What Assessors Actually Look For in an Excellence Submission

There is a common belief that an excellence submission is a writing competition: whoever writes most beautifully wins. It isn’t true. I have seen beautifully written submissions score modestly, and plainly written ones outscore them by a clear margin.

The difference was never the language. It was the structure of the evidence.

In this article I explain how an assessor actually reads — not as we imagine them, but as they work — what earns points and what removes them, and how the material is built before anything is written.

First: how does an assessor think?

Before discussing criteria, it helps to understand the assessor’s own position. They are not an audience to be emotionally persuaded, nor an adversary hunting for errors.

An assessor is someone who must award a score and then justify it to a panel. That second constraint governs everything they do.

This is why every sentence without evidence puts them in a difficult position: they may believe it personally, but they cannot defend a score built on it. The natural outcome is that they mark down — not because they were unconvinced, but because they have nothing to support their conviction.

Once you understand this, the objective of writing changes. You are not writing to impress an assessor; you are writing to make it easy for them to award and defend a score. That simple reframe changes an entire submission.

An assessor reads for four things

1. Is there an approach, or a series of scattered decisions?

The first thing tested is not the result but whether a deliberate approach sits behind it. An organisation that explains how it decided — on what basis, with which data, and who was involved — reads as mature. One that describes only the outcome reads as fortunate.

The practical difference in writing:

Weak: “We launched a leadership development programme.”

Strong: “Succession data analysis showed 42% of leadership positions would become vacant within five years, while only 19% of internal candidates were assessed as ready. The leadership development programme was designed to target that specific gap.”

The second sentence does not describe an activity; it describes a reason. And the reason is what proves an approach exists.

2. Was the approach actually deployed, and consistently?

A written policy is not a deployed one. Assessors look for traces of deployment: how many units were covered, what proportion, over what period, and what happened in the cases that did not go to plan.

The most neglected element here is coverage rate. “We trained 400 employees” is an absolute number that means little alone. “400 of 430 employees in the relevant units — 93%” means a great deal.

The difference is not cosmetic. The percentage tells the assessor deployment was comprehensive rather than selective, and that the organisation knows the size of its own target population in the first place.

3. Are results measured and compared?

This is where most submissions fall. An isolated result does not convince. A result shown with its trend across three years, and benchmarked against a comparable organisation or sector standard, moves from claim to evidence.

Four types of comparison, from weakest to strongest:

• Against target: “we achieved 91% against an 85% target”. Acceptable, but it relies on a target you set yourself.
• Against yourself over time: “rose from 78% to 91% across three years”. Stronger, because it shows direction.
• Against a published sector standard: “against a sector average of 83%”. Stronger still, because the source is external.
• Against a named reference entity: “against 8 minutes at a reference entity in the same sector”. Strongest, because it is concrete and verifiable.

You do not need the fourth type for every indicator. Having it for two or three headline measures lifts the whole submission.

4. Does the organisation learn and adjust?

This is the most overlooked. Submissions presenting unbroken success with no challenge and no course correction read as unconvincing. An organisation that shows what did not work, and how it adapted, earns more trust — not less.

Note that the criterion is not “did you make mistakes?” but “do you have a system that detects and corrects them?” Those are entirely different questions.

An organisation writing “the quarterly review showed a deviation, we analysed it, we adjusted, the indicator improved” is not admitting weakness — it is proving a working feedback loop.

Enablers versus results

Most excellence frameworks split assessment into two halves: enablers (how the organisation works) and results (what it achieved). The most common structural error I see is conflating them.

Enablers answer: how do you set strategy? How do you manage resources? How do you design services?

Results answer: what was achieved? In which direction? Compared to what?

The frequent mistake is filling the results section with activity descriptions (“we implemented…”), or filling the enablers section with result figures and no explanation of method. Either way the submission loses points in both halves.

The simple rule: in enablers, the core construction is “we do X because Y”. In results, it is “the number, its trend, its comparison”.

How to build an indicator worth presenting

Not every number is an indicator. One that survives assessment has five properties.

1. A precise definition. What exactly does it measure? “Completion time” starts when and ends when? Does it include customer waiting time or only processing? An undefined indicator admits multiple interpretations, and an assessor will assume the narrowest.

2. A specified source. Which system produces it, and who can reproduce it? An indicator that cannot be regenerated is not evidence.

3. A time series. One data point is not enough. At least three show direction and rule out coincidence.

4. A comparison. As above.

5. A clear link to the initiative. If the indicator improved in the same period as another change, the assessor will ask which caused it — and you should already have answered.

