Lesson 2: Building Islands of Innovation 


Welcome to the second in our five part email series about Lean Retail.  In the previous lesson we talked about identifying your assumptions and ideas.

Here's a brief recap:

Each aspect of your retail business (marketing, finance, operations, technology) should be broken down into the current drivers of your success (assumptions) and the changes you believe could drive this success forward in the future (ideas).

When truly ‘running lean’, each of these assumptions and ideas is then treated as an hypothesis that must be tested and proved. Every new initiative is a trial, until the data tells you otherwise.

Once again, we love to live lean so please help us improve the Lean Retail 101 emails by telling us what you think! You can reach us via hello@theleanretailer.com or on twitter or Facebook.

Lesson 2: Building Islands of Innovation 

In this lesson we discuss the following questions:

How do we separate the great ideas from the merely good?

How do we explore the merits of these new ideas in the most cost- and time-efficient manner possible?

How can we run scientific tests in an unscientific environment?

This lesson is broken out into two steps: 

  1. Prioritization: Knowing What to Build 
  2. Designing Experiments: Knowing How to Build

Prioritization: Knowing What to Build 

A big part of the job of a Lean Retailer is knowing which initiatives are most worthy of your limited your time and resources.  

If you’ve completed the exercise from the last lesson you should be set with a big long list of assumptions about your current business and ideas for possible new initiatives. These should be broken out according to the different aspects of your business, i.e. marketing, tech, finance, etc.  It’s now really important that you go through these lists, select the best and turn your ideas into an actionable roadmap of new real-world changes. 

Now, the whole idea behind build-measure-learn is that we can only really learn what’s ‘best’ by building tests and measuring the results. However, we have to start whittling down this list somehow.  A large part of this early stage prioritization is down to intuition but there are a few questions we can ask ourselves to help get started:

1. Will this grow my bottom line? 

Go through your lists and put a dollar sign next to any initiative that you think could increase sales or cut operating costs.

2. Will this make my customers smile?

Go through your lists and put a smiley face next to any initiative that will improve the experience customers have when they visit your store.

3. Can I get this done in the next 30 days?

Go through your lists and put a clock next to anything you believe you could get a trial up and running for inside the next 30 days.

In an ideal world, of course, you’re looking for initiatives that meet all three of these criteria!

Designing Experiments: Knowing How to Build

At the heart of the build-measure-learn cycle is a faith that we can implement changes in our businesses and accurately measure their impact. The problem is, this can be very hard to ensure.  Sure, sales went up following your expensive refurbishment - but how much of that was really down to the new paint job? Maybe it was that new sales person you hired? Or maybe that new condo building on the corner increased foot traffic on your street? The answer is that we simply can’t recreate the science lab out in the real world. 

That doesn’t mean we can’t take meaningful steps to bring some scientific rigor to our businesses innovation. 

The best kind of business learning is cheap and quick to access and leads to consistently replicable actions. There are two key concepts that I believe can help retailers find these drivers of success and do so cheaply and consistently: the Minimum Viable Product; and what I call, ‘Islands of Innovation’.  

Minimum Viable Product (MVP)

"...it's really hard to do big and good simultaneously... you can either do something small and good and then gradually make it bigger, or do something big and bad and gradually make it better. And you know what? Empirically, starting big just does not work."  

- Paul Graham, Founder of Y-Combinator

In the technology startup world, a Minimum Viable Product (MVP) refers to a new product with a limited feature set that provides ‘just-enough’ functionality to enable the developer to test the market for their new ideas.  

The first version of Gmail was literally written in a day. It wasn’t a version with all the amazing filters and intelligent usability we expect from it today.  It was simply good enough to let the people making it measure the demand for a better email system. 

There are lots of examples of technology startups testing the market without actually committing huge time and resources to building a fully featured, fancy product.  For example, the daily-deals giant, Groupon started out life as a simple WordPress blog with a widget that sent out their deals as PDFs. It’s now worth billions of dollars.  

