“I’m probably on the other side of the party, where I think data answers everything,” says Willix Halim, Senior Vice President of Growth, Freelancer.com, in a chat with this author.
Halim might even trust data more than his own intuition. “I believe that even things as vague as user experience, the data would tell you how it is.”
“If you can’t A/B test things,” he adds, “then there’s really no point of building those features.”
A/B testing allows developers or marketers to compare two versions of a particular interface by showing two variants — in this case, A and B — to visitors at the same time.
One set of visitors would see Page A, while the another set would see Page B.
This test checks for conversion, a sign that visitors are doing what the team wants them to do; and if Page A has a better conversion rate, which means that there are more people purchasing items or booking flights — or in Freelancer.com’s case, posting jobs — than Page B, the team might launch the former instead of the latter.
“Otherwise, how would we know if it works or not?” asks Halim.
Freelancer.com also spends a lot of time on machine learning and funnel optimisations — jargon that really mean “using algorithms and data to predict outcomes” and “improving conversion rates” respectively.
The team tests all important pages — from the home page where all users go to every time they sign in to the page users visit to post a project or browse jobs — three or four times a week to make sure that users get the best experience possible surfing the marketplace.
What are you sacrificing?
My hypothesis was that when Halim and his team prioritises growth, they are bound to sacrifice something else, even if the sacrifice came as a by-product.
He tells this author, “If you do it in a balanced way, then you’re not sacrificing a lot of stuff. Yes, a lot of stuff that you will be doing are sort of directed towards short term growth, in terms of funnel optimisation and so far it has resulted [in you] sacrificing the long-term strategy.”
But even this need not be the way. “[I]f you actually plan your resources well, then that wouldn’t happen,” he says.
Every team under him consists of one product manager, one data scientist, and three or four software engineers. The product manager has three or four headline metrics to hit by the end of every quarter.
He adds, “When you’re banking on a long-term feature improvement and stuff like that, it’s going to take a lot of time. You can essentially use data to justify that by building an MVP… or hack through it and test it with data before you spend … time delivering it.”
‘SVP for Everything’
Halim might be the Senior Vice President of Growth, but in a startup, his role is more like Senior Vice President of Everything. He oversees many divisions from Engineering to Product Management to Talent Recruitment to Customer Operations.
“For core teams like product management, product engineering, and data science, I still go through the hiring [and] most of them still have to go through me as a last interview to make sure that the culture is there,” he says.
But how does he know if someone’s growth-driven? “Pragmatism,” he says. “Yes, there are a lot of smart people but they have to be pragmatic.”
Employees have to be pragmatic in terms of the opportunities and things they will be able to do at the company. For example, if someone thinking of joining Freelancer.com’s data science team wants to go straight into machine learning instead of working on wrangling and cleaning up data for analysis, that would mean that they do not have a realistic view of the job.
Want to hear more from Willix Halim on the next big trends in the world of data science? He will be speaking at Echelon Thailand this November 26 – 27. Join us as we connect the Mekong region through Bangkok, Thailand in two action-packed days. Learn more about Echelon Thailand here.