Tax filing can be a chore. It’s complicated, time consuming, and people often miss out on refunds they are entitled to. Taxfix aims to change that. It uses tax experts and an intuitive and clear UI to make tax filing better for everyone.

Launch an MVP in a new market, based on extensive user and market research.

Create a seamless experience that encourages uptake of a new brand.

Easy to use and personalised enough that users believe they are getting a superior service over the free government tax filing tool.

A/B iterations on drop off points

One of the biggest challenges the team faced was timing. The tax season in The Netherlands is only 2 months, which makes collecting enough statistically viable data and making data-based iterations super difficult.

Early on in the project we had decided to go with three versions of the same app, with elements such as registration, payment, and tax questions in different places. We soon dropped two of the flows to focus on one.

Following this, the biggest drop of points were at registration, a ‘tax expert’ screen, and payment. Using the data collected from quantitative testing, I worked with the designers and project lead to come up with alternatives. The data also showed that most users preferred to use a mobile, rather than a desktop, so we focused our iterations on a mobile-first design.

Originally, we placed a lot of focus on the ‘tax expert’ and what they could do to help users with their return. However, quantitative surveys showed that users were confused about the role of the expert, didn’t think they needed one for a simple tax return, or thought it would make it too expensive.

As a result of the surveys, I chose to focus more on the steps of the process and how the expert will help at each stage. I also changed the text on the button to be action-orientated.

When the first iteration didn’t yield the desired results, again I focused on the messaging. This time taking the emphasis away from the expert entirely and instead focusing on the process.

Final result:

Unfortunately the MVP did not experience a good uptake from users. This is believed to be because of several factors, including a lack of need for the product and competing with the free government tool.

The Incubator team prides itself on working quickly to investigate the opportunities in new markets. In my time there I wrote the in-app copy for two MVPs (minimum viable products) and two LMTs (live market tests) as part of the product team. An example of the screens from the Mexican LMT is shown below.

Taxfix is a German product. It is built for a German audience, and it does very well in Germany. I was hired by the Incubator team, which led the international expansion efforts, to lead content design for new markets. Part of this included using my skills to come up with tone of voice guidelines that would resonate with the new audiences.

Using quantitative and qualitative research, as well as the brand identity, I wrote four different guidelines for the four markets I worked with - the Netherlands, Brazil, Mexico, and Australia.

In the Dutch market, the research identified that users prefer a formal yet conversational approach. Think more contractions, but less colloquialisms. As tax filing is very personal, the voice and tone I chose conveyed knowledge and guidance, to foster trust, as was serious when it needed to be.

Although each market was getting roughly the same product - a tax filing app - the nuances and culture for each one meant it was necessary to create a strategy. One of the strategies I worked on was for the Australian market, where users were comfortable with more colloquial language but wanted to be reassured their tax return was correct and they wouldn’t incur penalties. The Dutch users, on the other hand, were more comfortable with a higher level of formality.

Localisation is tricky, and at Taxfix it was no different. There was no process so to speak, which meant it was often an afterthought and rushed right before launch. To solve this I worked with the localisation manager to implement a process than worked alongside the agile process already in existence. With some trial and error I successfully introduced a staggered approach to localisation that fit in the 2 week sprint window before content was handed to the developers, thus solving what had been causing holdups and frustration.