Have you ever considered why it is that perfectly good ideas for products have quietly disappeared, while mediocre ones have come to dominate their market? The answer is likely in their growth model.
Product-led growth is something that I have written about a couple of times already, because it really can be the key to success. When the product is just so compelling that its popularity spreads from user to user, and they are finding it meets their needs so well that purchases and upgrades are made without specific prompting, growth takes on its own momentum. Several interesting things happen when a true product-led growth approach is implemented: innovation is expedited, spend on advertising and customer support can be controlled, more users are retained, and users spend more.
Why product experimentation is important
Central to product-led growth is the testing cycle and how it guides the direction of travel. With dependable data on how users respond to new features, it’s possible to make truly informed decisions about product development, and achieve the associated benefits. The advantages of a product-led growth approach can include:
Speeding up the development cycle, to innovate faster and more effectively (partly by honing focus, and dropping unsuccessful ideas quickly)
De-risking projects, by enabling better understanding of users wants and needs
Capturing a true picture of users’ preferences, to provide a solid basis for decision-making
Finding the best solution before sinking significant resources into actually building it
Democratising the decision-making process; good ideas can come from anywhere within the company
Why is a culture of experimentation not just the norm?
With all its benefits, product-led growth might seem an absolute no-brainer. But if that is the case, then why isn’t everyone doing it, and doing it successfully? The answer is, commonly, that they are not carrying out the volume of testing to achieve the insight required to make a difference. There are a few reasons for this, such as:
Risk. There are some sectors in which this approach would not be at all appropriate, since getting it wrong would result in catastrophic consequences. For example, nuclear power station control systems are not the place to begin experimenting in a cavalier fashion! However, this type of barrier to product-led growth only applies to a small number of cases, of course.
Cost. On a related note to the point about risk, if you can’t afford the cost of failed experiments (even a tiny ‘failed’ experiment), then you can move a lot more slowly and try to get everything right. Of course, this is a risk in itself as you might not hit on the right solution, or you may get beaten to the finish by a competitor.
There is no downside to just doing it. This is another rarity, but in cases where there is no cost to building the whole thing, then it might be better to just do it, and not bother experimenting.
Size. Very small companies might have insufficient people to build an experimentation culture in any kind of structured way (though that’s not to say they can’t adopt the mindset).
Apart from these notable exceptions (some of which are quite rare), the obstacles to a full product-led growth approach are often more about shared behaviours, beliefs and values. The predominant mindset in business is still that failures are wasteful, and that predictability and efficiency are important, which makes managers and leaders reluctant to allocate resources to anything experimental. But of course, without experimentation, innovation is hampered.
With the right approach to experimentation, innovation can be both faster and better, so the (controlled) risks can certainly be worth it. Overcoming the aforementioned barriers can be a challenge, but one that is worthwhile tackling – and the way to do that is to tackle the prevailing culture within your organisation.
How can you embed a culture of product experimentation?
The short answer is ‘make experimentation part of the everyday.’ To do this, broadly speaking you must create an environment that encourages creativity, where data trumps opinion, and where anyone can have an idea and run a test (not just those within R&D, cross-functional teams are pivotal in maximising chances of success).
Establish a context in which people are free to experiment, and provide support to help people get to grips with ‘failing on purpose’. Make sure they are aware that it’s okay to fail, and that the learnings achieved by experimentation are more important than getting things right first time. When there is psychological safety within the team, and they know that there is no negative outcome triggered by getting it wrong, they will not be afraid to try new things, resulting in a much-improved capacity for creativity and innovation.
Build an experimentation team responsible for viability, comprised of a product manager (who also owns responsibility for customer value), a product designer (who owns usability) and a technical lead (who owns feasibility).
Encouraging curiosity across the workforce is critical. How can individuals use their knowledge and experience to suggest improvements or new features? Give them a way in which to do that, and set out how to show that they have moved towards your organisational goals. Be explicit about what the goals are, and how you will measure them.
It’s important to encourage a culture in which anybody can have a good idea, from the admin assistant to the CEO; this diversity of perspective supports the best possible chance of coming up with the right user-centric improvements. Leaders must be humble, and recognise that just because an idea didn’t come from the top, doesn’t mean it isn’t a good or insightful idea.
Once these steps have been followed to set the tone for a new culture of experimentation, you should create a process for people to follow when they have an idea they want to experiment with. The following steps should all be considered and included:
Set a company-wide challenge in the form of a question that you want people to answer: would doing X drive more Y?
Create a hypothesis document to record and share thinking. Our Blue Ocean Insight hypothesis template is a handy tool for this, drop us a line to receive a copy. Treat each hypothesis as a ‘fail’ until proven otherwise, which helps to ensure a data-led mindset, rather than getting too excited about the idea itself.
Decide on the goals and objectives for your experiment (the BOI template helps with this).
Decide on a small number of variables. The ‘small’ is key; you can’t play with 100 variables and try to tweak them all, as you won’t know which one made the difference.
Decide what metrics you need to prove the idea is failing and then collect those signals (if you can’t prove a failure, then it is probably a success). So-called ‘pirate metrics’, ARRR play well here, as do HEART, while a mix of leading and lagging indicators can also be worthwhile – though watch out for ‘vanity metrics’, Goodhart’s Law, and confirmation bias. Also be aware of ‘the tyranny of the marginal user’, which can result in products actually becoming less useful, because the focus has been too heavily on customers simply using a product, rather than how well it fulfils their needs.
Decide when you will have enough data to make an informed decision (and consider when to bail from the test if it tanks).
Consider ethics. Although inserting a formal policing or ethical review board process is to be discouraged (it creates a bottleneck, and detracts from the team feeling empowered), some awareness and oversight is wise, to avoid a backlash such as experienced in 2012 when Facebook decided to undertake some tests that were effectively a kind of emotional manipulation.
Check that you will get actionable answers. Think about the value of knowing the answer to your question. If you had the answer, what action would you take? If you aren’t sure, it probably doesn’t tell you much. Will you understand why a user likes feature A over feature B? If not, what will you do with the experiment?
Identify your Ideal Customer Profile (ICP) and understand it (or them, if segmenting); try a combination of these tools:
User personas
User testing
Analytics and data analysis
Social media listening)
Competitor analysis
Customer Support interactions
Beta/Lab programmes
Once you have your ICP [Ideal Customer Profile] and your question, try testing your hypothesis and see if the data support your hypothesis.
Once you have actionable answers, take the actions that had been planned out during the design of the test.
The efficacy of the product experimentation approach is evidenced by companies that have made the leap of faith involved in undertaking tests without knowing whether they will fail or succeed, and found they could accelerate their development cycle and provide a better user experience: Amazon, Microsoft, Google, and Booking.com – and many, many more – are all examples that prove product experimentation works.
If you’d like to talk in more detail about how to nurture a product experimentation culture, and reap the rewards of the product-led growth model, send us an email at info@blueoceaninsight.com