Freelancer.com’s Rainmaker Willix Halim on the Key Ingredients to Optimized Growth
“Live and breathe hiring people smarter than you.” That was one of many nuggets of wisdom from Willix Halim, Senior Vice President of Growth at Freelancer.com.
In his SydStart presentation - 17 Million Users Through Data Science and Funnel Optimization - That advice became even more memorable when he added a tinge of humour to it saying, “...because I want to play on Facebook all day and all my guys will get 100% growth.”
Hiring the best has proven to be a highly effective strategy since Freelancer’s growth team, under Willix’s leadership, has been one of the company’s main drivers for reaching 17 million users. In his meaty presentation, Willix discussed data science, funnel optimization and several other key principles to achieving exponential growth for startups.
Join us for a valuable fireside chat with Willix next month
Avoid the death spiral of paid marketing
The death spiral is earning less from your customers than the amount spent on paid marketing to acquire those customers. It simply looks like this:
Lifetime Value (LTV) < Cost Per Acquisition (CPA)
Willix advised startups to avoid this like the plague because instead of achieving growth, your startup will eventually just run out of money.
Fun fact: Freelancer only uses 15% of its revenue on marketing while some unicorns use between 80% and over 100%.
Instead of throwing away heaps of money on a poor marketing strategy, Willix said your growth should come from two things:
1. Data science - the process of extracting insights from data.
2. Funnel optimization - consists of everything from the moment a user goes to the website until they get what they want from the website.
Willix also enumerated Freelancer’s rules of thumb when it comes to growth:
First Rule: Don’t be most companies!!!
The initial step to being novel is to hire the best people, the creme de la creme to build an A-team. “Don’t compromise talent.” Willix emphasized.
Fun Fact: Fifty percent of Freelancer’s data science team have PhDs - in the real sense and it also stands for Passionate, Hungry, and Driven.
Next, Willix pointed out that companies only track three things: traffic, users and revenues. His advice for startups is to track everything. When something goes down, you will be able to quickly pinpoint why.
Track your competitors and your funnel. To effectively track your funnel, Willix introduced AARRR, or the Pirate Metrics.
A - acquisition. How do you get people to your website?
A - activation. Do they sign up as a user?
R - revenue. Do they make money for us?
R - retention. Will they post a second project? (Or use our services again?)
R - referral. Will they invite other people to the site?
R - ressurection. If they don’t use us anymore, will they come if we resurrect them after 6 months of dormancy?
You also have to A/B test everything. With this, you get to test correlation and causation. When new features are built, you will know the success and failure metrics after running an A/B test.
Second Rule: Conversion rate is everything and everywhere.
Every page or screen in your website should have at least one purpose or call to action (CTA). That’s why you should plot your pages’ conversion rate.
Third Rule: Be fast!!!
Optimize the speed it takes to perform CTA. When startups work on making this happen fast, the conversion rate will go up and everything else will go up. One thing Willix encouraged startups to take to heart, “Being fast is always good.” Make it a habit.
Willix discussed the growth sprint where product teams are put into a room for a period of time and their work will focus on achieving growth.
These are the rules of growth sprint:
1. There will be 3 sprints a day for 15 days.
2. Each sprint is a minimum of ONE growth experiment.
3. Each experiment card will show [EXPERIMENT] in black, and [RESULT] in red. ONLY THE SPRINT MASTER CAN ADD CARDS TO THE SPRINT BOARD.*
Fourth Rule: Get the machines to help you.
Willix explained the concept of machine learning, which is the study of artificial intelligence to recognize patterns and understand data to predict from it. With machine learning, you can quickly detect fraud or predict if a code is going to be effective and in turn yield profit.