1.4 What Kinds of Growth Can You Engineer?

You can engineer growth for digital and physical products and even services.

The most obvious application of growth science is in software, which includes popular categories for innovation such as mobile apps and cloud-based software. Growth science is particularly enabled in this space by excellent tools for conducting quantitative experimentation on the design of these interactive products. Product owners can test variations of design (e.g. which button color gets more clicks) as well as functionality (e.g. does adding a particular function increase key engagement metrics such as total time spent using the app).

Similarly to product development, a lot of marketing for physical products and services is done online where a great ecosystem of experimentation tools also make it possible to optimize marketing channels and messages. For example, it is now very easy to test variations on messaging in email marketing or to compare the effectiveness of channels such as pay-per-click advertising to email marketing.

Beyond tools, there is rich community of growth hackers, marketers, and product designers that share ideas and best practices. In our opinion, all these factors have led to the advancement of a holistic approach to growth in the software space, which other industries can certainly emulate.

The growth science methodology can just as effectively be applied to physical products and services. As mentioned above, the big advantage with interactive products is that it’s easier to make changes to the product on the fly, and while this might not be as feasible with physical products, we can certainly experiment with marketing, customer support, sales, and other aspects of operations.

Even though it’s more challenging, we absolutely conduct experiments on physical products as well as services. With physical products, you don’t necessarily have to test variations of the actual product to see which version is most likely to resonate with customers. For example, let’s say that you are going to be selling sunglasses and you are not sure with which styles and colors you should launch. You could simply create variations product pages for various combinations of styles and color in your online store and measure which ones register the most number of clicks on the “Buy” button. Voila, you just tested product design of a physical product!

We could also try variations to how a service is delivered to test what works best. For example, let’s say that you are an executive at Goliath Bank. What if I told you that you could grow your business by optimizing how your employees deliver banking services to customers at bank branches? Let’s say that my hypothesis is that customers would be happier if they performed more straight-forward tasks such as depositing checks on Goliath Bank’s mobile app rather than going through the trouble to travel to a branch and standing in line. We could test various ways to encourage late adopters to try making check deposits via the online app. For example, we could try showing tutorial videos on monitors at some bank branches while patrons wait in line. By randomly assigning which bank branches show the video tutorial, we could see if it gets customers to make more check deposits via the mobile app. Along with usage, we could measure customer satisfaction and really see if these changes are making customers happier with your service.

While this book focuses on businesses that build software products, the growth science methodology can be applied to practically any other kind of product or service offering.
With physical products and services the main challenge is finding ways in which one can try variations of product or service implementations.

Be sure to check back next Monday to learn what we’ll be covering in the remaining chapters of the book. New sections of Growthzilla are published every weekday.

1.3 Growth Science Pillar 3: Optimization Across the Entire Customer Lifecycle

Optimize across the entire customer lifecycle to reach the greatest growth.

The biggest mistake that business leaders make in trying to drive growth is focusing on just customer acquisition through marketing. This is a problem because their customer retention and engagement are usually suboptimal, which means that the resources that they leverage to increase acquisition are squandered when those hard-won customers abandon the product right away. This is called the “leaky bucket problem,” which perfectly captures the futility of this approach. For this reason, the third building block will be learning how we can drive growth by optimizing each stage of the customer lifecycle– the third pillar of growth science.

Exclusively focusing on optimizing acquisition through marketing is unfortunate because it has led businesses to miss out on the full potential of growth optimization. The main thing to understand is that the effects of growth optimization are multiplicative and improvements compound. This means that you can often accomplish greater total growth by making many little improvements rather than just one big one. In particular, it’s better to optimize the entire customer lifecycle including acquisition, engagement, and retention. Let’s explore how this works.

Let’s imagine that you are a very good marketer and are able to improve your customer acquisition by thirty percent by experimenting with various marketing messages and different channels such as online advertising and inbound marketing. That’s great progress, but you can do better.

Diagram showing limited effects of optimizing just acquisition and retention.

You realize that many of your hard-won customers are abandoning your product right away, so you talk with the customer support folks and realize that many new customers have trouble with the initial setup. You work to improve the customer support for new customers and you are able to increase how many customers your business retains, which translates to another thirty percent growth. Since those two optimizations are compounding, you will increase your overall customer growth by a total of eighty-two percent. You don’t have to stop there. You can try optimizing customer engagement with better product design, retention with more targeted marketing, and so on.

If you look at the diagram below, you can think of the full universe of possibilities as having nine dimensions:

  • Improving customer acquisition through better:
    • marketing,
    • product design and implementation such as sharing features, and
    • operations such as sales support.
  • Improving customer engagement through:
    • marketing such as more accurate targeting,
    • product design and implementation, and
    • operations such as training.
  • Improving customer retention through better:
    • marketing such as promotions,
    • product design and implementation such as fewer usability problems, and
    • operations such as pro-active customer support.


