Credibility is not earned by doing good. It is earned by being checkable. Four tests you can apply to your own CSR claims before you publish them.
The uncomfortable starting point
Credibility is not earned by doing good things. Plenty of companies do a great deal of good and get no credit for any of it. Credibility is earned by being checkable.
That distinction annoys people, and it should, because it feels like it rewards paperwork over substance. It does not. It rewards substance that somebody outside the company can confirm. The good work still has to happen. It just does not count for anything externally until it can be verified by someone who is not you.
Once you accept that, the difference between a claim that builds trust and one that reads as greenwashing becomes much easier to see. It is not about tone, and it is not about how much you spent.
The four things that make a claim credible
It is specific. A number, a place, a period, a scope. “We reduced water use” is not a claim. “We reduced water use at our two Dubai warehouses by an eighth against 2024” is a claim, because it can be wrong.
It is dated. A reader needs to know what period the figure covers and when it was measured. Undated claims float. They also age badly, because a genuine achievement from four years ago starts to look like a current claim, which is a real reputational risk even when nobody intended to mislead.
It is attributable to somebody outside the company. Somebody who does not report to your CEO put their name to some part of it. That could be a certification body, an assurance provider, a partner organisation, a public authority or a customer. Internal sign off is not attribution.
It is repeatable. You could publish the equivalent figure next year, and the year after, without dreading it.
Any claim that clears all four is defensible. Most published CSR claims clear one or two.
The tells
If you want to know how your communications read to a sceptical assessor, look for these. They are the patterns that trigger doubt, whether or not anything is actually wrong.
Round numbers with no baseline. “We planted 10,000 trees.” Against what target, over what period, and how many survived? Round numbers with no comparison point read as marketing, because that is usually where they come from.
Photographs instead of figures. A gallery of a volunteering day is evidence that a day happened. It is not evidence of a programme. When the photographs are doing the work that a figure should be doing, readers notice.
A claim that appeared once and never returned. This is the strongest single signal of a one off. If your 2023 report celebrates an initiative and the 2024 and 2025 reports never mention it again, the reader concludes it stopped. Usually the reader is right.
Language about intent rather than result. “Committed to”, “focused on”, “striving toward”, “on a journey to”. These describe posture. None of them describe anything that happened. A page full of intent language with no completed actions is the clearest greenwashing tell there is, and it is extremely common in otherwise honest reports.
The repeatability test
If you only apply one test, apply this one. It is the sharpest, and it is free.
Take the claim. Ask whether you would be comfortable publishing the same measure, on the same basis, every year for the next five years.
If the answer is yes, publish it. The claim is proportionate to what you actually do, and it will build a record over time, which is where credibility genuinely comes from.
If the answer is no, the claim is already too strong. Not dishonest, necessarily. Just larger than the underlying activity can sustain. Something in it was a one off, or the good year flattered the figure, or the measure was chosen after the fact because it looked best.
The five year question catches almost everything. A claim you would not want to repeat is a claim you should not have made in the first place.
This test is worth applying before any award submission, any tender response and any published report, because all three create a record that people will read back to you later. Anything you claim can be checked against what you claim next year. That is not a hypothetical risk. Recognition schemes increasingly require winners to keep reporting against their submission for several years afterwards, which turns an overstated claim into a commitment you have to keep meeting.
Third party validation, in ascending order of weight
Not all external validation is worth the same. Roughly, from weakest to strongest:
- A partner or supplier confirming your account. Useful, but they have a commercial relationship with you.
- A customer or community organisation putting something in writing. Better, because they gain nothing by saying it.
- A professional body or certification scheme. Now you are being assessed against published criteria that you did not set.
- Independent assurance over your data. Somebody tested the numbers rather than the narrative.
- A public authority or award body that ran its own evaluation. The strongest, because the evaluation was designed by somebody with no stake in your result.
The pattern is distance. The further the validator sits from your payroll, and the more their own reputation is exposed by endorsing you, the more the validation is worth. Ranking your own evidence against this scale is covered in more depth in how to prove your CSR actually happened.
The conclusion most companies get backwards
Now the uncomfortable part. The safest communications are the specific ones, and most companies do the opposite.
The instinct is understandable. A specific figure can be challenged. A vague statement cannot, so it feels lower risk. In practice it works the other way round. A vague statement carries no evidence, so a sceptical reader discounts it entirely, and you have spent credibility to buy nothing. A specific figure carries its own defence. Even if somebody queries it, you have a basis to answer.
The companies that get into genuine trouble are almost never the ones publishing detailed, checkable numbers. They are the ones publishing warm language over a programme nobody documented, who then get asked for proof.
What to do about it
Take your last published CSR communication, whether a report, a website page or an award submission, and mark every claim in it against the four tests. Specific, dated, externally attributable, repeatable. Count how many clear all four.
For most companies the answer is a small fraction, and the fix is not to claim less. It is to go and get the evidence that was always there and was never collected at the time. That is nearly always possible for the last year or two, and nearly impossible beyond that, which is the argument for fixing it now rather than at the next submission deadline.
How ESGweise helps
We run claims and communications reviews. That means taking what you have published or are about to publish, testing each claim against the four criteria, and telling you which ones will not survive contact with an assessor. Where a claim fails, we identify what evidence would rescue it and whether that evidence can still be obtained.
See our sustainability strategy and assurance readiness services, and our overview of social audit standards for how external verification works in practice.
If you have a submission or a report going out soon, review it before it goes, not after. Speak with our team.
Conclusion
Doing good does not earn credibility. Being checkable does. A claim earns trust when it is specific, dated, externally attributable and repeatable, and the five year repeatability question catches most of the ones that are not. The counterintuitive part is that precision is safer than vagueness, because a specific claim comes with its own defence and a vague one comes with none.
Frequently asked questions
What makes a CSR claim credible?
Four things. It is specific, meaning it contains a number, a place and a scope rather than an adjective. It is dated, so a reader knows what period it covers. It is attributable to someone outside the company, whether an auditor, a partner organisation, a certification body or a public authority. And it is repeatable, meaning you could publish the equivalent figure next year without embarrassment.
What is the difference between greenwashing and a badly written CSR claim?
Intent, mostly, and from the outside the two are indistinguishable. That is the practical problem. A reader assessing your communications cannot see whether a vague claim is deceptive or just carelessly drafted, so both are discounted the same way. This is why precision matters even for companies acting entirely in good faith.
Is it safer to keep CSR claims general?
No, and this is the most common misjudgement. General claims feel safer because there is no number to be caught out on. But they also carry no evidence, so they read as marketing and are discounted or challenged. Specific claims are defensible precisely because they can be checked, which is what makes them worth publishing.
What counts as third party validation?
In ascending order of weight: a supplier or partner confirming your account, a customer or community organisation putting it in writing, a professional body or certification scheme assessing you against published criteria, an independent assurance provider testing your data, and a public authority or award body recognising you after its own evaluation. The further the validator sits from your payroll, the more the validation is worth.