How to Drive Growth Through Experimentation.

At last week’s growth community group at Fishburners, Expert in Residence and Managing Director at ikaros, Daniel Lohrmann was in conversation with Justin Bohlmann, the Head of CX and Conversion at leading fintech scale-up uno Home Loans.

With a diverse career spanning the Australian Army, media, publishing, lead-gen and e-commerce, Justin brings a contagious flair to the topics of growth and experimentation, with a relentless drive to test and optimise everything; from the way he talks, to internal processes, to growth experiments across the funnel.

1 – CX and UX are interchangeable, just focus on optimising your customer experience

Justin stressed the need to consider the customer experience from end-to-end; to guide them to make a decision and overcome objections at all touchpoints, from when they first see the ad, to when they’re on the landing page, and through every stage of your funnel.

2 – Against big competitors, you need out-of-the-box ways to acquire customers

For uno Home Loans, traditional acquisition channels were incredibly expensive (as they were bidding against big banks). They needed a growth channel that was digital, scalable, and that portrayed their core USP of transparency. From this, an internal product loanScore was born (a free ‘check-up’ tool to see if your home loan is in good shape). 

This was built in a relatively traditional framework: dedicated sprints for development, wireframes + mock-ups, testing and feedback, and MVP launch. Once optimised, the live version of this product ‘hack’ led to incredible product-led growth, significantly out-performing other expensive acquisition channels for a fraction of the cost.

3 – Set yourself up with the right tools

Justin explained that experimentation requires a robust marketing and analytics stack. His recommendations? Optimizely/Google optimise for A/B testing, Amplitude for product analytics, and Segment to pipe your data through all tools as a single source of truth.

4 – Experiments, big bangs and minimum valuable products

Justin went on to explain that when it comes to experimentation, you need to separate narrower experiments (e.g. focusing on specific conversion elements within the funnel), with ‘big bang’ experiments like loanScore that are best described as ‘high-risk, high-reward’ (e.g. hasn’t been done before, requires heavy, cross-team resource allocation). Across both of these experiments, he provided a set of practical guidelines.

Growth Experimentation 

Justin described these as smaller day-to-day experiments. Justin runs this with two teams, acquisition and CX. He starts with an ideation and prioritisation stage, with experiments listed out in a spreadsheet and defined using RICE methodology. Everyone across the business is encouraged to submit experiments, though prioritisation is handled by the core growth team. Uno’s robust analytics stack means they have a good handle on their funnel. There is already clear visibility on where the biggest drop offs are and where their biggest opportunities lie, so they can focus on those areas with experimentation. Each of the experiments in the spreadsheet has a problem statement, an objective (e.g. improve conversion from home page visit to account) and a primary metric (e.g. visit to account).

As the conversion funnel is very specific, he highlighted the need to get your primary metrics right, but also to look beyond it when you need to; you need to look at your leading indicators as well. For example, in his early days with uno, his focus was acquiring as many accounts as possible. One of his core avenues was targeting people who wanted to know how much they could borrow. The data quickly revealed, however, that people who wanted to know how much they could borrow didn’t necessarily want a loan. 

An entire dashboard for the experimentation process exists in an internal confluence board, including statements and objectives from the ideation spreadsheet, whether the experiment was a success or failure, optimizely links (for A/B tests), screenshots, and experiment reports from Amplitude. In this sense, Justin has an entire dashboard & database of past experiments, so that his new and current team members can go through and check what’s already been done. Finally, a growth channel exists in slack, allowing company-wide visibility on what tests are being run and whether they’re successful or not. 

Big Bangs

For all types of companies, but in particular younger start-ups, Justin recommended focusing on Big Bangs every quarter; putting time aside for design sprints where you can chase particularly bold experiments, strategies or products that you hypothesise can have astronomical impacts on your growth. 

For smaller start-ups with less visitors, minor optimisations will make little difference. He stressed that experiments should be bold; e.g. instead of changing a button colour on the home page, you should be testing three wildly different versions of the home-page with unique messaging. 

For product-led Big Bangs, he recommended following a typical MVP process (mock-ups, interviews, build, test, optimise) but to make sure you build a minimum valuable product, not a minimum viable product, as you aren’t going to make a big bang if you aren’t impressing users from the get-go.

5 – Justin’s key takeaway:

Stick to a process, be organised, document everything you do, be bold and brave and test the big ideas, not the colour of a button. 

Leave a Reply

Your email address will not be published. Required fields are marked *