By nature entrepreneurship is closely tied to experimentation because a successful outcome of an endeavor cannot always be known in advance. To fuel success, leadership must be ready to take risks, venture into the unknown, test ideas and adjust the next steps according to market feedback. This exploration doesn’t guarantee success, but if done right will help steer the path towards it.
Experimentation and testing define an iterative course of discovery that is very effective for driving growth and nurturing new business ideas. One can only fully understand the true potential of an idea until it’s tested in practice. This is why many leading companies have intuitively adopted experimentation as a tool to survive and thrive. Data-driven companies like Spotify, AirBnb, Netflix and Uber are known to breathe experimentation to optimize user experiences, test new ideas, products, features or pricing models.
Therefore businesses need to adopt a researcher’s mindset as a standard for testing and verifying activities on a daily basis, take only proof-based action and make a culture of experimentation part of their core values.
In practice, let’s say an e-commerce enterprise wants to evaluate whether its customers prefer discount coupons, loyalty cards, or reward points, they experiment and test how these variables interact with customer retention rates and choose the best performing option for the next campaign.
Evidence gathered from tests should be used as signposts that guide your business on its trajectory towards successfully implementing changes to optimize its value chain and make impactful decisions.
Embracing Failure is Key
Tests often fail - and that's a good thing. A failed hypothesis or idea should not be considered a failure as it provides value and learnings that show the path forward. Embracing a culture of experimentation means getting used to failed tests as they represent valuable lessons learned that help linking the process of discovery not with failure, but with efficiency due to the newly discovered knowledge.
Automating Experimentation to Tap the Hidden Potential in Big Data
The global Big Data market is predicted to grow to $103 billion in revenues by 2027. According to Forrester, around 60-73 percent of data gathered in various ways goes unanalyzed. As more and more companies become aware of the hidden value in data, experimenting will grow in importance to help companies determine exactly where those opportunities lie.
As more and more data gets generated every day, augmented analytics is essential to extract the juice from the data surge. Companies are eager to scale optimization, and always-on experimentation can provide that through an agile process.
The Effect on People
A culture of experimentation is vital for people in two ways:
1. Optimization tools facilitate decision-making by providing clear numbers that bring everyone from your team and customers on the same page. Metrics are powerful persuaders - they boost people’s confidence about why they’re doing what they’re doing.
2. When taken beyond the scope of sales and marketing, the culture of experimentation dissolves staff inertia by stimulating creativity and competitiveness and encourages an entrepreneurial thinking. An entrepreneurial team member is constantly developing ideas and testing opportunities for improvement or growth. This is what you want to encourage as a leader as it will ultimately lead to success.
How to Implement a Culture of Experimentation
There is no one-size-fits-all to how you start nurturing a culture of experimentation, but here are some best practices and simple guidelines that can be helpful to try.
1. Start by assigning an Experimentation Officer (EXO). This can be a temporary or permanent role. If you assign someone who will drive the process forward, you will have a responsible person to answer staff and customer questions.
2. Create a structure or a methodology about how to go through the stages of experimentation. This will help your staff adopt the new mindset quickly. It’s easier if everyone on your team knows what’s their task in running the tests.
3. Make a difference among experimentation, innovation and scaling. By making the distinction, you will get rid of the premise that experimentation must bring instant results. Start by experimenting - search for value. Next, innovate - create value by actually turning test results into practical outcomes. In the end, scale - put those outcomes in the wider company landscape.
4. Keep major data challenges in mind when you run the experimentation process. For example, how trustworthy or representative is your data?
5. Foster competitiveness in your staff to encourage innovation. Gamification and competition support a broader culture of experimentation. Make experimentation fun. Innovation and creativity include an element of play and doing it in the old boring ways just won’t cut it.
6. Experimenting needs a new language paradigm based on an entrepreneurial vocabulary. Abandon words such as expenses and losses and welcome words such as opportunities and potential. If you remain stuck in a mind frame that requires only predictable outcomes, your business won’t grow.
7. Use numbers and timelines when presenting in front of stakeholders. Stakeholders love risk, but not recklessness. Be mindful of the language you use. A solid proposal should show calculated risks by including concrete action about a defined improvement within a specific timeframe.
The Right Tools Will Make It Thrive
A properly implemented and deeply ingrained culture of experimentation can only thrive once appropriate tools are accessible across the organization. Being able to generate actionable results at pace is crucial for the adoption of an immersive experimentation culture.
This is why modern experimentation solutions need to work with metrics that can be directly tied to a business impact and overcome the challenges that come with traditional A/B testing methodologies, such as slow iteration cycles, not being able to peek into tests, limitations to the number of testable variations, automation, inconclusive and misinterpreted results and dependencies on restricted expert resources.
Modern data science methods and models specifically tailored towards a businesses KPIs deliver exactly that - they address data issues at scale and promise to draw business relevant insights automatically. Some key features are:
1. Experimenting with many variables on different granularities in parallel to multiply the knowledge that can be extracted at the same time.
2. Explorative experimentation - as opposed to the traditional confirmative approach of validating hypotheses - to uncover correlations and reveal hidden insights in your dormant data.
At Admetrics we developed the experimentation engine Quantify that supports automated explorative experimentation on any dataset. It’s by orders of magnitude more efficient than traditional testing methodologies while providing many additional benefits that enable always-on experimentation and a culture of experimentation.