What Proof Can You Show Before You Have Case Studies?

What proof can you show before you have case studies?

Four things, all of which you already have. A demonstration of the thing working, an artefact only someone who has done the work could produce, a constraint you are willing to name, and a person willing to attach their name to the claim. None of those requires a customer, and all of them beat a logo wall.

The mistake early stage companies make is treating proof as prestige. They go looking for something impressive to display, find nothing, and conclude they have a proof problem that only time can solve. What they actually have is a doubt they have not identified.

Proof is not a display. It is the specific thing that answers the specific reason a buyer is hesitating. Once you frame it that way, a company with zero customers has plenty to work with.

Why does a logo wall fail hardest at this stage?

Because you do not have one, and the substitutes are worse than nothing. Investor logos, accelerator badges, and the tools you integrate with are all attempts to borrow credibility from someone whose credibility was earned doing something else entirely.

A buyer evaluating whether your product will work for them learns nothing from knowing who funded you. Worse, a page that leads with borrowed marks signals that the page has nothing of its own, which is the exact inference you were trying to prevent. I have written before about why customer logos are not proof, and the argument is even stronger when the logos are not customers.

The one exception I would make is integration logos placed where they answer a real question, which is compatibility rather than credibility. Show them next to the integrations section, not in the hero, and you have turned a prestige play into useful information.

What counts as proof when you have no customers to cite?

Start with the demonstration. A recording of the product doing the hard part, unedited, with real data rather than a polished happy path, is stronger than most testimonials. It is falsifiable, which is exactly what makes it persuasive. A buyer can watch it and judge for themselves rather than taking your word.

The second category is the artefact. A sample output, a real report, a template, a teardown of a public example. Anything a person could only produce by genuinely having done this work. When I want to show what I do, the most convincing thing is not a claim about capability, it is an actual analysis of a real page with real findings.

The third is the named constraint. Saying plainly who this is not for, or what it does not do yet, buys more trust than any endorsement, because it is costly to say and nobody fakes a limitation. A buyer who reads an honest constraint immediately upgrades their estimate of everything else on the page.

Can you use your own team as endorsers?

Yes, if you disclose the connection, and the rule here is not merely etiquette. The US Federal Trade Commission's endorsement guidance states that if an ad features an endorser who is a relative or employee of the marketer, the ad is misleading unless the connection is made clear.

So an employee talking about why they built something is legitimate content, provided nobody could mistake them for an independent customer. Put their role next to their name. That single line converts a potentially deceptive endorsement into a credible founder story, which at this stage is one of your stronger assets anyway.

If you sell outside the United States, check the rules that apply where you advertise, because I am describing one regulator's guidance and not a global standard. The underlying principle travels well regardless: undisclosed connections are the fastest way to turn proof into a liability.

What about a design partner or a pilot under NDA?

Use it anonymised, with the shape of the client rather than the name. A mid sized logistics company in Europe with a twelve person operations team tells a buyer far more about fit than a name they have never heard of would, and it is usually permitted where naming is not.

The trick is to be specific about everything except identity. Industry, size, the problem, what changed, how long it took. Vagueness is what makes anonymous proof feel evasive, and vagueness is a choice, not a requirement of the NDA. I have written about running an anonymous case study in a solo practice, and the method transfers directly to a startup with one design partner.

Ask for permission properly and early, in the agreement rather than as an afterthought. A clause covering anonymised reference use is easy to agree at signature and awkward to request six months later when you actually need it.

How do you show a result without overclaiming it?

Give the number its denominator and its context. The FTC's guidance is instructive even outside advertising law: where an advertiser does not have proof that an endorser's experience represents what people will generally achieve, the ad must make clear what the generally expected results of the same approach are.

Translated into plain practice, that means a single good outcome presented alone is a claim you probably cannot support. The same outcome presented as one of a stated number, with the conditions under which it happened, is honest and usually more persuasive. Buyers are not naive, and a number with its context reads as confidence rather than as a highlight reel.

When I describe the automation I run for Ajust, I can say it has delivered more than 25,000 cases, helped over 400,000 people, and saved upward of 50,000 hours, because those are the actual figures for that system. What I cannot do is imply that every project produces those numbers, because most projects are not that system. Same discipline, applied to yourself.

What should you never do here?

Fabricate any of it. This should not need saying and yet it is the most common failure at the early stage, usually justified as a placeholder that will become true soon. The FTC is direct about the extreme case: where likes come from non existent people or people who have no experience using the product or service, they are clearly deceptive, and both the purchaser and the seller of the fake likes could face enforcement action.

The same logic covers the softer versions everyone tells themselves are fine. A testimonial written by the founder in a customer's voice. A star rating with no underlying reviews. A dashboard screenshot with invented numbers. A case study for a pilot that has not concluded.

Beyond the legal exposure, the practical risk is that this is easy to discover and impossible to walk back. A buyer who catches one fabricated element correctly discounts everything else, and you will never know which deals you lost to it because nobody tells you.

When does this stop being a problem?

Earlier than you think, and not in the way you expect. The transition is not from no proof to case studies. It is from generic proof to specific proof, and the trigger is your first customer who will describe a concrete before and after in their own words.

One of those is worth more than five polished case studies produced later, because it is early, specific, and obviously not manufactured. So the work in the meantime is not to wait. It is to set up the relationship so that conversation happens naturally, which mostly means asking about their situation before you started and writing the answer down.

By the time you do have a library, the discipline you learned here still applies. The best case studies are the ones that read like the demonstration and the named constraint, rather than like marketing, and I have covered what makes a case study actually convert separately.

What should you do next?

Write down the single sentence a buyer says to themselves when they hesitate on your site. Not the objection they voice on a call, the private one. Then ask which of the four categories answers it: a demonstration, an artefact, a named constraint, or a named person.

Build that one thing this week and put it next to the claim it answers rather than in a section of its own. One piece of proof aimed at the real doubt outperforms a page of prestige aimed at nothing in particular.

If you are early and not sure which doubt is actually stopping people, reach out. Working out what a page is failing to answer is usually a quick conversation and it changes what you build next.