Isolating the variable: the technique that separates strong submissions

This is an advanced point but it makes a large difference. When you claim your initiative improved an indicator, an alternative explanation always exists: perhaps everything in the organisation improved that year.

The practical way to rule that out is presenting a control indicator: “satisfaction for this service rose from 78% to 91%, while the entity’s overall index remained flat at 85% over the same period.”

That single sentence moves the claim from correlation toward causation. It requires no complex statistics — only that you thought about the question before the assessor asked it.

How to choose which initiatives to submit

A question that precedes writing and determines its outcome: which initiatives go into the submission?

The natural instinct is to choose the largest, the newest, or the one closest to leadership attention. That is usually the wrong choice.

The right criteria are three.

1. Availability of evidence. A mid-sized initiative with complete data always beats a large one with gaps. The assessor does not know the scale of your ambition; they know what you prove.

2. A completed cycle. An initiative launched three months ago does not qualify, however promising. Impact needs time to appear and be measured. The ideal candidate is one to three years old: long enough to show results, recent enough to remain relevant.

3. A present owner. If the person who led the initiative has left the organisation, defending its details in a session will be difficult. This practical consideration is frequently ignored.

And a final rule: three initiatives told in depth beat ten mentioned in a line. Space is limited, and depth is what gets assessed.

Writing the “approach” section so it earns its score

This is the most poorly written section, because teams assume the task is to describe what they did. The actual task is to prove that what they did was deliberate and grounded.

The structure that works has five sequential elements.

The trigger: what made this a problem worth intervening in? Preferably a number rather than an impression.

The analysis: how did you identify the root cause? Which data did you use? This element is what distinguishes a methodical organisation from a merely busy one.

The options: which alternatives were considered, and why was this path chosen? Naming a rejected option and the reason for rejecting it noticeably raises credibility.

The design: how was the initiative built, and who participated in designing it?

The monitoring mechanism: how would you know it was working — and what would have made you stop it?

The last element is the most powerful and the most rarely included. An organisation that defines a stop condition in advance proves it manages by evidence rather than by emotional commitment.

Benchmarking when you believe you have no comparators

“We are a unique case with nobody to compare ourselves to” is among the most repeated sentences, and it is usually untrue.

Five comparison sources available to almost any organisation:

1. Internal comparison between units. If you have ten branches, you have ten comparison points. The highest performer is your internal benchmark, and the gap between it and the average is a story in itself.

2. Comparison over time. You against yourself three years ago. The simplest comparison and the most neglected.

3. Published standards. Federal reports, national customer satisfaction surveys, sector studies. Much of it is freely available and goes unused.

4. Functional rather than sectoral comparison. You are not obliged to compare against a similar entity. If you measure contact centre response time, contact centre benchmarks apply to you regardless of sector.

5. The approved target. The weakest of the five but better than nothing — provided the target was documented in advance, not set retrospectively.

The absence of comparison is a choice, not a fate.

The supporting evidence file: the part nobody reads that changes everything

Most frameworks allow supporting evidence attachments. Many organisations either ignore this entirely or attach a hundred random files.

Both extremes hurt.

Good attachment is selective and organised: one clear piece of evidence per headline claim, named in a way that ties it to its location in the submission — for example “3-2 System report: completion time 2024”.

The benefit is not that the assessor will read everything; usually they will not. The benefit is that when they doubt one number, they find the proof in seconds. That single experience raises their confidence in the rest of the document.

The reverse is equally true: one claim they looked for and could not verify makes them question everything else.

The site visit: where alignment is tested

In many frameworks the submission is followed by a site visit. Its core purpose is simple: verifying that what was written matches what actually happens.

Three observations from visits I have attended:

The frontline employee is the most important source. The assessor will ask someone who had no part in preparing the submission about the same process. If their answer differs from what was written, the entire document loses credibility at once.

Systems get opened in front of you. You may be asked to display the screen from which a number was extracted. Confirm in advance that the figure is directly extractable, not manually calculated in a side spreadsheet.

Documents must be dated. A policy without an approval date, or minutes without signatures, reads as a document prepared for the visit.

Real preparation is not tidying the premises. It is ensuring that what you wrote is what happens. And where a difference exists, it is far better to state it in the submission than to have it discovered.

Why does a “good” submission score in the middle?

This puzzles teams most: the submission is organised, the initiatives are real, the language is clean — and the score is average.

The reason is that assessment systems do not reward the absence of errors. They reward maturity level. Take the same initiative and see how it reads at four levels.

Level one — description. “We have a leadership development programme.” The assessor records that an activity exists. The score is low because nothing proves it was designed or effective.

