What is a stages of excellence assessment?

A Stages of Excellence assessment is a self-assessment of one area of a business, such as pricing, strategy or the operating model. For each practice it sets five observable sentences, one per stage, and the respondent picks the one closest to what happens.

The stages run from the weakest practice to the strongest. The pattern of answers shows where the business stands, and the report suggests improvements for its largest practice gaps. onestepright's checks are written for businesses of about A$50 million (US$35 million, £25 million) in revenue and up.

Strategy consulting firms have used the format since the early 2000s. onestepright's versions are built from more than 30 years of strategy consulting and over 1,000 engagements.

The five stages

Most established businesses land in the second or third stage on most dimensions. In our experience across hundreds of engagements, they also sit one stage left of where their leadership believed they were.

The top stage is not the right target for every practice in every business. Some practices only need to be sound, and the effort is better spent where a stage move changes the result.

Why observable sentences beat survey questions

A typical maturity survey asks you to rate your agreement with statements like "we have strong pricing governance". Nobody can verify that, and answers tend to drift upward.

A Stages of Excellence grid offers descriptions that can be checked. For example, one cell in the pricing grid reads: "Discounts are approved by whoever the salesperson asks." Anyone in the building can confirm whether that sentence is true.

The wording is deliberately unflattering. When the true answer and the comfortable answer differ, only the true one produces a useful result.

How an assessment works

1. Answer. Each of 20–30 practices has five stage descriptions. Pick the closest. About ten minutes.
2. Profile. Results roll up into a stage profile for each dimension. In pricing, for example, the dimensions are strategy and value, price setting, price getting, governance, and data and performance. Common score patterns map to named archetypes, each with a recommended first move.
3. Move. The output is a short list of suggested improvements. It takes the two or three practices that sit lowest, the more basic first. For each, it says what firms typically do to move one stage right. The respondent then chooses which to act on first.

Where the method comes from

The Stages of Excellence format descends from Philip Crosby's Quality Management Maturity Grid (1979) and the capability maturity models developed in the 1990s. Consulting practitioners adapted it for strategy and commercial topics through the 2000s. onestepright's grids are built from more than 30 years of strategy consulting and over 1,000 engagements. Their source is the proposals, diagnostics and delivered work of a strategy consulting practice.

How this differs from a maturity model

Almost every published "maturity model" keeps the five levels and drops the sentences. The respondent gets level names such as Initial, Managed, Defined and Optimising, and a 1–5 slider per topic. A slider invites people to flatter themselves. A Stages of Excellence grid keeps a plain, checkable sentence in every cell, so the accurate answer is easier to give. Three other differences matter in practice:

• Results stay with the respondent. Most free assessments are a funnel: a quiz, an email wall, then a sales call. Here the report belongs to the respondent. Nobody else sees it, and no conversation happens unless the respondent asks for one.
• Benchmarks say how many firms stand behind them. Comparisons are anonymised aggregates. They are released only when a cut holds enough comparable firms for the comparison to mean something. Each carries a band, such as 30+. The alternative is a consulting firm's private database, where nobody can see who they are compared against.
• The wording comes from real engagements, not a question bank. Each cell describes something a consultant has observed in a business. The grids are built from more than 30 years of strategy consulting and over 1,000 engagements.
• Each grid is tested on real cases before release. Real businesses are placed on a grid one practice at a time. Some come from past engagements, and some are listed companies read from their public record. Where the sentences do not separate them, the wording is rewritten. The fund deal check was tested on 105 public buyouts, in an exploratory, AI-assisted retrospective test. It raised a warning on 43 of the 51 that later failed, and on 14 of the 54 that succeeded.

How the fund deal check was tested

We tested the fund check on 105 public buyouts: 51 that later failed and 54 that succeeded.

AI agents with web search scored each deal, one agent per deal, on an earlier draft of the check's questions. We then ran their answers through the check's current rule. They were told to use only what was public on or before the day the deal was agreed. Where the record was silent, they left the question unscored.

A failure is a poor result for the buyer, judged from what happened after. A success is a good result for the buyer. One person chose both lists, knowing the outcomes. The first lists lean to well-known deals from 2005 to 2008.

About a third of the answers rest on the agents' own knowledge or on a source that could not be opened. That knowledge includes what happened later, so those answers are not a record of the day. Each one is marked in the evidence file. Two other agents re-scored twelve of the deals without seeing the first scores, and reached the same verdict on eleven.

The rule for a Stop verdict was settled before the last 36 deals were chosen, and their scorers were not told which had failed. On those 36 the check said Stop on 9 of the 18 failures and on none of the 18 successes. Across all 105 it said Stop on 26 of the 51 failures and on 1 of the 54 successes. The wider warning in the figures above was adjusted after all 105 were scored, so it was tuned on the deals it is reported on.

What this does not show. Every outcome was known before we scored it, even where the scorer was not told. We have not yet tested a private company, and most mid-market targets are private. A forward record starts with the first deals scored after launch.

Frequently asked questions

Is a self-assessment reliable? It measures self-observed practice, which is exactly what most leadership teams lack. It is not an audit. Practitioners who use the observable-sentence format see it reduce the inflation that makes ordinary surveys unreliable. That is an observation from practice. No controlled test has measured it.

How is it scored? Each practice maps to a stage from 1 to 5. Results are reported as stage names and patterns. The check gives no mark out of 100, because the point is direction.

What is a good result? There isn't one. The useful result is an honest one. A solid middle score across the board, with two rows at the bottom you didn't know about, beats an inflated high one.

Does a grid assume one way of running a business? Each grid says what it assumes. The group questions in the strategy check start by asking how the centre runs the businesses.

A holding company or a financial-control group is then asked questions that fit that model, about capital and about what the centre costs. Other groups answer questions on how the centre gives each business a role and moves capital between them.

How do benchmarks work? Responses are anonymised and aggregated. Comparisons are shown only once a segment has enough similar businesses in it. Each carries a band such as 30+. The band is rounded down, so it never claims more businesses than there are.