Do you know what it takes to achieve exponential startup growth like Facebook? For reference, the chart for user growth at Facebook during the first few years is below. This is the classic “up and to the right” chart that we all want. So how do you achieve this kind of exponential growth at your startup? We need to learn to use the scientific method in business.
To answer that question, we first have to go back in history to the discovery of penicillin and the Theory of Relativity. What does these two scientific breakthroughs have in common with Facebook? They all came about from using a proven process or system that is hundreds of years old.
This process is known as the scientific method and you probably have heard about it. Yet very few startups are actually using it.
You could use the scientific method to test your assumptions around your product, business model, customer segment, marketing channels and much more. This is a much better alternative than blindly trying the latest growth hack hoping that it finally helps you achieve traction.
This post isn’t about tactics. You can find the latest tactics through a Google search. This post is about figuring a system that you can use to test different tactics until you find what works for your startup.
How to Use the Scientific Method in Business?
The scientific method breaks down into a series of steps that you repeat over and over again as you try to test a specific assumption.
Step 1: Brainstorm a Hypothesis
The first step is to come up with a hypothesis that we can test. For startups, these hypotheses can look at things like marketing channels, customers segments or anything else that you’re unsure of.
Some examples may be:
- Are young females, aged 20-25, our early adopters?
- Can Facebook Ads (after optimization) provide us with a better conversion rate?
- Do my users want X feature?
Step 2: Develop Predictions
The next step is to take the most important hypothesis and develop a prediction of the outcome we are expecting. Let’s take the hypothesis regarding early adopters as an example. We’ll predict that if this true, then we should see the following results:
- These users won’t mind that our product is unfinished and buggy
- These users will be willing to pay a monthly fee to use our product
Step 3: Test Hypothesis
We can now go and test this hypothesis (young females as early adopters) through experiments. These are some examples of experiment:
- Interview 20 users through customer interviews
- Drive traffic to a landing page asking for a free trial
- Host a workshop
You want your experiments to be as clear as possible and it helps to have small goals e.g. interview 20 people. You then collect as much data as you need for each experiment.
Step 4: Analyze Experiment Data
The last step is to analyze any data that we collected during our experiments and figure out if our initial hypothesis and prediction was correct. In this stage, you may be analyzing:
- Interview responses
- Analytics data
If your prediction was wrong, you can come up with a new prediction and start the process all over again. If you think your hypothesis is wrong, you can start with a new hypothesis and start the process all over again.
As you do more experiments, you will become quicker at coming up with hypothesis, designing experiments and analyzing the data.
Running experiments and being wrong sucks. There’s no way around that but spending your energy building the wrong product sucks even more.
Here’s a quote from Richard Feynman, on how you should approach the idea of being wrong:
“If the guess disagrees with experiment, it is wrong. In that simple statement is the key to science. It does not make any difference how beautiful your guess is, it does not make any difference how smart you are, who made the guess, or what his name is — if it disagrees with experiment, it is wrong.” – Richard Feynman
How Facebook and Hubspot Use the Scientific Method To Achieve Exponential Growth
Let’s now look at how high growth companies like Facebook and Hubspot use the scientific method to grow.
Former VP of Growth for Facebook, Chamath Palihapitiya, developed the following model for understanding growth:
The model looks at the following three areas:
1. Top of the funnel: traffic sources and conversion rates (signup, email newsletter, etc)
2. Magic Moment: this is also sometimes called the A-ha moment when your users realize the value behind your product.
3. Core Product Value: this is when you product starts solving a real problem for your users.
You could use the above framework to prioritize what to focus on and then use the scientific method in business to improve each area. Let’s imagine that you’re unsure of your Core Product Value which means there’s no point in driving more traffic to your site yet.
Instead, design experiments (e.g. customer interviews) to figure out your Core Product Value. Once you you feel confident in your Core Product Value, you could move on to other areas like driving more traffic and improving the conversion rate of website.
Let’s look at how Hubspot views growth.
The steps in this growth model are as follows:
- Brainstorm ideas (hypothesis creation)
- Prioritize ideas
- Test (experiments & predictions)
- Implement experiments
- Analyze experiments (final step in scientific method)
The Hubspot growth model follows the scientific method quite closely and they have added a few steps for clarity. They have also designed templates and documents for managing all of their experiments which can easily be replicated by any startup.
Finally, let’s look at the Lean Startup methodology by Eric Ries.
We can also see the the scientific method behind the Lean Startup movement. The process of coming up with ideas, testing them and measuring the results of your tests is a fundamental concept when it comes to building startups.
If you have any questions or comments, you can message me on Twitter @ugarteruben
I find most companies are stuck with high-level metrics and they aren't able to properly understand what actually drives user growth for their web and mobile products. To do that, you need the right data and the right tools.
If this sounds like your situation, then you should download our free tracking plan (and tutorial video). This is the document that you should create before you ever implement tools like Mixpanel, Amplitude, Segment, and Intercom. Click the image below to download your own free tracking plan (and tutorial video).