Customer feedback will allow us to direct our attention to what our current users want improved. Using an in-app survey, I got feedback from users directly.
Copying my automation portfolio completely to backtest my current strategy
Better automated portfolio/strategy testing, better date handling (mark/exclude invalid tokens, etc).
advanced calculations. For example: what would the test look like if I had DCA'd $1000 a month to the index?
Inform me why backtesting failed or cannot be performed!
It is too slow
I just wish it worked
Can’t backtest coins before they were listed on an exchange
Doesn’t work / show results
Unclear how to use it
Can’t search assets
Backtesting is slow to set up / use
Can’t build portfolio from backtest settings
Over half of our users are performing backtests, and with this much contact, it is important that they have a good experience. It is clear that the tool was too confusing, too slow, and key features were missing.
In remote usability testing of new users, it was clear that the tool was very unorganized. It was unclear where to start in the process, what some actions like "copy portfolio" and "fee optimization" did, and what items like "shrimpy total" and "holding total" were referring to.
Another problem is that we had two different backtesting tools that functioned very differently, and produced different results. Let's take a look at the previous designs.
It was clear, our backtesting tool was too complicated for our users, and it is easy to see why. These tools suffer from information overload. With the in-app tool (left), visual hierarchy and layout is confusing. How does this tool work? What do you do first to get started? Because the (english speaking) eye moves from right to left, the first item a user encounters is the graph with sample data loaded into it. Where did this data come from? Moving further left, the ctas are not clear. What do I need to do to achieve my goal?
The marketing backtest tool is organized in a better way. It outlines the steps and is more easy to follow. However, it is so simple that it is ineffective for our users. They can not select rebalancing timelines or backtest with fee variables. Also, there were a limited number of assets to choose from.
As the developers got to work on the speed and asset selection on the backend, I got to work on creating wireframes that made the backtest tool a lot easier to use.