Diagram showing the effects of optimizing across acquisition, engagement, and retention.

If you had made even modest gains along all nine dimensions, those gains would accrue to tremendous total growth. For example, even if you improved growth by ten percent in each of the nine areas, that would result in 136% total growth, which is a lot higher than improving just acquisition by thirty percent.

Growth science is iterative, data-driven experimentation to optimize the entire customer lifecycle through marketing, product design, and operations.

Fortunately, some companies are starting to realize that growth optimization spans far beyond just marketing and product design as well as beyond just acquisition or retention. Those companies are systematically applying iterative experimentation to marketing, product development, and operations to improve customer acquisition, engagement and retention. All three of these pillars together is what makes up modern growth science.

Be sure to check back tomorrow for the next section: 1.4 What Kinds of Growth Can You Engineer? New sections of Growthzilla are published every weekday.

1.2 Growth Science Pillar 2: Optimization Across Disciplines

Growth is multidisciplinary

While corporate structure continues to silo departments, growth has quickly become an interdisciplinary practice requiring functions such as marketing, product development, customer support, sales, and even logistics. The third pillar of growth science is employing all of these different functions to spur customer and revenue growth. It is important to recognize that growth isn’t just driven by marketing but rather by multiple functions that are interrelated and must work together to have a positive impact on the business.

Sean Ellis, who coined the term “growth hacking,” was hired at Dropbox right as the company had decided to open the product to the general public. The biggest driver of growth at Dropbox at that time was when the company created a referral program giving existing customers additional storage space if they got their friends to try the product. While the strategy and vision for this kind of initiative might be driven by marketing, implementation often relies on the product, such as a website or app, that makes it possible to send and track invites and to reward users. That means that this tactic relies on both marketing and product development functions, and the two teams must work closely together to successfully implement it.

Uber’s customer support is another example of how different functions within a company work together to affect growth. One of the first times that I used Uber, I requested a ride, and the driver never showed up to pick me up. What made things worse was that I cancelled my ride after waiting for a long time and was charged a cancellation fee. Needless to say, I was furious contemplated deleting the app. Uber could have easily lost me as a customer then, or at least I might have been more likely to use a competitor such as Lyft. However, Uber made it extremely easy to request that the cancellation charge be reversed. It took only a few taps of my finger, and the refund request was submitted. Not only that, I was granted the refund that same day. Uber was able to retain me as a customer through great product design and lightning fast customer service.

In many regards, job boards hint at the interdisciplinary nature of growth science, and you can often see postings for “growth hacker,” “growth manager,” “growth engineer,” and “head of growth.” These titles did not exist a decade ago, but now the demand for these kinds of roles is ballooning. Business leaders created roles such as  “growth manager” and “head of growth” because those responsible for these efforts needed to oversee across departments and traditional roles that focused on a narrow domain didn’t fit. What is critical to understand is businesses are better able to identify new strategies when they understand how all functions impact growth.

Be sure to check back tomorrow for the next section: Growth Science Pillar 3: Optimization Across the Entire Customer Lifecycle! New sections of Growthzilla are published every weekday.

1.1 Growth Science Pillar 1: Iterative Experimentation

Meme: Never stop testing, and your growth will never stop improving.

The origins of growth science could be traced back to David Ogilvy, one of the most well known ad men of the twentieth century, who proclaimed back in 1963, “Never stop testing, and your advertising will never stop improving.” Ogilvy’s statement embodies two important concepts–continually test and improve–that have paved the way to modern growth science.

The idea that advertising is not a static creation is critically important because challenges one to continuously improve tactics. Although this philosophy was popularized in advertising it is just as relevant in product development, operations, and even sales. Since the 1960’s some marketers have embraced this ethos by iteratively trying variations of tactics, channels and messaging. Moreover, the practice of iteratively implementing improvements has become more commonplace in other business functions such as product development, customer support, and sales. However, It’s not enough to keep trying different tactics and implementations.

The second key concept that Ogilvy proposed is actually testing new variations and improvements by carefully measuring associated outcomes. Once again, this practice is not just limited to marketing activities, and one can test variations of product implementation, operations, and sales. It’s not surprising that many business leaders have adopted both iterative experimentation and rigorous testing across business functions, and it’s now just as common to test variations of product design (particularly in the digital realm) as it is to test different advertising messages.

This paradigm shift has been supported by an ever burgeoning ecosystem of data collection and analysis tools that have allowed companies to track key performance indicators for new customer acquisition, such as click-through rates on online and email marketing, as well as for engagement and retention of existing customers. Now, one can easily test variations to marketing emails or to webpage designs to see which ones lead to the greatest gains in key metrics such as the total number of new sign ups or daily active users.

