By its very nature, account information or transaction data is rich in customer insights but often comes in a highly unstructured format. Operating with this kind of data requires a great transaction categorisation engine.
Here are your options if you have a need for such a product:
build your own transaction categorisation engine;
buy the default engine provided by your account aggregator;
or buy a specialised transaction categorisation engine.
There are benefits to each approach, but the right strategy depends on a multitude of factors. Here's a short checklist to help you along the decision making path:
Time to market
How soon do you want to have a finalised solution that uses categorised account data? This might be the most important question to ask. Building an in-house transaction categorisation engine can take anywhere from several weeks (an engine that identifies a few keywords in transaction data) up to several months (a more sophisticated engine). Additionally, you need to allocate time for constructing the solution on top of categorised data. If timing is important, it's best to use an existing engine.
Do you have the support of your team or are you planning on building the engine on your own? Development hours easily add up and the true cost of building an engine is not just the sum of the hours spent, but also the alternative cost, i.e. what you or your team could be doing instead.
This is a hidden cost of having your own transaction categorisation engine — it needs to be periodically updated because transaction data is constantly changing. The maintenance includes (1) monitoring uncategorised transactions, (2) taking random samples from categorised data for quality control and (3) making engine adjustments. These can be added to development expenses.
Data preparation anxiety
Research suggests that cleaning data is the least enjoyable part of data science. Transaction categorisation is very much that — data cleaning. Are you excited to spend the next few weeks on working out categorisation rules and all the exceptions to those rules? Or are you more excited to work with categorised data to build the solution you need in the first place?
Whichever solution you choose, Nordigen's data experts are always a call or an email away — having built our own transaction categorisation engine from scratch makes us uniquely qualified to help others make the best of their account data. So let's talk: email@example.com
Here’s what we've been up to recently.
A while back we shared a list of 5 fintech podcasts we listen to at Nordigen. This time we wanted to share a shortlist of experts that we follow and frequently read to stay on top of fintech news. While there are many finance and tech professionals in the fintech industry, we've shortlisted our favourite thought leaders that we find to be truly exceptional and engaging.
With Nordigen's API Platform, you can start building data-driven services - powerful apps, onboarding flows and seamless customer experiences today. The platform guides developers and businesses to installing data aggregation, assessing potential risks and onboarding their customers. In this article, we will guide you through 5 easy steps of turning raw account data into valuable insights.