How Fave built and scaled a million dollar wheel spin tool

We’re a team of hackers.

We like to run experiments.

We love to build things.

We pull the plug if a feature doesn’t work, invest and scale if it does.

That’s pretty much how our million dollar spin to win feature begun.

Conception

It started with an idea from our Growth Planner, Aaron Tau. Our deal users love excitement, campaigns, games. Could we build a games tool for users to drive deals growth for the upcoming festive season?

The 3 of us (including Edison, our then newly minted software dev) white boarded the concept in an hour. Version 1 was to hardcode the bulk of the designs, flows, logics. The premise was simple, a sticky animated icon on the home page, which linked to a page to spin a wheel to win promo codes. Level 1 was to validate if users would play, win a promo code, and be incentivised to subsequently use it. Engagement + conversion (two of our key success metrics).

Mini gamification of winning promo codes

The Dirty Hack

Rolling out version 1 wasn’t easy. We had to consider many use cases. For example, we had various permutations of promo codes that could be won — deals, pay, categories, cashback, discount, max caps etc. We also wanted to build in the concept of variable rewards and allow for winning probabilities without overcomplicating the architecture. From whiteboard to production, the process took us a few weeks.

We’d be lying if we said V1 went without a hitch. We probably rolled back at least twice. Fortunately, we were forward thinking enough to ensure all our key features were managed using feature flags. So we could easily switch them off without going through our standard review and deployment process.

Validation

Despite the glitches and bugs, we had over 130,000 spins over a 2 week period. Our hackish-hardcoded-unrefined wheel spin tool had contributed to over 27,000 vouchers sold during the period. Not too shabs.

Scaling

We ran this wheel spin “hardcoded style” with similar results for a couple more campaigns til it became too time consuming to support the various country demands. A problem! But a good one. That’s when we decided to automate — hardcoding had to go!

Over the next few weeks our engineering team mapped out a framework for our mini games tool. Everything would eventually be customisable by our campaign managers. We’d be able to run up wheel spin campaigns in under 5 minutes as opposed to weeks. The framework had to be scalable, flexible to support custom designs, timings, marketing copies, and campaign managers could apply strategies to drive retention and cross-sell whilst keeping within budget. We would be completely hands off going forward. A dream.

Our game page and “AngPao” layout (frameworks courtesy of Loh Jia Jing, aspiring rockstar)

We did it! But 1% Done

Our teams were pretty happy with the results. We hacked it out the first couple of rounds to validate then tweak and optimise the process. Building scalable tools meant the team could then scale further on other projects, to drive more growth.

Our games tool was repurposed for key global events to drive excitement and traffic. Jia Jing and Edison worked long nights to make a baby called FaveCup.

Since then we’ve been trying to run projects in a similar manner — hacking the quick and dirty, validating our MVP, invest and scale for growth if succeed/ discard or pivot if failed.

Its not all sunshine and roses with driving new features and experiments, but what’s key is the process of learning, validating and growth — not just for Fave as a company, but for the individuals within our team.

Here’s to 10X and beyond.

(Team credits: Jia Jing, Edison, Azuan, Asad, Irene)