Hey @RareAkuma,
Thanks for your thoughts. Here is some background information on the experiment: the worry behind this thread is one I share. I personally rate everything on the full slider with notes, aromas and most of the time in-depth comments. The 100-point system (0.1-10.0) is not going anywhere and I’d be the first to push back if it ever was.
But here’s the part you don’t see from the outside. When we looked at the numbers, only a small share of our active users had ever submitted a single rating. Not “less than I’d like”, a fraction. The old form was working great for power-users of the app, and for me, and at the same time it was a wall for almost everybody else. A new user opens a rum and cannot decide whether it is a 7.3 or a 7.9. They don’t taste like that yet. So the form scared them off and the rating never happened. For years. This insights from the data analysis was also confirmed in many survey interviews with non-power users.
That’s the problem the 4-button row tries to solve. It’s not “RumX simplified”. It’s “RumX has a way in for the user who isn’t there yet”. The four buttons are not a separate scale either. They map to the labels the 100-point scale has built in (Bad / Okay / Good / Very Good are real anchor points on our scale, that’s how the descriptors work. Full background HERE). So a quick rating is the same scale, just picked at four anchors instead of all hundred.
Some things you should know:
- Every rating in our system is tagged with where it came from. Full tasting, quick rate, onboarding flow, imported, all separately marked. That means we can treat them differently anywhere we want. First place we did this was the feeds. After @Boletus pointed out that quick ratings were clogging the Global feed, we filtered them out within a day. He was right and it was the obvious call.
- How quick ratings flow into the global score of a rum is a question I’m actively sitting on. I don’t want to announce a formula before I see how the system actually behaves over a longer window. Too early, too many assumptions. What I can tell you: the option is open. Flag them visually, weight them lower, leave them out of the average entirely, all of those are on the table because the data is tagged for it.
- The full tasting form is unchanged. Slider, aroma tags, notes, photos, all of it. The quick row is added on top, not in place of.
What I’d ask from you, since you’re the people who actually understand how a rum score should feel:
Pay attention over the next few weeks to a rum you know well. If a score starts feeling off in a way you can pin on quick ratings, post the bottle here. Concrete rum, concrete current score, concrete reason it feels wrong to you. That’s the evidence I need to decide the aggregation question properly. “Scores will inflate” I already worry about, same as you. What I am missing is the specific bottle that proves it. 