@Cindy237 @Julie394 @Kelly149 @Linda108
I’ll caveat this with a, I don’t work for ABB but I know a little about computers...
Firstly, ABB CS won’t be able to answer your questions because the logic built in to the system will be completely unknown to them. They don’t know the answer, and realistically, they probably don’t even know who does know the answer. Generally speaking though...
The search algorithms and listing prioritisation will be designed to be “smart”. That is, they will adjust rankings based on what it thinks will maximise that person’s experience.
If you have a single outlying review, I had 1 bad retaliatory review out of 130, the algorithm buries it at the bottom. If it was 1 out of say 15, it may not do that as it determines that 1/15 is more relevant than 1/130. It determines that future guests don’t want to see the outliers - if the others are more helpful it will promote those.
In terms of searching, if your IP address or device with cookies or whatever, always searches for your own listing (maybe you’re checking how it looks), then it remembers and puts your listing first as it assumes that’s what you want - even if that particular time you wanted to check out the local competition. However, if someone else arrives to your local area, and they always look for “whole places run by Superhosts” whether they apply a filter or not there will be an element of content management to promote the places it thinks they want to see.
The same process happens in search engines. Try searching a bunch of articles about a particular tv franchise, and eventually when you search for a famous actor their cameo role in that franchise will come up before their bigger roles in others because it assumes that’s what you mean.
It’s the way computer designers try to make platforms improve user experience because each of us searches slightly differently and wants different answers. They are simply trying to give what they think are the best possible answers to everyone who uses them.
Regarding negative guest reviews, it could be argued that host-guest conflicts reduce repeat business, therefore it makes business sense to promote listings that it perceives to have more amenable hosts, and it helps those more tolerant hosts get more bookings too. It wants to try to manage out hosts who are ‘bad for business’ and that’s fair. But there is a massive HOWEVER...
What if by pure bad luck you get three bad guests in a row? What happens long term with this policy? Are hosts being incentivised to give bad guests good ratings and therefore perpetuating the problem?
Algorithms can’t read you mind, and they are only as good as their designers, so it’s understood they aren’t perfect but that they hit more than they miss. Unfortunately, without knowing the logic it’s hard for us hosts to know what’s expected of us, and it’s inevitable that sometimes we will be hurt by them.
I’m not really sure what to do to solve the problem though, short of asking @Stephanie to find the right team and feed back scenarios where it doesn’t work so they can improve the programming.
