01
02
03
04
05Sections
Fronterra develops and operates forest carbon and biodiversity projects across Peru, and when it came to us it had credibility to spare. It had four projects spanning more than 1.6 million hectares across the Amazon and Andes, long-term concessions, projects listed on Verra and substantial prior experience in carbon-credit transactions. Most of that proof was buried. Project information was hard to find, and the audience had to assemble the case themselves. That was the problem we were asked to solve, and our Fronterra case study opens on it. Nothing needed inventing. The evidence a buyer or investor would ask for already existed, and it sat where those readers were unlikely to look while they were deciding whether to keep reading.
We called the rebuild "Proof before promise." The problem behind that phrase is one instructional designers have studied for decades. In a 1992 study, the educational psychologists Chandler and Sweller found that learning suffers when related information has to be mentally integrated across separate places. A buyer reading a website is a different reader from a student in a classroom, so the commercial application is an analogy. The design principle carries over intact. Keep the evidence that determines whether a claim is believable near the claim itself.
Proof loses force when the reader has to assemble it
Technical organisations often build material in two halves. The story comes first and the evidence arrives later as charts, studies, certifications, pilots and references. The reader meets a performance claim on page three, a commercial claim on page six and the proof for both in an appendix or data room they have not opened yet. They then have to remember each earlier claim and match it to the right proof.
The opposite error puts the evidence first. Numbers appear before the reader knows which question they answer, and a page of figures without a claim to attach them to is as hard to use as a claim with no figures. Good evidence architecture gives the reader a reason to care about the proof before asking them to inspect it, then shows it at once.
Place proof at the first reasonable doubt
The strongest location for a proof point is usually the first moment a reasonable reader would ask how the company knows. A product page claiming performance under a demanding condition should show the test, data or certification there. A deck claiming accelerating demand should place the commercial signal on the same page. A company claiming biological information can be read from the air should demonstrate that capability before moving to outcomes further downstream. The result is a rhythm of claim, proof and implication, and the reader never has to suspend judgement for several pages while waiting to learn whether the headline is supported.
A claim register finds the places where that rhythm breaks. Our piece on evidence-based storytelling sets out the full register, and our essay Why Good Evidence Still Fails to Move a Decision covers the cases where placement is not the problem at all. The two fields that matter here are where the proof first appears and where deeper verification lives. Set side by side for every consequential claim, they expose two different problems. Some claims have adequate evidence placed too late, which is a layout problem the team can fix. Others have no adequate evidence at all, and that belongs back in the business, research or commercial process, because copy can narrow a claim and cannot manufacture the proof it lacks. Brighter Future will not lift a claim into a headline before its proof can sit beside it. A qualification that changes what a claim means sits in the same sentence or panel as the claim.
Use labels that preserve commercial meaning
Different evidence supports different decisions. A laboratory result can support a technical claim without demonstrating demand, a paid pilot shows customer engagement without proving repeat purchase, and a letter of intent records a form of buyer intent and does not create recognised revenue. For operating and investor material, five useful commercial labels are actual, contracted, forecast, target and pipeline. Actual has happened and been recorded. Contracted has been signed and may still await delivery or recognition. Forecast records management's current expectation, target records the outcome the organisation wants, and pipeline records a possible opportunity at a stated stage.
Technical claims need equivalent distinctions. Measured and modelled, observed and attributed, pilot and commercial deployment, independent verification and company-reported result all tell the reader what kind of confidence fits. A label placed beside a number often does more work than another paragraph, because it stops the strongest-looking figure on the page from quietly acquiring a stronger status than it has earned.
Give deeper readers a route without forcing everyone through it
Most readers do not need the full method at first contact. A useful evidence system has three depths, and Fronterra's rebuilt site shows all three carrying one claim. At the first depth the claim is the portfolio itself, four projects across more than 1.6 million hectares, the figure a reader should meet before anything else. At the second depth each project has its own page, and as we set it out at the time, each one had to answer the same questions quickly. The six questions are where the project is, what is being developed, what stage it has reached, what methodology sits behind it, who is involved locally and what scale is being built. Those pages give each audience a consistent route through status, methodology, numbers and operating context.
The third depth is diligence, where the Verra listings and the long-term concessions can be checked by a reader whose decision requires it. One rule held the three together. We aligned figures and claims across the website and the deck, so the argument holds from first impression into diligence. A buyer can judge relevance from the first depth, a partner can test a single project at the second, and an investor's analyst can verify the third without finding a different number at each stop.
When a team's proof sits pages away from its claims, our science communication work moves each proof point to the first reasonable doubt, from the homepage through to diligence.
Investor material needs evidence to become more specific as the stakes rise
An investment case usually begins with the market change and proposition, then moves into technical proof, commercial proof, operating evidence, team credibility and the milestone the next capital is intended to unlock. Each result should sit against the investor uncertainty it retires.
Koralo, a company that co-ferments microalgae and mycelium, had five kinds of evidence, market, commercial, scientific, technical and operational. Putting all of it in front of an investor at the first meeting would have buried the argument, so we sequenced it and the proof got more specific as the claims did. Our Koralo case study records that call. The story those proofs supported had itself changed when Koralo's opportunity moved, which our piece on repositioning after a pivot follows.
Controlled access should preserve the claim
Some evidence belongs behind diligence. Confidential customer names, unpublished scientific results, commercially sensitive performance data and personal information may require controlled access. Public material stays credible when it states the evidence class accurately and gives the appropriate reader a route to verify it. We applied the same rule when writing up Koralo. Its scientific and nutritional material stays confidential, and our case study shows how that material was organised and leaves out what it says. A line such as "independent test results available under NDA" does the same job on a company's own site.
Silence creates ambiguity about whether the evidence exists. Overstatement creates a worse problem once the reader reaches diligence. In our interview with Nick Lyth, published in May 2024, the founder of the UK climate angel syndicate Green Angel Ventures lists "exaggerated claims not holding up to due diligence" among the traits that lead his syndicate to reject an investment. He is describing one angel syndicate's screen, and the trait he names is the cost of a claim that ran ahead of its proof. Our guide to presenting first-of-a-kind (FOAK) risk and our piece on what 16 climate investor interviews teach founders carry investors' own words on one-sided messaging and on what they expect to open after a term sheet.
The working question is the same on every page. Decide what the reader needs to believe at that point, then place the strongest evidence you can defend beside that belief. The argument usually gets shorter, because the reader no longer carries unsupported claims forward to resolve later. Fronterra is the plainest case of it. The projects, the hectares, the Verra listings and the concessions all existed before the work began, and the work was to move them to the point where each reader's doubt arrives.
Sources and further reading
Tell us what needs to move.
Bring the brief if it is clear. If it is unclear, tell us where the work is stuck.
Bring us the problem