Algorithm Update – Part I
See part II here – Score Scale Update
To learn more about the current algorithm, check this out.
ResellerRatings pioneered the rating algorithm space by being one of the first companies to not post strict averages as the rating score. That is, our score curve worked on evaluating the current shoppers experience buying from that store, while also taking into account reviews from the past.
One of the most critical things the score needs to represent is ‘If I were to buy from this store right now, what would be my expected experience’. It’s often hard to wrap that into a singular score, which is why we have authentic reviews, verified content, and provide other indicators such as percentage of positive reviews and historical ratings for shoppers to explore.
Over time, as we collect hundreds of thousands of reviews, we run into new situations. Edge cases, weighing in lifetime reviews, vs current scores. Consider the following – should a merchant who has had bad reviews in the past be forever punished, when they have worked on improving their experience? How do you represent that at a glance? Should a merchant who stopped collecting reviews with us, score stay frozen if no more reviews are collected – but what happens to reviews as they age and a new more recent few scores come in that are the polar opposite score? The algorithm curve works on all these situations and we often run models on tweaks based intense internal discussions.
Today, we are working through one of our largest algorithm updates in recent history. We like to give reviewers and merchants alike a heads up. As we prepare to announce more regarding the algorithm update, simply put we will be notifying retailers of how their scores may be affected.
If you have any questions, let us know.
We’re excited about these changes and believe they represent a better score that is easier to understand and better represents the experience. We will provide more details on the changes in the coming weeks.
Grammarly premium features upgrade
Why users still use to read news papers when in this technological globe all is existing on web?
my web blog: Grammarly premium features upgrade