At the beginning of the month we talked about our new Popular Juices sorting, the first iteration to incorporate our user weighting system and have since been running live tests against the data. Today we’re happy to announce the new rating system is LIVE and is used anywhere a juice’s score is shown and when sorting juices by popularity. We expect continual iteration on this algorithm, but the feedback we’ve received points to the results being a closer reflection of the community’s current sentiments than the old method provided.
We also are excited to announce that we have started the testing phase of our new tools for business owners. These tools will let the owners of the fine businesses on JuiceDB update information related to their business, products and deals.
If you’re interested in knowing more, read on as we talk a bit about the new tools for businesses and some conclusions on weighted juice reviews.
We just started the testing phase of our new business management tools to gather some feedback. We are looking for businesses who are willing to experience our tools and then have conversations with us about it’s usability, usefulness and feature set.
In these new sets of tools, verified business owners will be able to:
If you’re interested, feel free to drop us a line and a bit about what parts you’re interested in, email@example.com. We will provide instructions and start up a conversation about what you feel these tools need or what’s just too difficult to do. These will soon be open to the public and anyone who owns a business will be able to participate.
When we initially started talking about this, it was a mixed bag of reactions from the community. It’s most likely that we were unclear about what this actually is. The weighting of a juice review (and it’s rating) changes how it impacts a juice’s overall rating. That’s it! Nothing about the reviews are altered, and if a juice is given a 1, it’s a 1 displayed on that review for the world to see. This change is more about how juices get ratings than what you’re contributing.
So why have we gone through so much trouble just to tweak these numbers? Our goals were as follows:
Every system can be “gamed” or “cheated” somehow, ours included. If you’ve been a part of our community long enough I’m sure you could list of a handful of products that made it on the lists that shouldn’t have been there. Since the start of 2.0 (and ramping up development of JuiceDB) we have been taking a closer look at how these things are being used.
Weighting helps us trust those involved in the review process. It’s not to say anyone’s untrusted, simply that others are trusted more because they’ve proven it with the actions they’ve taken on the site. It’s an automated way we can help reduce future “gaming” without hindering the community’s right to review things.
There have been a lot of strange anomalies on JuiceDB’s juice lists over the years. Here’s a common example of what we saw behind the curtains:
A new juice would popup on the juices list of “Popular this Month”, it had 10+ reviews in the last 10 days. Almost all of the users had written a single review, which was for this product. A few of them had a few more reviews, but they were all products of the same brand. From the admin tools the details make them look like real, different users. It still smells fishy though, who are these people? It’s similar to a reddit thread that appears and a bunch of 2 day old accounts all start talking about how great the subject matter is.
We can’t just remove these users and we can’t say they’re definitively fake and remove them. There’s a number of situations that this could be legit behavior. It seems somewhat likely that a few Brick and Mortar store employees stumble across JuiceDB, think is awesome, and encourages a bunch of regulars to give it a shot. These same employees probably also have favorite juices that they’re selling to customers. Then all of the sudden out of left field a bunch of reviews for a flavor unheard of appear. This kind of stuff happens.
A bunch of reviews for an unknown flavor (if legit) certainly isn’t a bad thing. Some people might want to know if a local flavor in rural nowhere tastes awesome. It just might not belong on the front page of our juices list. This occurrence is a microcosm of opinion, which is awesome, but it’s not the global sentiment.
Currently the juices pages on JuiceDB are meant to be the best that everyone can get their hands on. No one should expect their favorite local flavor included on the best of all time. The team’s favorites sure aren’t, but we recognize that the flavors listed are indeed of quality. Like every other popular review website out there, you also won’t like everything you see, but a large number of people actually will.
We haven’t reached what we’d consider global sentiment by any means, but we’re inching closer. We have a number of steps in mind that we plan on taking to further refine our juice ratings system and expand it into even more useful tools. The goal is to have a website with an opinion of it’s own that closely matches the community as a whole, but never one individual or limited group.
It’s a lofty goal, one that will take a lot time.