In an earlier article, we made the case that Experimentation Gives You Innovation Superpowers and that companies with a culture of experimentation empower employees to innovate. Our example was Google who has maintained their search engine dominance by continuously raising the bar on search. Google has built infrastructure and tools to make it easy to run experiments that keep making their search engine better.
Now your reaction might be - “But that’s Google! They have tons of money so of course they can afford a culture of experimentation and all the technology investments required to sustain it.” Well, your company may not be as big as Google but you probably can’t afford to not have a culture of experimentation. Making product decisions that rely on gut and opinions is not a recipe for long-term success. The good news is that any company without much more than a curious mindset can start experimenting today.
In Eric Ries’ The Lean Startup, he describes how by running customer experiments we maximize our chances of success by removing uncertainties on whether our product will succeed or fail. Every experiment that yields customer-validated learning gives us greater confidence in our idea or provides us new insights on how to improve our idea (or if appropriate, abandon the idea). Running these experiments can be done more cheaply and efficiently versus investing the time and effort to build out the full product, while crossing our fingers that it will succeed when launched.
So let’s say you have an idea for a product you believe millions of customers around the world will love. Here are a few baby steps you can take to start acquiring customer-validated learning:
Recruit a unit-of-one customer
Can you recruit just one customer to sign up to be an early user of your product based on its description? This may seem like a ridiculously low bar but you’ll be surprised how many ideas fail this simple test (if you’re having a hard time finding one customer, maybe your idea isn’t so great after all). We call this customer a “unit-of-one” customer and encourage all teams to find their unit-of-ones before they start developing their product.
Your unit-of-one customer will often provide you invaluable early feedback. When teams are debating how to approach a feature, a unit-of-one customer can be a tiebreaker and provide real world use cases. The unit-of-one customer can eventually be a helpful reference to other customers and help you understand what the real market for your product is.
A team at Intuit had an idea for a mobile-based cash register for mom-and-pop shops in India they called ShopOwner. The product would replace hand-written receipts which are prevalent in small shops. Engineer Mithun Shenoy recruited his mother who ran a beauty salon to be their unit-of-one customer. One of the surprising insights she provided the team was one of the most important benefits of ShopOwner is that it made the business look much more professional (colloquially described as “Hi-Fi”) which reduced customer bargaining.
Dry Market Tests
A dry market test gauges interest in a product concept before the product is ready by presenting marketing material to potential customers and measuring how many demonstrate an interest in purchasing. A customer can demonstrate interest by a range of commitments including clicking on a “Buy Now” link, signing up for a waiting list, or paying an actual deposit.
Dry market tests are one of the leanest experimentation techniques given the only effort required is to develop marketing copy and finding customers to run the experiment on. Companies from just about any industry can use this type of experimentation. Even non-technical employees can quickly assemble web landing pages using a service like Unbounce and run an ad campaign using Google or Facebook. Services like Nielsen’s BASES can provide more sophisticated forecasting for product concepts.
A company can enable the running of dozens of dry market tests by setting aside a modest budget to use some of the services mentioned. Having a dry market specialist in the organization to coach teams on the best approach can ensure consistency in results and maximize learning.
Concierge MVP
In a concierge minimum viable product, employees deliver the value proposition manually. By manually performing the tasks that constitute what the product will eventually offer in an automated way, employees can test how much the value proposition matters to customers and the best way to deliver that value. Also, it’s much easier to run several iterations of experiments given no product development is required to make design changes.
An Intuit team was tasked with improving a call center’s Voice Response Unit (the automated telephone answering system to route customers to the right agent). They found out that it took several weeks to reprogram the system which would be way too long to experiment with different approaches. They came up with the idea to have real people replace the VRU. When customers heard “Press 1 for billing questions”, they had no idea they were hearing a live person who would route the call manually. Using this approach, the team was able to quickly test several permutations - no programming required!
These experimentation baby steps will get your company started on your journey to developing an experimentation culture. Developing a mindset of data over opinions is the biggest shift. After that, the organization will be bought in to start making investments in developing experimentation infrastructure that make experiments more rigorous and easier to run. Google didn’t generate their awesome infrastructure overnight - that took years of continuous investment. If you want to develop this innovation superpower, start with these baby steps today!
Jeff Zias and I are writing the book Grassroots Innovation in the Enterprise to be published this fall. Do you have stories to share about how you’ve used experimentation in your company? We’d love to hear from you and share some of the best stories we gather in our book.