Level two — approach. “The programme was designed on a succession gap analysis showing only 19% of candidates were ready.” Now there is a basis. The score rises.

Level three — proven deployment. “Deployed to 93% of the target population across three cohorts over two years, with individual tracking for each participant.” Now there is evidence of rollout rather than a pilot.

Level four — compared results and learning. “Readiness rose from 19% to 58% against a sector average of 41%. The first cohort revealed weakness in one dimension, so the associated training module was redesigned before the second cohort.”

Notice the initiative never changed. What changed is the depth of what we proved about it.

Most “good” submissions stop at level two. The distance between them and advanced submissions is not more writing effort — it is data collected at the time.

The questions assessors actually ask

From reviewing multiple sessions, these recur in different wordings. Prepare for them specifically:

• “Where did this number come from?” — by far the most common. The expected answer: system name, period, and who extracts it.
• “What would have happened without the intervention?” — tests whether you understand causation or only correlation.
• “How do you know the improvement came from your initiative?” — where the control indicator earns its place.
• “Was it applied to everyone or to a sample?” — tests deployment.
• “What did not work?” — a direct question in many frameworks. No convincing answer reads as concealment.
• “What will you do next year?” — tests future readiness, and the answer should build on what you learned rather than on general ambition.
• “Who owns this process after the project ends?” — tests sustainability. “The project team” is a weak answer.

Notice that most do not ask about the achievement, but about your knowledge of your achievement. That is exactly what professional assessment distinguishes.

Ten recurring failure patterns

1. The activity submission. A long list of what the organisation did, with no measured impact.
2. Isolated numbers. Impressive percentages with no baseline and no comparison.
3. The absolute claim. “First of its kind” with no defined scope or source.
4. Unbroken success. A submission with no challenge anywhere — read as not candid.
5. Competing versions. Each department describing the achievement differently.
6. Breadth over depth. Twelve initiatives in one line each, instead of three documented.
7. The enabler–result gap. Excellent methodology described with no result proving it worked.
8. Unstable terminology. Two or three names for the same indicator with near-but-not-identical figures.
9. Decorative language. Opening paragraphs carrying no information, consuming limited space.
10. The translated submission. An Arabic version literally translated from English or vice versa, with terms that do not work in the target language.

The self-assessment: reaching the assessor’s findings before they do

The best investment before submission is a rigorous self-assessment session. Not a language review — a genuine simulation of what the assessor will do.

The method I recommend takes a day and runs in three stages.

Stage one — blind scoring. Give the submission to three people inside the organisation who had no part in preparing it, and ask each to score every criterion with written justification. Explain nothing beforehand.

The value is not the score itself but the spread. If three readers range between 40% and 80% on the same criterion, the text is ambiguous — and a real assessor will lean toward the lower end.

Stage two — claim tracing. Pick ten numbers at random and ask for each to be evidenced within five minutes. Any number taking longer, or requiring a question to an absent person, is a number at risk.

This exercise reveals more than expected. In most sessions I have run, three or four numbers out of ten failed this simple test.

Stage three — the deletion test. Read through and mark every paragraph that could be removed without losing information. Then actually remove them.

The volume that can be cut is often surprising. The freed space goes into deepening the remaining initiatives — which is what genuinely raises the score.

A simple readiness indicator

A quick test I use: calculate the proportion of sentences containing a number, a comparison or a reason, against total sentences.

In weak submissions it sits below 20%. In advanced ones it exceeds 50%.

It is not a scientific measure, but it gives an honest signal within minutes: is your submission saying something, or describing something?

Written submission versus live presentation

Many treat the presentation as a summary of the submission. That misunderstanding costs points.

The written submission proves. The presentation explains and is tested.

A jury does not want to hear what they read. They want three things the text cannot show:

• Does the team genuinely own the decision? It shows in the first question outside the script: “why this path and not the alternative?”
• Are the numbers live or memorised? A presenter who turns back to the slide to read a figure loses trust even when the figure is correct.
• Does the organisation speak with one voice? When two presenters answer in slightly contradictory ways, a jury registers it immediately.

Preparing: the hard-questions drill

The best preparation is not presentation coaching but one session dedicated to difficult questions.

Gather the team and ask a colleague from outside the initiative to pose the ten harshest questions they can think of. Do not defend — only record. Every question without a documented answer is a gap still open, and you still have time.

One important note: “I don’t know, I will send you the exact figure” is a strong answer, not a weak one. Guessing is far worse. Juries remember a wrong guess for a long time and do not remember an honest acknowledgement.

Who should be in the room?

