A Losing Streak
Why Duolingo always sets me up to fail. (And how it can not do that.)
In 2018, I lost my 50-day long German Duolingo streak. I was so disappointed, I never picked up the app again. In the past month, I’ve been interested in learning German again. Today, I lost my 25-day long Duolingo streak. Here’s why I’m still going to continue, this time.
What’s wrong?
There are two methods that Duolingo employs to notify the user of their habit. Neither of these methods seem very effective.
Firstly, their graph of the past week is inevitably dotted with crests and troughs. This is a disappointing portrayal of habit. A horizontal graph is nothing to aim for either; it has been drilled into our minds that graphs must go upward. This is silly, since learning a language requires consistency rather than gamification in terms of increasing the points from the previous session.
Secondly, no matter my usual consistency with the language lessons, I am spared no thought regarding my regularity, and immediately thrown back into the initial state that I’d entered, and that which all other newbies and casual learners inhabit – the 1 day streak. This is extremely disheartening, to say the very least.
What really matters?
The bigger picture. I want to learn a language, after all, and not win a competition. Duolingo should reward the habit and not the streak. It should not psychologically punish the occasional slip of mind or extremely busy day.
After all, the longer the streak, the worse you’d feel to lose it. What is the end game anyway? Bragging rights? A Guinness Record? …Death?
And no, introducing a ‘Streak Freeze’ is not the perfect solution to this, it is only a temporary dupe remedy, which highlights the irritating nature of their streaks, almost as if they understand the annoyance. It also imparts a false sense of maintaining a streak.
What’s the remedy?
A fitness website I’ve used – Fitness Blender – does this portrayal better. Working out is also like learning a language; it requires the same level of dedication for the payoff, and the results would show neither from a single day’s intensity nor from nothing but a daily grind with no skips in between.
Fitness Blender displays the consistency throughout the month. This is satisfying to see and it gives a better idea of how much we’ve worked towards the skill in the recent past, as a visual texture.
So what am I suggesting?
I created my own version of this, as a conceptual visual. Here is a calendar month of 31 days, on which there rest 31 yellow circles. These circles represent every day the user has completed their set target (these could be points, or a number of new words/micro-skills learnt.)
However, this is an ideal scenario. Accounting for hiccups, we could suggest that a (hardworking) regular user may have a calendar that looks like this.
Now imagine we are more watchful of the amount of work completed in the day, as well. Rather than a binary *worked* and *not worked*, we could have a system that shows the amount of work the user has done, relative to their day’s target. This can be done by decreasing the opacity of the yellow circle. We can call this brightness the ‘luminance’. Technically, there can be three states of luminance — 40%, 70% and 100%, depending on the work done that day. And so, we have an updated work calendar that looks like this.
But you wanna brag!
Duolingo has a tiny vanity mark for any public interactions, where you can see, along with the languages being learnt, the streak of the user that is commenting. Check out this user’s streak!
It’s possible some hyper-diligent users may want to keep this feature. For this, I have a different solution. Rather than show the current streak that the user is on, without any reverence for their history on the app, the app could display the total luminance of the user since the beginning of their journey. For example, If they have missed a hundred of the thousand days they’ve been on the app, the luminance is still 9.0.
For a visual example of this, see the next section.
How would this all look?
Here’s a couple of bare bone examples of such an interface.
So are streaks just useless?
Streaks are cool. They’re fun. I think they’re an interesting edition for an app like Snapchat; nothing is really lost if the user quits the app out of rage.
But with a language app like Duolingo, it might be better to focus on the learning aspect rather than the app itself. It is definitely a problem if the interface of the app hinders the learning rather than accelerates it.
