90% of Sales Reps Had Split Performance Data. Nobody Knew It.

The Setup

The same $15M/year coaching company. After the rep leaderboard went live, we started auditing the underlying identity layer in more detail. What we found made a few of the CEO's past decisions suddenly make sense.

The Problem

The reporting looked clean on the surface. Every rep had numbers. The numbers seemed plausible. But when we pulled a random sample and walked each rep's data back through the systems, we kept finding the same thing: their performance was split.

A rep named Amanda showed up in the CRM as "Amanda K." and in the payment processor as "Amanda Kowalski." Another rep's phone changed mid-quarter and the new number created a second profile. A third rep was hired under a nickname, switched to her legal name three months in, and effectively started over in the data.

The data didn't lie. But it was telling two stories about every other rep.

What We Found

Of the 50 reps active in the reporting window, 45 had identity mismatches across systems. Ninety percent. Most were minor — a missing middle initial, a formatting difference. A handful were substantial enough to change performance rankings.

One rep's two profiles combined showed 142 leads worked and $17,602 in revenue. Split across the two systems, she looked like two different people: a setter with 142 leads and zero revenue, and a closer with one lead and $17,602. The ops lead had been using the second profile to argue this rep was top performing. The first profile, under a different manager's view, had her flagged for termination.

What We Built

A rep identity resolution layer that sits between every source system and every reporting view. It uses a deterministic match (email, phone, payroll ID) with a fuzzy fallback (normalized name, start-date proximity, funnel assignment). When a match is ambiguous, the layer flags it for a human — it never guesses silently.

Every new hire gets a canonical ID on day one. Every system is configured to write that ID alongside whatever internal ID it uses. The resolution layer becomes invisible the moment it's in place: from then on, every query and every report uses the unified identity.

The client's ops lead spent 10 minutes reviewing the flagged ambiguities. That was the entire setup cost.

The After

Two weeks after implementation, the CEO ran a cleanroom review of the top-20 and bottom-10 performers. Four reps moved out of the bottom tier into the middle once their two profiles were unified. Two reps previously in the top ten dropped out when the duplicate credit was removed. No performance decisions made in the six months before that review had been wrong — but three were about to be.

The Marketing Director at this client, on seeing her numbers for the first time on a clean identity layer, sent us a note: "Now that we have better insight into next-day appointments, I'm starting to work in 2x per day promos."

That decision — doubling promotion cadence on the funnels with strong next-day show rates — moved revenue on its own.

Closing

This kind of work isn't glamorous. It's the layer underneath the dashboard that the CEO never sees. But every real reporting system depends on it. Most teams we meet have never audited it.