Most sales training fails for a reason nobody diagnoses correctly. Progression of Performance® is a four-stage behavioral-change model — Learn, Practice, Rehearse, Perform — built on a distinction cognitive science settled decades ago: knowing a skill and executing it under real conditions run on two different memory systems. Declarative memory holds the facts. Procedural memory holds the execution. Most sales training trains the first one and stops there. Progression of Performance® is a registered service mark, Reg. No. 7836888, and it is the framework The Tonkin Group uses to close that gap inside professional services firms.

Training That Produces Knowledge, Not Behavior

Every sales leader has watched this happen. A rep sits through training, recites the framework back perfectly in the room, answers every question correctly. Then the rep walks into a real client conversation and freezes. The instinct is to blame motivation, or decide the rep is a bad fit for sales. Neither is usually true. This is not a motivation problem. It is a training-design problem, and calling it anything else delays the actual fix.

Most training programs are built to produce a rep who can execute a skill in a controlled, zero-stakes drill. That is real progress. It is not the same skill as executing it against a client who is skeptical, distracted, or pushing back, with money and reputation actually on the line. Law firms, CPA practices, and IT consultancies hire technically excellent people who were never trained to sell in the first place. That gap gets wider under this model, not smaller.

The Four Stages

Stage 1: Learn

ChatGPT Image Sep 15, 2026, 04_49_35 PM (2) The rep receives explicit instruction: what the skill is, why it matters, what correct execution looks like. This stage is declarative. It answers “what do I need to know,” and nothing more. It should not pretend to answer anything else.

Stage 2: Practice

ChatGPT Image Sep 15, 2026, 04_49_35 PM (3) The skill gets isolated from everything else and drilled in a zero-stakes setting. A law firm partner practices a single discovery-call question, alone, with no client on the other end of it. This builds raw mechanical repetition, separate from the pressure of a real conversation. It is also where most training stops — mistaking fluency in the drill for readiness in the room.

Stage 3: Rehearse

ChatGPT Image Sep 15, 2026, 04_49_36 PM (4) Real stakes come back. A CPA rehearsing a fee conversation now does it against a person actually playing a skeptical prospect, pushing back, asking the hard question. The rep has to self-diagnose in the moment instead of recalling the right answer. This stage separates “can recite it” from “can do it.” Most programs skip it entirely, jumping straight from a scripted drill to a live client call.

Stage 4: Perform

ChatGPT Image Sep 15, 2026, 04_49_36 PM (5) The rep executes in the real world, with an actual client and actual consequences, and documents what happened afterward. Reflection is not optional here. Skip it, and Perform just becomes another rep of Practice with higher stakes attached. The learning stops compounding.

The Diagnostic Principle

Here is the part most training misses. When a rep fails at a given stage, the fix is never at that stage. Failure at Perform means Rehearse ended too early. Failure at Rehearse means Practice ended too early. The instinct is almost always to add more of whatever stage the failure showed up in: more live reps, more in-the-moment coaching. The actual gap sits one stage back. Diagnosing that correctly is most of the work. Once it’s diagnosed, the fix itself is usually the easy part.

Where This Comes From

Progression of Performance® is built on Squire’s (1992) work distinguishing declarative and procedural memory, and on Anderson and Krathwohl’s (2001) revised Bloom’s Taxonomy, which separates Remember and Understand (Learn) from Apply (Practice) from Analyze and Evaluate (Rehearse) from Create (Perform). Dr. Tom Tonkin’s own record backs the same thesis. Leading sales enablement inside Oracle’s business, he measured the trajectory of 746 reps over eighteen months and traced $50 million in revenue above forecast, most of it margin, to reps who trained past the drill stage instead of stopping there. Progression of Performance® formalized that pattern years later — first prescribed as Dynamic Questioning in 2015, then built out into the four-stage model it is today. His 2026 Journal of Selling study on sales enablement and KPI achievement, and his doctoral research on leadership authenticity, sit on the same foundation: real behavior change takes more than information transfer, and most organizations stop investing exactly where the return actually starts.

Where Your Team Sits Right Now

If your team can recite the framework and still freezes with a real client in the room, that is a diagnosable gap, not a motivation problem to manage around. It is almost always one stage back from where the failure shows up. A twenty-minute Revenue Systems Snapshot call is the fastest way to find out exactly where your own team’s training exits early. Book time directly on the calendar.

AI Information