Picture a classic meeting-room scene: the CEO opens a sleek sales funnel in Bitrix24 and freezes. At the "Meeting Scheduled" stage, 14 deals sit proudly, while at the next step—"Invoice Sent"—there are 31. It looks absurd: the company sent twice as many invoices as it scheduled meetings. The head of sales shrugs guiltily and opens a secret Excel spreadsheet where the numbers suddenly add up.
You are probably thinking: "The CRM glitched again!" But software code has nothing to do with it. The creators of the analytics service Маржай broke down in detail how standard dashboards mislead businesses. The issue lies in a methodological mix-up: a visual kanban board answers "where deals are sitting right now," whereas managing a business requires analyzing a continuous flow.
Let's examine the 4 key traps that turn management analytics into a funhouse mirror.
The Snapshot Trap and Stage Skipping
The first systemic failure is trying to measure conversion from a snapshot. Kanban shows the balance of deals at the current second. Suppose 158 clients passed through "Meeting Scheduled" during the month, but by the end of the period most moved forward, leaving 14 unhurried buyers stuck at that stage. If you calculate conversion from the remaining balance, you will see a grotesquely understated step efficiency.
The right approach is to measure end-to-end flow from the change-history log. The experts cite a telling example of a window-installation company (8 reps, 441 inquiries in a month):
- Taken into work: 370 deals (83.9%);
- Meetings scheduled: 158 (42.7%);
- Meetings held: 120 (75.9%);
- Invoices sent: 82 (68.3%);
- Paid: 57 (69.5%). Once we measure flow instead of a slice, the main bottleneck stands out immediately: the transition from taking a lead into work to scheduling a meeting converts at just 42.7%. That is where the company loses most of its revenue.
The second trap is stage skipping. A rep calls a repeat client, quickly issues an invoice, and drags the card straight to "Paid," bypassing intermediate statuses. If the system does not normalize the path, conversion at skipped steps drops while subsequent ones shoot past 100%. There is only one fix: any deal at stage N must automatically be counted as having passed all stages from 1 to N-1.
Duplicate Deals and the Graveyard of Lost Leads
The third problem is mixing cold leads with repeat sales and duplicates. A client submits a form on the website, then writes in chat, and then calls by phone. If the CRM fails to merge contacts by phone number, the rep takes one deal into work and marks the other two as lost. Marketing conversion artificially drops threefold. Just as dangerous is mixing first-time leads with upsells to loyal customers: instant payments from returning buyers mask problems in primary acquisition. The solution is automatic duplicate merging and moving repeat customers into a separate, isolated pipeline.
The fourth trap is bogus loss reasons. In most companies, the "Loss Reason" field is either optional or stuffed with a catch-all "Other" option. As a result, up to a third of the database goes into the trash without a clear explanation. Proper discipline requires a mandatory choice from 6–8 clear reasons divided into client-side ("Chose a competitor," "Too expensive") and internal failures ("Could not reach"). In Маржай's analysis, the "Could not reach" category eats up to 30.6% of all incoming inquiries—a direct signal to tighten the SLA on first callback time.
If you let a funnel run without systemic control, the CRM turns into a warehouse of forgotten contacts. In our piece on auditing stalled deals in a CRM, we broke down exactly how much lost revenue can be recovered simply by cleaning out false statuses. Manage the flow, not static cards on a screen.

