12 votes

Ethoscore beta round two - and thank you!

For anyone that missed round one, the Ethoscore Project is a conscious consumerism platform that's in invite only beta. Invites are only available to Tildes users. Reply to the top level comment below for an invite.

There have been a lot of updates since my first post, many of which are thanks to feedback from Tildoes. I'm so grateful for all of the feedback, it was incredibly helpful.

Here are some highlights:

  • New scoring UI. Quite a few people commented about scoring, one of my takeaways was that the old display gave people the impression that scoring was meant to be, essentially, perfect. Another takeaway was that scoring kinda took center stage and obscured the rest of the functionality when, from my perspective, scoring isn't the most interesting part. I'm not sure how much the changes will help, we'll see! Thanks to: @MimicSquid @Autoxidation @Kilcundas @fnulare @atomicshoreline for scoring related comments and feedback. See below for more notes about scoring.
  • Company types i18n. @Lia pointed out that the company types were US-centric (public, private, etc.). I decided the best way to solve that was to run the various company type designations through a selection of pre-AI global english text corpuses, including GloWbe. It turned out that the types were already pretty close. 'Public' company won globally at 52% usage, so no need for a change there. The runner up was 'listed' at 33%. Similar for 'private' which won at 92%. 'Cooperative' and 'nonprofit' also won. The one designation that needed an update was 'government', which I've changed to 'state-owned' which came in first at 62%.
  • Company locations. Locations now include the country in all cases rather than just some. Thanks to @xothist for the feedback.
  • Companies page table sort. I've expanded this in various ways including: Added 'employees' to the custom sort column so that it's possible to sort by an approximation of company size. Thanks to @fnulare
  • Companies page table. Filters now include tags and company type. Also added a tags column (when screen size allows). Thanks to @Lia and @Kilcundas
  • Company profile pages. Headquarters location now shows up earlier (thanks @xothist). There's now a 'lists' stub that shows all lists a company show up on (thanks @CannibalisticApple).
  • Homepage trending companies. The display now includes lists and tags. Thanks to @Lia and @Kilcundas and everyone who commented on scoring.
  • New scoring impacts. This is ongoing, researching scoring impacts takes time and a lot of companies continue to lack full scoring impacts. In some cases the baseline scoring gets it right without impacts, in a lot of cases it doesn't. That's ok! Scores will keep improving, especially when users start to contribute scoring impacts. And scores aren't everything, lists can do all of the work for many usecases without scoring mattering very much at all. Maybe a couple dozen more companies (along with 100's of brands) now have full scoring impacts since round 1, including all companies on the Champions and Villians lists. Thanks to @atomicshoreline @fnulare and @Autoxidation for comments about scores that looked wrong.
  • New companies added. This is also ongoing, thanks to @Lia for the new company suggestion.
  • B Corp certification now positively impacts all metrics instead of just ethics. This one wasn't based on specific feedback, just the dawning realization that b corps are consistently better than the alternatives in every case I've researched so far.
  • Local Vaultwarden calls no longer trip bot protection. Thanks to @F13
  • Mobile safari bugfix. Thanks to @umbrae
  • Metric badge clicks on company pages now link to impacts details when available. Thanks to @user who didn't make it into my notes!
  • 'Still in Russia' tag added. The tag (and a companion list) covers all companies in the database that are still doing business in russia. Thanks to @crulife
  • Opt in email updates. Thanks to @Kilcundas. You didn't suggest it, but your comments reminded me I hadn't done it yet. RSS updates are on the list!
  • Barcode scanning improvements. Right now there are about 3 million barcodes covered, with fallbacks for codes that aren't.

There's also a new blog post: Fossil fuel against the world (you'll need an invite read it, still trying to avoid bot traffic for now). This post is based on the research for the Climate Deception and Climate Policy Obstruction tags which I did just before round 1. The goal is to make it easier to identify and avoid companies that have worked against sane climate policy and action. You might be interested in these @crulife

Thanks also to everyone else who participated by testing, commenting or DMing, including: @Eric_the_Cerise @ThrowdoBaggins @phoenixrises @SleventhTower @Mikie @ajwish @jmpavlec @ilTofa @Sycamore @vord @Drynyn @ogre @polle @Deely

Sorry if I missed anyone, there were a lot of signups from invite links I posted in the thread before I realized that was drawing a lot of bot traffic. Any of those who never commented or messaged are probably missing.