Team composition determines the ceiling of a submission’s quality more than anything else. The common error is building it from the excellence department alone.

An effective team has four types.

Someone who knows the work. The initiative owner. Without them you will write an approximate account that collapses at the first detailed question.

Someone who knows the data. A person from IT or analytics who can extract and verify figures. Their presence from day one saves weeks and prevents claims built on numbers that cannot be extracted.

Someone who knows the framework. A person who understands the language of the criteria and where points are awarded.

Someone who knows nothing. The underrated role. An intelligent person from outside the initiative whose only job is to ask “why?” and “what is the evidence?” at every sentence. They are the closest available model of an assessor.

The absence of the second type is the most common reason submissions stall midway. The absence of the fourth is the most common reason submissions look excellent internally and fail externally.

The feedback report: a neglected asset

After every cycle, organisations receive a feedback report setting out strengths and improvement opportunities. In my experience this is the most wasted institutional asset there is.

What usually happens: the report is read once, filed, and the next cycle starts from zero.

What should happen:

Convert every improvement opportunity into an item with an owner and a date. A feedback report is effectively a free action plan written by external experts.

Classify the observations into two types: those about performance (the work itself needs improving) and those about evidence (the work is good but did not show). The second type is cheaper and faster to fix, and usually accounts for half the observations.

Review it well before the next cycle — not in the final month, but at the start of the documentation cycle.

Organisations that improve steadily cycle after cycle do exactly this. It costs nothing but discipline.

Time and resources: a realistic estimate

A practical question always arises: how long does a good submission take?

The honest answer: the writing is less work than you expect, and the data is far more.

The typical effort distribution for a mature submission:

• 50% gathering and verifying data. Extraction, reconciliation, resolving contradictions between systems.
• 20% analysis and building comparisons.
• 20% writing and review.
• 10% assembly and supporting evidence.

Organisations that fail usually invert these: 70% on writing and 10% on data, producing an elegant text with no foundation.

This also explains why a submission cannot be accelerated by adding writers. The bottleneck is rarely writing speed.

Excellence work and daily operations

A common misunderstanding deserves addressing: that excellence work is an additional burden on top of the real work.

That is true only when it is managed badly — as a separate seasonal project.

Managed properly, what excellence frameworks ask for is exactly what good management needs anyway: knowing your baseline, measuring impact, documenting decisions, benchmarking, and learning from deviations.

An organisation doing these things because it wants to be well managed will find itself assessment-ready automatically. One doing them because an assessment is coming will find them a burden every time.

The difference is motive, and the long-term outcome is entirely different.

A pre-submission checklist

• Every headline number has a baseline and a known source.
• Every indicator is shown across at least three time points.
• At least one external comparison exists for headline measures.
• Every uniqueness claim has a defined scope and a source.
• There is a clear section on what did not work and how it was handled.
• Terminology is consistent throughout (use a one-page glossary).
• No decorative opening paragraph consuming space.
• Arabic and English versions match numerically, exactly.
• Someone uninvolved read it and marked every sentence they did not believe.
• Every initiative has an owner who can defend its figures verbally, without slides.

The most repeated mistake

The most common error I see is not a shortage of achievement. It is confusing output with impact.

“We trained 400 employees” is an output. “Transaction completion time fell by 31% after the training, affecting 60,000 customers annually” is impact.

The first describes what the organisation did. The second describes what changed in the world because of it. Assessors award points for the second.

And if you do not win?

A closing point, more practical than it sounds.

Many organisations treat an assessment result as a final verdict on their worth. That is inaccurate, and costly for teams.

The result measures something specific: your ability to evidence your performance against a particular framework at a particular moment. It does not measure the value of your work, the capability of your team, or your real impact on the people you serve.

This distinction matters for two reasons.

First: it prevents the false conclusion that “excellence is not for us” after a single result. A first cycle is usually a diagnosis, not a competition.

Second: it directs energy correctly. An organisation that reads the result as a diagnosis extracts a plan from it. One that reads it as a verdict extracts only discouragement.

Of the organisations I have worked with, the ones that advanced furthest were not those that won in their first cycle, but those that treated the first cycle as a baseline measurement — and built on it with discipline for three years.

Before you write a line

I always recommend one session before any writing begins: gather the data you genuinely hold, and put two questions against every achievement — what is the evidence? and compared to what?

Anything left unanswered by those two questions should not enter the submission in its current form. Anything that finds an answer almost writes itself.

And that is really the most important outcome of all this work: when the material is built well, the writing becomes the easy part. When the material is incomplete, no amount of writing can compensate — nor should it.