What Entrepreneurs Can Learn From the Scientific Method

Startup founders can take a lesson from the scientific method to improve their changes for business success.

Startups can iterate more effectively by adopting scientific methodologies found in the field of science and medicine. When the world of startups and medicine collide, there is a pivotal shift towards better experiments, data collection and data analysis. In many ways from a technical standpoint, the field of medicine is antiquated compared to startups, but when it comes to data collection, data analysis and proving causation, it’s the startups who are seemingly falling behind.

Let’s take a quick look at the steps for the scientific method, the basis for scientific discovery. It includes the following steps:

  1. Ask a question
  2. Conduct background research
  3. Construct a hypothesis
  4. Create/run an experiment
  5. Verify if the procedure worked
  6. Analyze data and draw conclusions
  7. Communicate results

My company applied the scientific method to improve the usability of our product, a software platform for self-checkout systems. In this example, our client uses our software to power their healthy, fresh food kiosk at over 150 corporate locations. Below is the iteration and observation we pursued:

Our Experiment and Findings 

During a couple of kiosk launches, my business partner and I observed queues forming as lunch time came around. Prior to launching, we had added data collection points when a user touches the screen to begin and when they complete their transaction. Upon analyzing three months of data across 100 sites, we found that the average checkout time was upwards of six minutes.

We then  began segmenting our data to identify ways to reduce checkout times. On the kiosk touchscreen, we offer two methods of checking out: users can either scan the product barcode, or manually add the products via a touchscreen menu. We found that over 70 percent of users checked out items using the barcode method. As it turns out, the average checkout time for the barcode users was close to five minutes flat; meanwhile, the checkout time for touchscreen users was close to eight minutes.

Our hypothesis for reducing the global checkout time was to revamp the user interface to hide the touchscreen menu until users clicked a button to access it. We also reduced the amount of images loading on the tablet. Unfortunately, we still required the touchscreen to check out products that did not use barcodes, such as fresh fruits.

We ran this experiment by creating two random groups of 20 sites, and left one group as a control. We then tested the new user interface on the experimental group for a month. When the month ended, we found that the control group remained steady in terms of both barcode usage and checkout times. The experimental group not only had an average checkout time of four minutes and 40 seconds, but barcode usage shot up to 93 percent at checkout. We found that the checkout times for the touchscreen option with non-barcoded items increased by 20 seconds. This highlights the importance of keeping a macro view of your business metrics as you execute these iterations.

Once this was completed, we drew up our findings and shared them with our development and design teams. This simple test created a ripple effect in the organization — not only did the experiment generate more key questions for the business, but it also created a culture comprised of data and accountability.

Using the Scientific Method to Improve Your Business

It is important to note a couple of things when applying this method to improve your business. In most cases, the more narrow the question, the fewer variables you’ll come across. When creating the hypothesis and experimental design, only focus on a few variables. After completing the experiment, run the data points — pre-optimization and post-optimization — through statistical significance algorithms. Make sure you remain unbiased and convey the findings to departments or individuals affected by your findings.

Proven research frameworks found in evidence-based medicine can help startups create a structure around innovation and iteration. The goal for these research frameworks is to prove causation: when a business can show causation between their product and clients in a positive way, it demonstrates just how impactful the product can be on their business.

Positive and negative findings are commonplace in the field of medicine. The same exact thing will happen within the world of startups, which will forever change the landscape of the industry.

Resources

What Entrepreneurs Can Learn From the Scientific Method

Startup founders can take a lesson from the scientific method to improve their changes for business success.

Startups can iterate more effectively by adopting scientific methodologies found in the field of science and medicine. When the world of startups and medicine collide, there is a pivotal shift towards better experiments, data collection and data analysis. In many ways from a technical standpoint, the field of medicine is antiquated compared to startups, but when it comes to data collection, data analysis and proving causation, it’s the startups who are seemingly falling behind.

Let’s take a quick look at the steps for the scientific method, the basis for scientific discovery. It includes the following steps:

  1. Ask a question
  2. Conduct background research
  3. Construct a hypothesis
  4. Create/run an experiment
  5. Verify if the procedure worked
  6. Analyze data and draw conclusions
  7. Communicate results

My company applied the scientific method to improve the usability of our product, a software platform for self-checkout systems. In this example, our client uses our software to power their healthy, fresh food kiosk at over 150 corporate locations. Below is the iteration and observation we pursued:

Our Experiment and Findings 

During a couple of kiosk launches, my business partner and I observed queues forming as lunch time came around. Prior to launching, we had added data collection points when a user touches the screen to begin and when they complete their transaction. Upon analyzing three months of data across 100 sites, we found that the average checkout time was upwards of six minutes.

We then  began segmenting our data to identify ways to reduce checkout times. On the kiosk touchscreen, we offer two methods of checking out: users can either scan the product barcode, or manually add the products via a touchscreen menu. We found that over 70 percent of users checked out items using the barcode method. As it turns out, the average checkout time for the barcode users was close to five minutes flat; meanwhile, the checkout time for touchscreen users was close to eight minutes.

Our hypothesis for reducing the global checkout time was to revamp the user interface to hide the touchscreen menu until users clicked a button to access it. We also reduced the amount of images loading on the tablet. Unfortunately, we still required the touchscreen to check out products that did not use barcodes, such as fresh fruits.

We ran this experiment by creating two random groups of 20 sites, and left one group as a control. We then tested the new user interface on the experimental group for a month. When the month ended, we found that the control group remained steady in terms of both barcode usage and checkout times. The experimental group not only had an average checkout time of four minutes and 40 seconds, but barcode usage shot up to 93 percent at checkout. We found that the checkout times for the touchscreen option with non-barcoded items increased by 20 seconds. This highlights the importance of keeping a macro view of your business metrics as you execute these iterations.

Once this was completed, we drew up our findings and shared them with our development and design teams. This simple test created a ripple effect in the organization — not only did the experiment generate more key questions for the business, but it also created a culture comprised of data and accountability.

Using the Scientific Method to Improve Your Business

It is important to note a couple of things when applying this method to improve your business. In most cases, the more narrow the question, the fewer variables you’ll come across. When creating the hypothesis and experimental design, only focus on a few variables. After completing the experiment, run the data points — pre-optimization and post-optimization — through statistical significance algorithms. Make sure you remain unbiased and convey the findings to departments or individuals affected by your findings.

Proven research frameworks found in evidence-based medicine can help startups create a structure around innovation and iteration. The goal for these research frameworks is to prove causation: when a business can show causation between their product and clients in a positive way, it demonstrates just how impactful the product can be on their business.

Positive and negative findings are commonplace in the field of medicine. The same exact thing will happen within the world of startups, which will forever change the landscape of the industry.

See Also: Networking Advice From Matt Wilson, Co-Founder and Adventurer at Under30Experiences

If you have insights like this to share,

and join us!