Finally, a note about scoring: If scores didn't exist at all, I believe the app would still serve the purpose of giving people an easier way to choose companies to support and avoid via the lists and tags features. Especially once there's a community creating and sharing lists. Or I guess at this point, if there's a community. I've spent a lot of time thinking about the scoring system since round 1 because for some users it was more distracting than useful. Tweaking the user experience so that the purpose and functionality of the app is more intuitive is still a work in progress.

As before all feedback is welcome, good or bad. And even just playing with the app without commenting at all helps. Seeing how people interact with it is invaluable. In round 1 there was exactly one barcode scan, so if anyone feels inspired to scan a few barcodes you'll have the undying gratitude of an internet stranger. Your chances of a scan giving you useful results will be higher with well known brands. There are currently about 15,500 brands and companies covered.

12 comments

  1. [3]
    sparksbet
    Link
    I just got my invite and did a quick browse. I didn't look into it back when you first released, so this is very much first impressions. I actually really like how the scores are laid out and how...

    I just got my invite and did a quick browse. I didn't look into it back when you first released, so this is very much first impressions.

    I actually really like how the scores are laid out and how much detail you get about where the scoring comes from. I didn't see the older version, so I don't know how it was changed, but I didn't feel like the aggregate score was overly emphasized, which I like. I felt encouraged to dig deeper into the sub-scores, which presumably is the desired behavior.

    I absolutely love the lists and tagging and would like to see it expanded. The addition of the "still in Russia" list/tag is exactly the kind of thing I want -- but now that that one exists, the lack of a list of companies contributing to the genocide in Gaza seems egregious. Ofc it's very easy to find other lists of those companies online, but that fact makes the lack thereof here look kinda intentional. But I get that you're a one-man operation, so I assume it's probably just a "I haven't gotten to it yet" type thing.

    I'm honestly wondering how you expect updating these scores to scale up, but it seems like you're aiming for a community model? This is a UI/UX thing, but when I clicked the "suggest improvements" link on the front page, just getting redirected to the list of companies was confusing. It would be helpful to have a clearer workflow for submitting changes/updates, something that walks you through what's necessary to submit. This may also help with my biggest worry, which is that if/when updates are handled by a larger community, the quality of the sources and detailed info presented on the site would go down.

    A lot of the copy on the website presents a thin veneer of political neutrality despite these all being highly political issues. The left/right score not being counted into a company's overall score is the biggest tangible effect of this, though it's not too bad as the overall scores aren't too prominent anyway. I think it's a little bit silly to pretend that these issues are valued across the political spectrum or that there is not a very direct relationship between ethics and political affiliation, especially when it comes to the types of issues this website concerns itself with. But I understand that this type of song and dance is probably necessary if your intention is to make this more mainstream and potentially monetizable someday.

    I do worry, however, that this somewhat artificial line between what's politics and what's ethics. Are Chic-Fil-A's actions that harm LGBTQ+ people less an issue of human rights or ethics than other companies harming people because queer people are a political issue? (From what I can tell, their company isn't listed on the website yet, but this is just an example that came to mind.) I very much want to pull my support from companies that are particularly harmful when it comes to queer issues, and I don't think a binary "left/right" scale that doesn't contribute to the overall score really captures this. Maybe this is a job for future lists, but in that case I'd definitely suggest heavily prioritizing expanding the number and variety of lists. In any case, I don't think the current scoring for "politics" is useful at the moment given that Tesla is rated "strongly left". Whatever metric you use for that score, it absolutely needs to take some things into account that it is not currently counting if it rates a company owned by one of the most famous toxic right-wing figures in politics as "strongly left".

    3 votes
    1. [2]
      ethoscore
      Link Parent
      Thanks, this is great feedback! I'm happy to hear that the scoring made sense to you, and that you found the tags and lists. I'm sure some people will have different experiences but that's exactly...

      Thanks, this is great feedback!

      I'm happy to hear that the scoring made sense to you, and that you found the tags and lists. I'm sure some people will have different experiences but that's exactly the goal of some of the recent changes.

      About Gaza, yes that was my first thought too when I was working on the Russia tag. If there's a Russia tag, why isn't there a Gaza tag? And yes I haven't gotten to it yet. It could be more generic "supports genocide" or similar. Any thoughts?

