Alton Brown is Smarter Than Most CEOs and TV Personalities

My father will tell me over and over, 'If you always do what you always did, you always get what you always got.'

That goes for:

• falling off skateboard ramps
• talking to girls
• spending / saving money
• buying cars
• work & my career in just about every aspect.

Every time you change something, you impact the end result.  In business, we decide we are moving to quickly to look at WHAT we changed that made the difference and more concerned with fixing the fallout or basking in the adulation of the win. 

It is time to explore the use of a scientific tool in the business world with all the data already sitting in a corner waiting to be tapped. We have experiments, we just don't document them and often, people are not honest about the variables, effort and thus results.

So where can we look to for examples and variations?  The news is now digital, there are plenty of business articles & books out there and of course there are blogs, us spewing our minds from the perspective of what we have experienced.  With that in mind, I am going to make an effort to add more retrospective views on situations and the point I think they changed direction, the variable.  This is loose on the scientific part since the variables are plenty in daily life and we are focusing more on the research.

What is the scientific method at its core?

A seven step process to getting where you want to be.

1. Question; build your 'what if'.
2. Research; utilize what others already have learned/experienced.
3. Hypothesize; If we <insert change here> then <result> will happen.
4. Test; In business this is more of an implement/take risk idea.
5. Analyse and Conclude; Determine if your change got the desired result.
6. Proclaim the result; If your hypothesis was true state 'I knew it all along' or, conversely, 'Here is what I learned'. Don't take the easy way out and say I would have been right but something else made it happen different.
7. Report on it; Who cares if you learned something, add to the greater good of the knowledge that exists in this world.

Want a more detailed, simplified, actual version of the Scientific Method, check out ScienceBuddies to occupy a little more space in your brain. I was inspired that this method does apply to the office by a great research project that was done involving personal happiness and giving people envelopes of money covered on Ted.com.

Here is how the steps can apply more to business life and not just the IT guys in the corner, although some of them can use a refresher on the scientific method. (I'm talking to you Mr. 'we are going to have to re-image your laptop' for any damn problem.)

Step 1 Question Development

This is the day to day decision making we do at work. What if...

• we hire a less skilled member and train them instead of paying for a sales-pitched expert?
• send the team home early every so often?
• do just enough testing instead of being thurough?

You have these questions coming at you every part of the day, these are not the problem. It is identifying the one you want to learn & teach from and tracking it, documenting it and growing from it.

Step 2 Research

You have the internet right? So fire up your modem (anybody else hearing 9600 baud sounds in your head) and start typing away. Do the research on your phone while waiting for a train, sitting in the meeting nobody is on time for, whatever. Searching on what you would use as the headline for your report would be a good starting point. "Is it worth developing a skill instead of hiring it in?' might bring a few decent results from your search engine of choice... I tried it, it does... go ahead, verify it, I'll wait.

Satisfied, ok on to...

Step 3 Hypothesize

This is where you take your question, stack your research on top of it and proclaim you are going to use your environment to determine just how genius you are. So from the ones above I would have:

• IF I hire a less expensive, semi-skilled employee and train them then I will have a productive worker that cost less in the first 2 years than hiring that skill set in.
• IF the team leaves early on slow days and let them work remotely then I can increase morale and get more out of them on a regular basis.
• IF we test basic functionality then the additional, rarely used features will go unconfirmed but we would save time.

That seems pretty straight forward so on to the next item.

Step 4 Test

In the business world, this would mostly consist of implementing the change or observing what happens AS the decision has been implemented.  Your organizations is rarely going to let you set up a control group and a hypothesis group of employees to do the same job.  This is where we are going to be questioned by the scientific community on the validity of our results but this is the best measure we have right now or implementing change so we roll with it.

Step 5 Analyse & Conclude

Look back at your belief at the beginning, look at what happened, OK, now look at it again and be honest with yourself.  You are likely clouding your judgement so go ask somebody else to confirm they saw what you saw.  This is the point you have to be honest with yourself and say, 'meh, that didn't work', without giving yourself the option to dismiss it because of another variable.

Step 6 Proclaim Your Result

This may be a personal proclamation as in an email to yourself or a note stating a recognized behavior or a discussion at a team meeting.

Step 7 Report It

Figure out how you are going to do this during these experiments / learning exercises / growth opportunities.  It may be as simple as writing a blog entry or journal page about it or something as formal as a report to a superior or colleagues for their understanding.  This will give you some sense of finality to it and a growth for next time.

Cost to Business of the Scientific Method


To perform true experiments inside of a business structure would be significantly costly and not tolerated by stakeholders, executives and clients; that is what colleges are for. What I am proposing is use a similar approach, acknowledge the multiple variables but don't dismiss every problem as NOT MY FAULT. At the end of every corporate experiment do some documentation so you, and maybe others can learn from it.