Growth teams have become a critical part of product-driven companies. In this article, I'll dive into how growth teams operate and the role which growth engineering plays.

The Prerequisite of Product Market Fit

The objective of a growth team is to take a product that has product-market fit (PMF) and scale product usage. Having PMF is a key prerequisite to growth. If your company tries to scale before finding PMF, it will likely succumb to the effects of premature scaling.

See Product Market Fit.

Understanding Your Conversion Funnel and KPIs

In order to scale product usage, you need to understand the customer journey of how your users discover your product to becoming paying customers. This is answered by defining your product's conversion funnel. In it's simplest form, there are five stages:
* Acquisition - how do users find your product?
* Activation - how do users have a great first experience and receive value from your product?
* Retention - how do users keep coming back to your product?
* Referral - how do users tell others about your product?
* Revenue - how do users turn into paying customers?

Source: Startup Metrics for Pirates

The key here is to measure the conversion rates and volume of users at each stage then run experiments to maximize the flow through to the following stage. For example, if you have 100 users visiting your website and 80 users sign up, your conversion rate for this stage is 80%. Note that this also means you have 20 users which dropped off - find out why they did not progress to the next stage. Read Moz's blog on Conversion Rate Optimization for more details.

For some additional reading, read How we put Facebook on the path to 1 billion users by Chamath Palihapitiya. It describes how Facebook focused all their efforts on optimizing three stages of the funnel:
* How do you get people in the front door. (acquisition)
* How do you get them quickly to that ah ha moment as quickly as possible. (activation)
* How do you deliver core product value as often as possible. (retention)

Implementing Analytics

To build your conversion funnel you first need to gather the necessary data through product analytics.
Implementing analytics is the responsibility of your engineering team. Some of the popular analytics tools used are:

• Google Analytics - a web analytics platform by google
• Mixpanel - a web and mobile analytics platform for tracking user events
• Segment - a platform for tracking and sending customer data to multiple data sources
• Snowplow - enterprise grade anayltics platform stored in a customer's data warehouse
• Full Story - analytics with user screen recording
• Ahoy - simple and roll-your-own analytics for Rails

Client vs Server Side Tracking: When implementing analytics, it's important to understand the differences between tracking on the client-side versus server-side.

Client-side tracking is done on the user's browser or mobile app which means it can track events such as tapping, clicking, or scrolling. A drawback of client-side tracking is that it is often blocked by ad blockers. Special implementation considerations also need to be made if you're using a front-end javascript framework such as React, Vue, or Angular - these frameworks do not trigger a full page reload when navigating to different pages and thus you need to adjust how tracking events are binding and firing.

Server-side tracking is done on the application's server which means it can track all events that happen on the server such as reads/writes to the database, page requests made to the web server, and billing charges made to a customers credit card. To help decide what type of tracking is needed for your application, read when to track on the client vs server.

Assembling a Growth Team

A growth team is a cross-functional team of product, marketing, data, and engineering that focuses on driving full-funnel metrics, for example, daily active users and monthly active subscribers. This team complements existing functional teams which are focused on their specific part of the funnel, for example, marketing focuses on awareness and acquisition, product focuses on activation and retention, and sales focuses revenue. For more details, read Growth vs Marketing vs Product by Brian Balfour

To optimize for rapid experimentation, a growth team ideally has its own dedicated team members. In the presentation How to build a growth team, Andrew Chen provides an example what this structure might look like:
* Growth PM - defines experiment roadmap and leads the team
* Growth Designer - creates quick designs for experimentation
* Growth Engineer - implements experiments
* Growth Marketer - optimizes marketing channels
* Growth Data Analyst - draws insights from data and experiments

Team Characteristics

There are certain characteristics which are needed to work on a growth team due to it's cross-function and rapid experimentation nature. Here's what I look for:
* Ability to rapidly iterate and experiment
* Versatile, resourceful, and entrepreneurial
* T-shaped individuals who have a wide base of knowledge with an area of specialization. Read this blog article by buffer about T-Shaped marketers.
* Discipline and persistence to trust the process with an understanding that most of your hypotheses will fail
* Ability to think from first principals and follow the scientific method

Running a Growth Process

Once your product has product-market fit, it's time to start thinking about growth. Steven Dupree of SoFi and LogMeIn describes growth as the "scientific method applied to KPI's".

By applying the scientific method to growth, we end up with a four step process:

Source: How to build a growth team by Andrew Chen

Forming Hypotheses

Hypotheses are the ideas your team wants to test. Craig Sullivan provides a well structured approach on how to write hypotheses:
1. Because we saw (data/feedback)
2. We expect that (change) will cause (impact)
3. We’ll measure this using (data metric)

Prioritizing Ideas

With so many ideas, it's hard to decide which idea to test first.
Sean Ellis from GrowthHackers recommends to use the ICE score.
* Impact: How impactful do I expect this test to be?
* Confidence: How sure am I that this test will prove my hypothesis?
* Ease: How easily can I get launch this test?

Rate each of these on a scale of 1 - 10 and take the average to determine the score. Stack-rank ideas by score and implement the idea with the highest score first.

Source: The Practical Advantage Of The ICE Score As A Test Prioritization Framework by GrowthHackers

Implementing Experiments

This is the key part which growth engineering plays - implementing the experiments. This means first setting up the foundation for rapid experimentation then afterwards building features.

Experimentation Tools

There are several experimentation tools that can help with A/B testing or feature flagging. Some of the popular tools are:
* Launch Darkly
* Optimizely
* Rollout
* Field Test (for Rails)

Marketing Automation Tools

There are also several tools to help with marketing automation. Some of these tools are:
* Drip
* Intercom

Analyzing the Results

You've run your experiments and gathered a bunch of data, great! The next step is to analyze the data and see the results. This is where your data team comes in.

Data teams typically take care of two things: maintaining data through data workflows and warehousing and analyzing the data using data visualization or business intelligence (BI) tools.

Data Workflows and Warehousing

Data workflows are used to manage a company's data pipelines and data warehouses are used to pull together all of company's data including CRM, ERP, product analytics, billing, supply chain, etc. These systems are typically managed and implemented by a data engineering team, however, it's important for a growth engineer to understand how a company's data is managed and surfaced into data visualization and BI tools.

• Airflow - programmatically author, schedule, and monitor workflows / data pipelines
• Meltano - data science life-cycle tool: Model, Extract, Load, Transform, Analyze, Notebook, and Orchestrate.
• Snowflake - cloud data warehouse
• BigQuery - Google's cloud data warehouse
• Redshift - Amazon's cloud data warehouse

Data Visualization and BI

These tools are used to help sift through, understand, and visualize data.

• Periscope Data - Data Visualization and BI Platform
• Metabase - Open source BI tool by Expa
• Superset - Enterprise BI tool by Airbnb
• Kibana - analytics and visualization tool by Elasticsearch

Rinse and Repeat

That's it! Running a growth process takes discipline. Rinse and repeat and continuously experiment to find the working growth tactics to scale your product.

More Reading

Here's some extra reading which I found helpful to understanding how to build a growth process:
* High Tempo Testing by Sean Ellis
* How to build a growth machine by Brian Balfour
* Sustainable Product Growth by Sequoia Capital's Data Science team
* Business Lessons about Growth from Andrew Chen
* What are growth teams for and what do they work on from Casey Winters
* Why retention is the silent killer by Brian Balfour
* Why Retention is King of Growth Strategy (talk by Brian Balfour) by Jonathan Crowe at OpenView Partners
* 7% a week growth by Paul Graham
* Don't Let Your North Star Metric Deceive You
* Don’t Become a Victim of One Key Metric
* Product-led growth