Founder of Dropbox, Drew Houston even managed to go one further. He ‘launched’ his revolutionary new product via a simple 3-minute demonstration video. It was a huge success growing his subscriber list from 5,000 to 75,000 people overnight.  The cool part - he hadn’t even begun coding the product. Dropbox is also now worth billions of dollars. 

The point is this - if you want to start selling baked goods, you don’t have to commit to installing a bakery in the back of your cafe. And if you think a particular part of Chicago needs a new jewelry store, you don’t have to put a downpayment on a property to find out.  The goal is to find quick and easy tricks for assessing the demand for your product.  

For an example of how this kind of thinking might work in practice, check out the article ‘Are you Smoke Testing your Business?’

Islands of Innovation 

The key to carrying out real-world tests is creating ‘islands of innovation’ - that is to say, isolated environments where we can implement a change and, in as far as possible, control the environment. 

When running an online business there is a simple process called Split Testing that can be really effective.  For example, a developer may believe that changing the color of a website’s sign up button to green will increase sign-up rates. Green means go after all. In order to test this theory, the developer simply redirects a small percentage of web visitors to a new page showing the new, green sign-up button. Comparing the sign-up rates from the two different pages, the developer can quickly gain data on which is the more effective. What’s more, the developer is able to take this risk knowing he is only experimenting on a small percentage of his customer base.  

Creating similar environments in retail is more difficult. However, there are a number of ways we can help control the ‘noise’ and risk around our innovation efforts.

1. Trial via e-commerce

E-commerce is increasingly becoming the local retailers’ new best friend.  Cheap, easy to set up and stylish e-commerce platforms allow merchants to access new customers and try out new concepts in a controlled manner. We can introduce new ranges, take risks on new products and get really great data on which to base our decisions.  

2. Off-Site Testing

More so than ever before, the chances are that the great new restaurant you love so much started out life as a food truck or as a stand at a farmers market.  These environments can serve as incredibly useful low-cost MVPs for new businesses or as useful product testing grounds for existing retailers.  

I love the story of Bang Bang Pies, a business in Chicago that started out life as a food truck.  Not only were the owners able to refine their product offering and find out which pies were most in demand, but the mobile nature of the food truck let them collect real-world data about successful sales locations. They are now set up in a brick and mortar store around the corner from their highest performing food truck location.

3. Time-Sensitive Testing

Introducing new innovations only on certain days of the week can create a less risky way of measuring demand for something new.  

Imagine a coffee shop owner decides that releasing the aroma of freshly baked bread throughout their store will increase sales of baked goods and that the aroma of freshly brewed coffee will increase sales of coffee.  Or similarly, imagine a clothing store believes that playing upbeat music will  get people amped up and increase sales.  By putting these initiatives in place only on certain days, these retailers can effectively create a control group against which to test the impact of our changes.  Also, if the bakery smell is off-putting and actually damaging to sales, we have only exposed only a small number of customers to the change.

4. Crowdfunding

There is no surer way of knowing that people want your store in their neighborhood than asking them to commit a small amount of money to make it happen. Kickstarter and Indiegogo are websites designed to enable those with new ideas to crowdsource the funds they need to make them a reality. They are not currently widely used for brick and mortar retail ideas but they could be an amazing way to test the demand for new store concepts. 


Some time in the next few days, step away from the counter, find a quiet corner and fill out the following two tables with your priorities and the initial experiment ideas you have.    


Knowing What to Build

Define what’s worth working on and what should be put off until later. 

Business Area Idea Why is this important?
Marketing Building out a Facebook Page More customers
Marketing/Sales Put large poster in window announcing 25% off Sunday Sale More customers, Increase revenue
Sales Play upbeat music to get people in a better mood to buy more More customers, Increase revenue

Download as a PDF to print and complete »

Knowing How to Build

Define how you will test your idea and how you will keep that test controlled and efficient. 

Business Area Idea MVP Island of Innovation
  How am I keeping this time & cost efficient? How am I controlling this test?
Sales Play upbeat music to get people in a better mood to buy more Instead of installing an expensive speaker system, bring in a stereo from home Play upbeat music in-store on Tuesdays and Thursdays for three weeks. Compare sales numbers with Mondays and Wednesdays.

Download as a PDF to print and complete »