Let's 'Move Fast & Break Things'


Extract from my medium article - Let’s 'Move Fast & Break Things

Amazon’s chief, Jeff Bezos said, “Our success at Amazon is a function of how many experiments we do per year, per month, per week, per day.”

If ‘a startup is a company that is designed to grow fast’ then a startup can also be viewed as a company that runs a lot of experiments. Growth is the main currency of a startup. It’s indicative that the fastest growing startups run more experiments.

Some companies and the number of experiments they run each year.

Netflix: 1,000

Intuit: 1,300,

Amazon: 1,976,

Google: 7,000,

P&G: 7,000–10,000,

Facebook - 10,000

and Airbnb : 36,000

I believe one of the top attributes that will be responsible for the most successful startups in Africa will be the companies that run the most efficient experiments (for customers, channels and products) every year.


If i were still in school then this would make sense.

We need to start giving advise based on our present realities and environment. Silicon valley has an environment that has helped most founders bounce back from failed experiments.

Failed experiments can cause serious damage and depression in this clime. We don’t have such luxury!

I have failed in 3 and it has cost me a few years and some good coins.


I’m assuming you have read the main article. If you have, perhaps I need to exhibit better clarity with expressing myself.

I’m not sure running experiments mean to be involved in one big giant operation, moreso, technology has made carrying it out really inexpensive and qucik for most. It can be as simple as testing a Call To Action, value proposition, color of a button, using a landing page or finding out how different users from different channels act/react when they get to your site.

It’s the learnings from the different little things that add up.

In an engineering example, rather than get your developers/engineers on your team to start coding a feature for the next 2 days, you can get one to commit 2 hours of her time to the test. If the test becomes successful, the engineering team can then work to implement new feature, if not, you have lost 2 hours of 1 dev instead of 2 days of the entire team, documented and learnt what doesn’t work for going forward.

Far more important is building a process, your own experiment management system. A process usually will work anywhere, it’s like a cookbook. Testing/ experiments are similar to and borrow from Lean Methodology framework.

The key objective of running tests/experiment is to save you time and money.

I’ll share an (overly simplistic) example in marketing
So you have a marketing budget of a N100,000 for the month for facebook ads. After completing your Audience research, would you stack your target audiences together and place your N100k as your budget and see what happens like you will with TV and radio?

Or after completing your Audience research, would you create different adsets with different target audiences and place a budget of N1,000 on each, and as the days go by, kill those not working, gradually increase the budget for the ones that work, optimize and scale?

If it’s the second, its an element of testing and experimenting. Finding a way to efficiently use and invest your most precious resources - time and money.

For me, that’s how I understand the meaning of experiments.