What Entrepreneurs Can Learn From How Scientists Validate
Picture a founder who tells her investors that customers love the new product. Ask how she knows, and you’ll usually hear about a few warm emails, a good meeting, maybe a spike in sign-ups the week of the launch. None of that is nothing. But very little of it would survive a lab meeting.
Scientists have spent a few centuries working out how to avoid fooling themselves, and the habits they landed on travel surprisingly well. You don’t need a lab coat to use them. You need to keep asking yourself, “How do I know this is true?”
A result isn’t real until it can be repeated
In research, a finding that worked only once is treated as a lead, not a fact. Someone else should be able to follow the same steps and get the same results.
When that doesn’t happen, the cost is high. A 2015 analysis in PLOS Biology estimated that around $28 billion a year goes to U.S. preclinical research that can’t be reproduced, and it put the overall irreproducibility rate above 50 percent. Researchers still argue about the exact figure, and not all of that money is necessarily wasted. But the authors’ biggest single cause, at roughly 36 percent, was problems with biological reagents and reference materials. In other words, the basic ingredients.
If a marketing channel brings in ten good leads in March but none in April, you might ask, “Was March just a good month, or does this channel actually bring in leads consistently?” Before spending more money on it, give it another chance and see if the results hold up.
Check the tools before you trust the measurement
Researchers often use lab research tools to detect a particular protein in a sample. If an antibody grabs the wrong protein as well, or doesn’t behave the way a study assumes, the picture looks convincing and means nothing. Just because an antibody is labelled for a particular target doesn’t mean it will detect it correctly. Researchers need to test it first. Entrepreneurs rely on instruments too, they just don’t call them that. A survey question can nudge people toward a certain answer.
Have you ever thought about what happens when your tracking tag counts the same visit twice? When this happens, you start trusting numbers that don’t tell the full story, and your decisions can go wrong because of it.
Before trusting a number, take a minute to check where it comes from and how it’s counted. It may seem like a small thing, but catching a mistake early can save you from bigger problems when it’s time to report results or make business decisions.
Always have something to compare against
Controls are the least glamorous part of an experiment and probably the most important. A control tells you what would have happened without the thing you’re testing. Without one, you can’t separate the effect from everything else going on.
Say you launch a new onboarding email and retention climbs 8 percent. Great. But was it the email, the new pricing page that shipped the same week, or just the time of year? Hold back a small group who get the old version, and you’ll know. Small businesses often skip this because they feel they’re missing out on potential sales.
Use methods that produce numbers, not impressions
Scientists also prefer measurements to feelings. When a lab wants to know how much of a specific protein is in a sample, it doesn’t eyeball it. It might use an ELISA test, which uses antibodies and a standard curve to measure how much of a target is present in a sample. Two people running it properly should get close to the same result, which is the whole point.
Write down what you did
Scientists write down what they do during an experiment so they don't have to rely on memory later. They write it all down so that when they come back to the experiment later, they know exactly what they did and what happened. In business, people often run a campaign, check how it did, and then move on. A few months later, they’ve forgotten the details. Just jot down what you tried, when you did it, and what happened. That way, you won’t have to guess what you did last time when you want to try it again.
Go looking for reasons you’re wrong
The habit that separates good scientists from the rest is that they try to break their own results. Peer review is built on the idea that colleagues will poke at your work, so you poke first.
For founders, that can mean asking the most skeptical person on the team to argue against the plan, or setting a number in advance that would change your mind. “If trial conversion stays under 3 percent after 500 visits, we rework the offer.” Deciding that before you see the data keeps you honest after.
Start small
You don’t need a research budget. Pick one decision this month, the one you feel surest about, and run it through the list. How is it measured? Has the measurement been checked? What’s the comparison? Could someone else repeat it?
You might find that you were right all along, or you might realise you never actually checked. Either way, you’ll know more than you did before.
