Combining science with practicality in segmentation
How do you categorize the app on your mobile phone? Is it based on color, function, frequency of use, or something else? Without realizing, you might have done many segmentation exercises in your life. At GoTo Financial, this is how we do it.
GoTo Financial is part of GoTo, Indonesia’s most wide-ranging tech ecosystem that seeks not only to increase the efficiency in the virtual lifestyle of those who already have access to the technology, but also to be the enabler for those who are yet to discover the benefit of the digital economy.
Committing to this mission In Indonesia requires a lot of understanding of the diversity of its people. High income urban residents, for example, will have very different needs from communities living in rural areas where 4G is still a dream.
This is where segmentation becomes crucial: it helps us prioritize and put resources on problems that matter the most, and focus on the customers with the most potential to benefit from these solutions.
External vs internal data segmentation: A dichotomy?
The GoTo ecosystem boasts an extensive consumer products portfolio, ranging from on-demand transport, food & beverage, groceries, eCommerce (Tokopedia), online payments, offline payments, up to financial services such as PayLater. We have abundant data points on how people move, spend, all the way to their lifestyle choices - we’re a paradise for the data freaks of the world !
To work on the vast amount of data points, our research team consistently engages those considered as stakeholders (customers, partners, merchants, non users) in order to know who they are and how we as a company can fulfill their needs and expectations.
We collect insight through primary and secondary research projects, the collection methodology as well as the benefits and disadvantages of each laid out in the following table:
|External Data||Internal Data|
|Collected via survey (qualitative & quantitative)||Collected via recorded app utilization & transaction activities, quantitative|
|Can be influenced by perception, memories||Real data based on usage|
|Time lag between occurrence and data collection||Real time data collection|
|Can obtain emotional / psychological aspect||Unable to record the emotional / psychological aspect|
|Can obtain the motivation behind the action||Unable to record the motivation|
|Can obtain insight about behavior outside the recorded apps||Only able to record activities within the apps|
|Can obtain other supporting profiling data||Profile will depend on registration requirements (e.g: KYC process)|
|Enable to record behavior / activities in other apps for benchmark||Can only do historical benchmark|
|Usually sampling-based, higher margin of error||Population-based, higher accuracy|
|More difficult to take action for targeting||Easier to take action for targeting|
|Costly & timely to collect data||Relatively easy to collect data|
The main tension is connecting those two different data sources. In talking about segmentation, many companies decide to take the easy way and only focus on internal data. This is when you still have much to analyze from your existing consumer base.
However, in focusing too much on internal data, companies are missing the “what other opportunities are out there” data points. TWithout seeing what is outside of your home, you won’t know how big the world is.
GoTo Financial rarely takes the easy path. Where we find something that is broken, we commit to fixing it. Where we see untapped resources, we mine it and turn it into something valuable. We combine the vastness of our internal data to understand our customers better with the new insight from external data to figure out how we can serve the untapped market better.
Practicality vs Science
Then how should we process that data? How can we identify and segment our customers? How can we find which customers are influencers to their friends?
That’s where a strong collaboration amongst the Business Team, Data Science (DS) team, Business Intelligence (BI) team, and Research team are needed. At GoTo Financial, we’re lucky to have people with expertise in all these areas. In brief, what we do are:
- Research team will do the data collection through surveys to complement the internal data.
- Based on the results, the BI team does multiple variables assessment and statistical modelling.
- BI and Growth teams do experiments to test the validity & reliability of the model
- Then together, we iterate until we feel comfortable to utilize the data in different campaign executions.
- Then this cycle repeats periodically to ensure we capture and address the latest trends.
With a wealth of data points, is it easier to segment our customers and figure them out? It is an ongoing process. Customer behaviour and preferences can change too. But by having a model to make predictions about our customers, we can start personalizing campaigns to people’s specific needs. And this is where science meets practicality at GoTo Financial.
There are so much more to explore for the benefit of all. Check this link to join our Community of Makers and be part of this change for good movement!