      It's more than I'm expecting out of beta testers, but if you were a theoretical user who cared enough to contribute, you could suggest a Gaza tag, along with a list of candidates it should be applied to. The idea being that the things people care about can inform what gets added and how.

      I'm honestly wondering how you expect updating these scores to scale up, but it seems like you're aiming for a community model? This is a UI/UX thing, but when I clicked the "suggest improvements" link on the front page, just getting redirected to the list of companies was confusing. It would be helpful to have a clearer workflow for submitting changes/updates, something that walks you through what's necessary to submit. This may also help with my biggest worry, which is that if/when updates are handled by a larger community, the quality of the sources and detailed info presented on the site would go down.

      I'm glad you brought this up. Yes, community contributions are a big part of the model for making it work at scale. That combined with better systems that I have planned but don't need quite yet.

      Agreed about the suggest improvements link, it should link to a guide that I haven't written yet. This is a good excuse to bump that up higher on my list!

      About quality and community contributions, as you can imagine I've thought about that a lot. Right now community revisions go through an admin approval process, which solves quality but not scale. In the wonderful event that the volume gets to be too high, I have some options in mind. The eventual goal is a system where community also reviews and fact checks contributions, with a trust system and maybe escalations for admin review. I have a bunch of ideas for what that would look like but it's premature right now. Medium term there's a "superuser" tier that will have permissions to do some level of submission/revision review and then long term will have a higher level of editorial control over the community review process.

      I'm resisting the urge to make a wall of text about how I'm thinking all of the above could work!

      A lot of the copy on the website presents a thin veneer of political neutrality despite these all being highly political issues. The left/right score not being counted into a company's overall score is the biggest tangible effect of this, though it's not too bad as the overall scores aren't too prominent anyway.

      What makes it seem like a "thin veneer"? Do you mean that it's impossible to be politically neutral so the claim itself is questionable? If so I don't disagree, depending on how you look at it almost anything can be seen as political.

      One of the reasons politics doesn't go into overall score is that a lot of companies aren't involved in politics in any easily measurable way, which makes it hard to incorporate into an overall score. The other challenge is how to translate politics into scoring at all. What constitutes politics that makes a positive score contribution versus politics that subtracts? And then of course, political values vary from culture to culture.

      So I decided to leave it up to people to decide how much they want to factor politics into their decisions. For example, in the US under Trump, aspects of conservative politics violate a lot of shared values most of the world can agree on (or so I like to think). Meanwhile in other countries "conservative" is a completely legitimate political viewpoint.

      I think it's a little bit silly to pretend that these issues are valued across the political spectrum or that there is not a very direct relationship between ethics and political affiliation, especially when it comes to the types of issues this website concerns itself with. But I understand that this type of song and dance is probably necessary if your intention is to make this more mainstream and potentially monetizable someday.

      You have a point there in that pretty much any stance bleeds into politics in some way. At the same time what that looks like varies depending on who you are and where you live. What I'm striving for is to focus on things that are fairly universal, for example "we should stop actively trying to destroy the environment we all depend on". But of course there are people who would disagree with that. It's tricky and there's no perfect answer to what counts as universal.

      I'll happily accept suggestions for how politics could be handled differently if you have them!

      I do worry, however, that this somewhat artificial line between what's politics and what's ethics. Are Chic-Fil-A's actions that harm LGBTQ+ people less an issue of human rights or ethics than other companies harming people because queer people are a political issue?

      This is a good example of where it's impossible to avoid taking a stance and I see that as a necessary compromise. I think LGBTQ+ issues are universal human rights issues along with any other kind of bigotry. So in the Chic-Fil-A example, human rights is where their abuses would show up. Some of it could end up in ethics too.

      But I understand that this type of song and dance is probably necessary if your intention is to make this more mainstream and potentially monetizable someday.

      I haven't tackled monetization yet, but yes being overtly political would make it harder. What I know for sure is that any sort of funding has to be the kind that can't influence the content of the site/app. I've thought about funding it by building something completely unrelated that's more directly monetizable and routing the profits to Ethoscore. Donations are an option, though I don't know if that will turn out to be viable. In any case I lean towards some version of not-for-profit.

