide.
Fastest fix: go to Reporting → Create Custom Dashboard, add a widget for each of those numbers, set the date filter to the last 30 days, and schedule it to email itself to you every Monday. A report you have to remember to open is one you will stop opening. For the longer list of numbers worth watching, read the metrics that actually matter.
Minute 10: Can You Actually Reach People?
Everything above assumes your messages arrive. Two settings decide that, and both fail silently: nothing bounces, nothing errors, the send goes nowhere.
For email, go to Settings → Email → Domain Settings and confirm SPF, DKIM, and DMARC all show green, then send a test through mail-tester.com. Anything below 7 out of 10 needs work before your next campaign.
For SMS, go to Settings → Phone Numbers. Each number should show as active and be attached to an approved campaign. If you have not completed A2P 10DLC registration, carriers are why your texts are not landing, and no workflow debugging will change that.
What good looks like
Three green authentication records, a mail-tester score of 8 or better, active numbers on a registered campaign, and a test that actually arrived on your own phone and in your own inbox. Trusting the sent log is how this goes unnoticed for months.
What to Fix First
You now have a list. Work it in this order. The top items are losing you money; the bottom ones are losing you clarity.
- Anything blocking delivery. Email authentication and A2P registration. Every other fix depends on messages arriving.
- Anything losing leads. Response time, unassigned records, and dead automations. These are active leaks.
- Anything slowing your team down. Required fields and notification noise. These cost hours and goodwill rather than deals.
- Anything hiding the truth. Stage definitions, duplicates, and reporting. Fix these and every future decision gets easier.
Then put the audit in the calendar. Systems drift: forms get renamed, people leave, someone builds a workflow that overlaps another. Ten minutes a quarter keeps a CRM healthy far more reliably than a rebuild does. And if your list came back long, that is usually one structural problem rather than ten separate ones. Start with the nine mistakes behind most of them.
Want Us to Run It For You?
Book a call and we will run this audit on your account, show you what is broken, and tell you what it takes to fix. Or hand it over entirely: our team builds and manages the pipelines, automations, and reporting for you.
Get Your Free Audit