What was once driven entirely by intuition can now be informed by experimentation and testing. Leaders across marketing, product development, and operations are forming hypotheses and conducting experiments to see which tactics yield the most positive results. This rigorous methodology is the first pillar of growth science and underlies the entire field allowing us to call it a “science.”

What is shocking is that after more than half a century, most marketers still do not make data-driven decisions according to a survey conducted by Google in 2017.  I can empathize with those that do not put this philosophy into practice. While the field has burgeoned, it has also become a lot more difficult to navigate. There are now countless tools that allow people to experiment with marketing and product design. The scope of growth science has also greatly expanded across disciplines–the second pillar of growth science– from marketing to product implementation and operations. The application of this methodology has also expanded to include not only new customer acquisition but also to engagement and adoption. The optimization of the entire customer lifecycle is the third pillar of growth science.

We will cover the other two pillars of growth science in subsequent sections. Hopefully by explaining the three elements of growth science and providing a comprehensive guide, we can empower you to successfully employ growth science to supercharge your business.


Be sure to check back tomorrow for the next section: Growth Science Pillar 2: Optimization Across Disciplines! New sections of Growthzilla are published every weekday. 

Chapter 1: An Introduction to Growth Science

I remember looking at a picture with colorful tiles overlaid over a mock webpage. In the upper left there was a red-orange tile and another one right below. Across the top and down the left-hand-side of the page there were more orange and bright green tiles. As my eye gazed across to the right and down the mock page, the tiles got blue and finally purple. I was at Google in 2005 where my job was helping website owners figure out how to implement Google AdWords ads on their site. The picture that I was inspecting was a heatmap, which was derived from aggregate statistical data showing which ad placements resulted in the highest number of clicks.

I was astonished. I realized that design on the web was a big statistical game. Users had tendencies such as where they tended to look first on a screen and what colors tended to catch their eye. It was just a matter of using experimentation to figure out what those tendencies were and designing products that played to those preferences rather than worked against them.

This idea that we can test the effectiveness of product design resonated with me. Having studied physics and economics in college, I had a strong preference for proof and hard numbers over subjective intuition. My academic background led me to adopting a data and research-driven approach to designing user experiences for interactive products like websites.

I soon discovered testing tools such as eye-tracking, which allows one to track what a user is looking at on the screen, as well as A/B testing, which allows one to compare how effective two versions of a page are in relation to each other. I also read results of experiments that others were performing with headlines such as “Placing the security logo on the upper left between the search box and the navigation bar increased conversions 8.83%.” 

Later, during my tenure as director of user experience in a Silicon Valley startup in 2007, I decided to use experimentation to drive product design. The startup was a document sharing site that was one of the most visited web properties in the world. I suggested to Senior Director of Product that we should try experimenting with the registration button on the home page. He laughed and told me not to waste my time, but luckily the founders were supportive of the idea, and we did it anyways.

One of the senior engineers (a brilliant gentleman that later became a very successful entrepreneur) developed our in-house A/B testing platform. First we figured that maybe moving the registration button to a more prominent place might help our conversions–the proportion of people clicking through to the registration page. So we moved the button from all the way on the right of the menu bar to the center. Lo and behold, our conversion rate increased by more than 70%! I was stunned that simply moving the button to the left of the page by a few hundred pixels (effectively about two inches on a monitor) made such a huge difference. I think we all were.

Given our initial success, we decided to keep going to see what other optimizations we could make through testing variations. We next decided to change the design of the button itself. Originally the button was a bland shade of blue, so we decided to try making it bright green based on the hypothesis that users would more readily notice a brighter color and would, thus, be more likely to click it. Another round of A/B testing revealed that the green version gave us about another 50% increase in conversion rate. This moment in time changed my life. At that moment, I thought that soon everyone would be experimenting and testing product design because it had such unbelievable potential to help businesses grow.

Over the next decade, two important trends occurred. First, leaders figured out ways to experiment with changes to not only product implementation but also to marketing and operations such as customer service. Second, many new tools appeared in the market allowing businesses to experiment with things like marketing messaging and even customer service process. Some companies have taken full advantage of the methods and tools to dominate their competitors. Unfortunately, these exciting advances have not yet reached a mass audience, and many businesses continue to fall behind.

In speaking to friends, clients, and colleagues, we heard the same refrain: the possibilities seem so numerous and the methods so opaque that many folks simply do not know where to start. In addition, many people had heard of terms like “growth hacking,” and incorrectly believed that they should replicate hacks that other businesses have used rather than adopting a broader process that they can contextualize to their business and implement to reliably grow their customer base and revenue. Our aim in writing this book is to give everyone an overview of the three pillars that make up growth science as well as practical tactics for growing your business.


Check in tomorrow for the next section of Chapter 1.