      I very much want to pull my support from companies that are particularly harmful when it comes to queer issues, and I don't think a binary "left/right" scale that doesn't contribute to the overall score really captures this. Maybe this is a job for future lists, but in that case I'd definitely suggest heavily prioritizing expanding the number and variety of lists

      Yes I think lists are a perfect for that. The hope is that people will create lots and lots of lists. It's really easy to do, there's no admin oversight at all people can just make them. Though of course toxic list content will be removed.

      I don't think the current scoring for "politics" is useful at the moment given that Tesla is rated "strongly left"

      Yes that's tricky. The politics metric is based on political giving and Musk doesn't route his spending through the company. So the company itself is indeed left leaning politically even if the owner and a lot of the user generated content isn't. Mostly Tesla the company doesn't participate in politics that much for a company their size. Or if they do they keep it hidden. The politics metric does have score impacts though, so well sourced impacts could change their politics score. The question is what would the impacts be based on?

      I think from a LGBTQ+ rights perspective the better metric to focus on is human rights. And of course lists bypass scores completely.

      On queer issues, is there anything else you'd like to see? Bearing in mind that it's early days and things are limited by things like Chick-Fil-A not having full score impacts yet.

      2 votes
      1. sparksbet
        (edited )
        Link Parent
        I think ultimately I agree with you that having a ton of lists labeling many different issues quite granularly is the best solution to a lot of the problems aggregating scores like that,...

        I think ultimately I agree with you that having a ton of lists labeling many different issues quite granularly is the best solution to a lot of the problems aggregating scores like that, especially for those like me who have particular issues they're especially passionate about. I think that's the right direction to move things in.

        I am still curious about the precise methodology for measuring political giving, though. Do only campaign donations and/or lobbying count? I was just watching a video about the current Nvidia debacle, and it was mentioned in the video that Nvidia made large donations to Trump's ballroom, seemingly in exchange for specific political favors. Yet Nvidia is rated as 92% left wing. I also think some transparency of where those ratings come from (in terms of just displaying which donations to what have influenced that score, presumably scraped from wherever you're getting that data from in the first place) would help with making it a little clearer why a company is rated a certain way on the left/right axis.

        And I think that the same philosophy that underpins combining the subsidiaries of mega-corporations is necessary for the owners of companies, at least when it comes to companies with owners that play visible, active roles in politics. You can't separate the actions of Musk from his companies any more than you can separate the actions of Nestlé from its many subsidiaries. In fact, one could argue that there's a much closer relationship in the case of the former than the latter.

  2. [2]
    MimicSquid
    Link
    Wait, I was the only person in the whole test who scanned something? It was just to show my wife how it worked. Which I guess is kind of the point. I showed my wife Mondelev's poor score, and she...

    Wait, I was the only person in the whole test who scanned something? It was just to show my wife how it worked. Which I guess is kind of the point. I showed my wife Mondelev's poor score, and she said she didn't need a website to tell her that a large food corporation was going to be shady as shit. For me it best serves a purpose when I run across something novel.

    2 votes
    1. ethoscore
      Link Parent
      Yep just you! Fortunately some of my personal invites have done a lot of scans so it's not entirely untested. Re: Mondelēz, yeah it's not a surprise that they suck. And they actually likely suck...

      Yep just you! Fortunately some of my personal invites have done a lot of scans so it's not entirely untested.

      Re: Mondelēz, yeah it's not a surprise that they suck. And they actually likely suck more than their current score, it hasn't had a full scoring run yet. The useful thing with a conglomerate like that is tying their 100+ brands to them (i.e. Nabisco). They also own brands you wouldn't expect like the highly health focused, erstwhile family owned Perfect Bar.

      2 votes
  3. [7]
    ethoscore
    Link
    Reply here for an invite (or send me a DM)

    Reply here for an invite (or send me a DM)

    1 vote
    1. [2]
      sparksbet
      Link Parent
      I'd be interested in an invite!

      I'd be interested in an invite!

      1 vote
      1. ethoscore
        Link Parent
        Sent, thanks for checking it out!

        Sent, thanks for checking it out!

        1 vote
    2. [2]
      TheDarkerZone
      Link Parent
      I'd be interested in checking this out!

      I'd be interested in checking this out!

      1 vote
    3. [2]
      all_summer_beauty
      Link Parent
      I'm hopeful that I'll have time to give this a look, I'd love an invite!

      I'm hopeful that I'll have time to give this a look, I'd love an invite!

      1 vote