How Itaú Unibanco used Experimentation In It’s Digital Acquisition Strategy
In Digital Marketing, the experimental process is not different from other fields of knowledge. Observation, hypothesizing, and testing are some of the tasks involved and can be beneficial to the banking industry.
In the 6th edition of the GrowthHackers Conference, we welcomed Robson Harada, Head of Growth Marketing at Itaú Unibanco.
This Brazilian bank has 50 million clients in over 18 countries. Beyond growth, client centricity, digital transformation, and efficiency are other priorities in its current strategic agenda.
Harada presented a case where, after growth experiments, he and his teams got a +67% conversion increase of high LTV clients in six months.
Keep reading if you want to know how they achieved this!
Financial market competition, at least in Brazil, has clearly been boosted in the last few years by factors such as:
- The rise of fintech.
- Other businesses providing financial services, such as Retail Companies.
- More strategic agendas, like more payment methods, Open Banking and Open Finance.
That’s why digital growth appears as a solution to face this challenge. And more specifically, the digital acquisition strategy.
There is a belief, in the banking industry, that “digital isn’t able to bring high-value clients compared to physical channels”.
That’s because the manager’s presence is highly persuasive. Generally, professionals in this field are trained to sell more products, deal with people, negotiate, and so on.
In spite of that, Robson investigated how to improve the quality of checking accounts by utilizing digital acquisition campaigns.
By testing and earning — knowledge, experience, and results — Harada proposed the following steps:
Let’s say you’re analyzing two CPAs (Cost Per Action) campaigns to find out what has a higher performance.
- CPA: 5
- ROAS (Return On Ad Spend): 20:1
You probably would keep Campaign A, right?
But the question is: are you really using the right metric to measure your growth strategy?
Assume, for example, that, by accessing more data, you discover that Campaign A had 10 conversions, with $950 of profit. Meanwhile, Campaign B obtained 30 conversions and $ 2700 of total profit.
So, campaign B performed better. That’s why choosing the right metric is so important.
Anyway, for Harada’s team at Itaú, the chosen metric for checking accounts was Offline-Conversion Cluster A.
This “Cluster A”, by the way, is a specific profile of a prospect — with a high credit score and a high predictive LTV.
And why offline conversion? This is because, in the traditional checking account journey, clients are separated by credit scores, predictive LTV, etc. These are the data extracted from proposals sent by leads who are interested in opening an account in Itaú.
At first, Robson’s team was setting up and managing the outcomes of multiple experiments manually, using an elaborate spreadsheet. But, as the process moves on, the document got more and more complex and time-consuming.
A better way to do this, according to Harada, is to use a platform built for the need of your experimentation team, like Experiments by GrowthHackers. He highlighted a significant difference using the software: easier processes and more clear and focused discussions about tests.
In this step, the marketer and his colleagues used three pillars to optimize Offline-Conversion Cluster A:
RTB (Reasons to Believe)
They tested different creatives, with distinct arguments (like strengths, benefits, and positioning) of Itaú.
Instead of publishing more and better copies, the team tested posts and ads that were more connected with high LTV clients — the right metric.
In this case, the message was about the benefits of an Itaú account, like the convenience, security, tradition, diverse portfolio of services, and other drivers.
Here, tests were performed with offline conversion and data integration. Basically, the algorithm was instructed to “give more” Cluster A prospects — the North Star Metric of the campaign.
But Robson points out that there might be other possibilities for your company besides paid media, such as inbound marketing, referral programs, omnichannel approach, etc.
The app flow was the main channel.
But they were forced, by Apple, to move to the web channel. So they decided to shut down the Itaú app on both devices, Google Play and App Store. For a smooth transition, it was used the PWA (Progressive Web Apps) method to emulate an app.
In simple terms, a PWA is an app that runs directly in the mobile browser.
The hypothesis of the experiment was: that digital marketing can bring in clients with the same — or even better — effectiveness than offline methods. And, finally, here are the results collected:
- Media budget optimization.
- Sharp bidding according to the predictive LTV of each cluster.
- More strategic use of Machine Learning and Artificial Intelligence in looking for the right and the highest value client.
- Increased quality of the acquisition, since the algorithms were instructed about who is the right prospect.
Among the benefits of the six-month experiment was an increase of 67% in conversions for cluster A, a duplicated ROI based on the predictive profit, and the use of digital channels as the main acquisition source in volume.
Robson ends his presentation at the conference with valuable learnings, which are:
1% Finished Mindset is Mandatory
Culture of Experimentation is A No End Journey
This mindset should lead us to be discontent and to always look for experiments that will help us reach our North Star.
Most of Your Experiments Will Fail
Since you learn — and earn — failing it’s fine.
Be Obsessive to Figure Out the Right Metric
It is always important to remember that the right metric will guide your growth marketing efforts.
Set the Best Cost, Not the Cheapest to Compete Boldly
Most of the time, money is essential. When you are in a very competitive landscape, economizing will probably allow someone to do the opposite — invest more and fill the gaps you left behind.
Competition Lives on Both Sides: Outside and Inside
Even though we must keep our eye on our competitors outside the arena, it is preferable to also understand what barriers they are facing. So, on behalf of a common good, you can engage everyone in the same direction.
Overall, having an experimentation process makes a huge difference, even if the product has historically performed better offline. But, to really enjoy this journey, is recommended to have the right tools to get the results and, in this case, reach bank innovation
How is your company managing experimentation to achieve business growth? Share your experience in the comments!